What Mother-Child Talk Tells Us About Children's Language, Social
What mother-child talk tells us about children’s language, social, emotional and behavioural development: A community-based study of slow-to-talk children
Laura Jane Conway BA (Hons) ORCID ID: 0000-0002-0986-2964
Submitted in total fulfilment of the requirements of the degree of Doctor of Philosophy
November 2017
Murdoch Children’s Research Institute Department of Paediatrics The University of Melbourne
i Abstract
Language difficulties, including Developmental Language Disorder (DLD), often co-occur with social, emotional and behavioural (SEB) difficulties in childhood. It is unclear when and how the association between language and SEB difficulties emerges. Parenting behaviours are targeted in early interventions for both DLD and
SEB difficulties, yet there is minimal research into whether similar parenting behaviours may promote both language and SEB development. An understanding of when the association between language and SEB difficulties emerges, and whether there are specific parenting behaviours that could be used to identify and/or provide intervention for children experiencing co-occurring difficulties is needed.
This PhD was embedded within an existing prospective community-based study of children who were slow-to-talk at 18 months, Let’s Learn Language and its follow- up, Language for Learning. Data were available for 197 participants. Language and
SEB skills had been measured at 24, 36, and 48 months, and mothers and children video-recorded during free-play at home for 15 minutes at 24 months. Maternal use of responsive behaviours during these observations, including expansions, imitations and labels, was examined in an earlier study. For this PhD, a coding scheme was developed to measure further maternal behaviours: praise, missed opportunities, successful and unsuccessful directives, and emotion and mental state talk. An existing coding scheme was modified to rate interactive behaviours: Supported and Coordinated Joint
Engagement, and Fluency and Connectedness.
The first aim of the thesis was to examine how language skills were associated with SEB difficulties at 24, 36 and 48 months. Linear regression models were fitted
ii with and without cubic and quadratic terms to examine the shape of the associations between receptive and expressive language scores and internalising and externalising behaviour scores. The analyses controlled for environmental (e.g. socioeconomic status) and biological (e.g. birthweight) risk factors for language and/or SEB difficulties.
Receptive but not expressive language abilities were associated with internalising behaviours across the spectrum of language scores, and with externalising behaviours at very low or very high scores. These findings suggest receptive language difficulties may be integral to the emerging association between language and SEB difficulties.
The second aim was to explore the associations between maternal and interactive behaviours used during free-play and language outcomes at 24, 36 and 48 months using linear regression analyses. Simple slopes were used to examine whether the associations between the maternal behaviours, including the previously coded responsive behaviours, and language outcomes were modified by interactive behaviour ratings. Successful and unsuccessful directives were associated with poorer concurrent and later language scores, and the association was weaker when mothers also used expansions. This suggests that responsive behaviours may protect against the effects of intrusive behaviours. Supported Joint Engagement, the degree to which mother and child focused on the same object at the same time while the mother scaffolded the interaction, was positively associated with language outcomes. Supported Joint
Engagement also modified how expansions and imitations were associated with language outcomes such that at lower levels of engagement, the positive associations were stronger. The degree of Fluency and Connectedness was also positively associated with language scores, and modified the association between directives and language.
Together, these findings suggest that when maternal behaviours and the interactional
iii context are aligned in a manner advantageous to language development, children may be more likely to learn language.
The third aim was to explore whether the maternal and interactive behaviours, including the maternal responsive behaviours, were associated with SEB difficulties.
Higher use of unsuccessful directives was associated with poorer internalising behaviours at 24 and 36 months, resonating with research showing that controlling parenting is associated with early internalising problems. However, there was no evidence that the other behaviours were associated with SEB outcomes. The effects of the maternal responsive and interactive behaviours might be specific to language development, perhaps by providing direct language-learning opportunities. In contrast, any associations between these behaviours and SEB development may be via more complex pathways.
The findings from this thesis demonstrate that there are specific linguistic features of maternal input and situational factors of the mother-child interaction that are key to child language development. These same factors did not appear to be directly associated with SEB outcomes, except for intrusive directives. Knowing that receptive language difficulties, which are difficult to detect, can co-occur with SEB difficulties between 24 and 48 months may be critical to earlier identification of receptive language difficulties. However, specific parenting behaviours may not be useful for identifying or treating children with co-occurring difficulties. Further research is recommended to continue investigating how to best support children with co-occurring language and
SEB difficulties, and how to provide advice to their families.
iv Declaration
This is to certify that the thesis comprises only my original work towards the degree of Doctor of Philosophy except where indicated in the Preface. Due acknowledgement has been made in the text to all other material used. The thesis is fewer than 100,000 words in length, exclusive of tables, bibliographies and appendices.
Laura Jane Conway
v Preface
Work contained in this thesis is submitted solely for the purposes of the PhD candidature. The student did not use any third-party editorial assistance in preparing the thesis. The student led the writing of the two manuscripts that have been accepted for publication and of one other manuscript currently undergoing the peer review process.
The co-authors provided intellectual input into the design, coding, analysis, interpretation and writing. The student was responsible for addressing all feedback and revising the manuscripts accordingly, with advice from the co-authors as needed.
Work towards the thesis that has been submitted for other qualifications;
In November 2015, I was awarded a higher distinction (H1) in the University of
Melbourne, School of Global and Population Health module, Linear & Logistic
Regression. The statistical skills acquired from this qualification enabled me to conduct the analysis for my thesis competently and with confidence.
Work towards the thesis that was carried out prior to enrolment in the degree;
Prior to commencing my PhD, I was a Research Co-ordinator within the
Hearing, Language and Literacy Group at the Murdoch Children’s Research Institute
(MCRI) for six years. I was responsible for the day-to-day management of the large- scale population-based study, the Early Language in Victoria Study (ELVS). I worked alongside the team running Let’s Learn Language and Language for Learning, within which this thesis is embedded. I gained practical experience of data collection, using the language and SEB measures used in this thesis. This work laid the foundation for my degree.
vi Publications during candidature;
During my candidature, I contributed to ten publications that are in various stages of the publication process. I conceptualised, analysed and drafted the manuscripts and contributed 75% of the work for publications One, Two and Three, which address two of the aims of my thesis and comprise Chapters Five, Six and Seven. Publications
Four to Ten represent work related to the broader subject areas within which my thesis is based, but are not directly related to my thesis. Publications Four to Seven have been published, and Publications Eight to Ten are undergoing peer-review. A signed copy of the Co-author authorisation form and the Declaration for thesis with publication have been submitted with this thesis for Publications One and Two.
1. Conway LJ, Levickis P, Mensah F, Smith KA, & Reilly S. Associations between
expressive and receptive language and internalising and externalising behaviours:
Findings from a community-based slow-to-talk sample. International Journal of
Language and Communication Disorders, 52 (6): 839-853.doi: 10.1111/1460-
6984.12320
Original manuscript: Submitted on 28 September 2016. Accepted 1 March 2017
2. Conway LJ, Levickis, P.A., Smith, J., Mensah, F., Wake, M., & Reilly S.
Maternal communicative behaviours and interaction quality as predictors of
language development: Findings from a community-based study of slow-to-talk
toddlers. International Journal of Language and Communication Disorders.
Original manuscript: Submitted on 9 December 2016. Accepted 2 October 2017
3. Conway LJ, Levickis, P.A., Mensah, F., Smith, J., Wake, M., & Reilly S. (2017).
Association between joint engagement and language in a sample of slow-to-talk
vii toddlers. Journal of Child Language
Original manuscript: Submitted on 13 December 2016. Invited to resubmit on 2
March 2017. Revision was submitted on 27 May 2017. Invited to do second major
revision on 17 August 2017, due 13 February 2018
4. McKean C., Reilly S., Bavin EL., Bretherton L., Cini E., Conway L., Cook F.,
Eadie P., Prior, M., Wake M., & Mensah F. (2017). Language outcomes at 7 years:
early predictors and co-occurring difficulties. Pediatrics, 139 (3), doi:
10.1542/peds.2016-1684
Contribution: Conducted the preliminary data analysis, drafted the Methods section,
and commented on manuscript drafts
5. Smith, J, Levickis, P., Conway, L., Eadie, P., Bretherton, L., & Goldfeld, S.
(2017). Associations between maternal behaviours and child language at 2 years in
a cohort of women experiencing adversity. Infancy, doi:10.1111/infa.12200
Accepted 1 Aug 2017
Contribution: Helped develop coding scheme, conducted piloting, and contributed
to data analysis and interpretation
6. Smith, J, Levickis, P., Conway, L., Eadie, P., Bretherton, L., & Goldfeld, S. (2017).
Concurrent associations between maternal behaviours and infant communication
within a cohort of women and their infants experiencing adversity. International
Journal of Speech-Language Pathology, doi:10.1080/17549507.2017.1329458
Accepted 8 May 2017
Contribution: As for publication five, helped develop coding scheme, conducted
piloting of scheme, and contributed to data analysis and interpretation
viii 7. Kefalianos E., Reilly S., Onslow M, Block S., Packman A., Bavin E., Conway
L., Vogel A., Pezic A., and Mensah F. The history of stuttering by 7 years:
Follow up of a prospective community cohort. Journal of Speech, Language and
Hearing Research. doi:10.1044/2017_JSLHR-S-16-0205
Accepted 1 July 2017
Contribution: Managed aspects of data collection and cleaning, read and
commented on manuscript drafts
8. Eadie, P., Conway LJ, Hallenstein B, McKean M., Mensah F, & Reilly S.
(2017). The impact of early childhood language impairment on the quality of life
in leisure and recreation activities of children at 9 years of age. International
Journal of Language and Communication Disorders.
Invited to submit a revision 21 August 2017, due November 2017
Contribution: Conducted the data analysis, read and commented on manuscript
drafts
9. Levickis, P, Sciberras E., McKean, C., Conway, L., Pezic, A. Mensah, F., Bavin
E., Bretherton, L., Eadie, P., Prior, M., and Reilly, S. Language and social-
emotional and behavioural wellbeing from 4-7 years: A community-based study.
European Child & Adolescent Psychiatry.
Major revision re-submitted 22 September 2017. Accepted subject to minor
amendments 2 Nov 2017
Contribution: Conducted preliminary data analysis, read and commented on
early drafts
10. Smith, K. A., Iverach, L., Mensah, F., O’Brian, S., Kefalianos, E., Conway L.,
Rapee, R., & Reilly, S. Stuttering and anxiety: A comparison of clinical and
ix community samples. International Journal of Speech-Language Pathology
Submitted October 2017
Contribution: Helped to conceptualise the research question and analysis plan,
read and commented on drafts
Presentations during candidature:
I delivered seven national and international presentations during my candidature.
Presentations one to four were of my thesis findings. Presentations five to seven were related to my thesis, but did not directly contribute to it.
1. Conway LJ, Levickis, P.A., Smith, J., Mensah, F., Wake, M., & Reilly S. Maternal
communicative behaviours and interaction quality as predictors of child language
th development: A prospective community-based study of slow-to-talk toddlers. 14P P
International Association for the Study of Child Language, Lyon, France 17-21 July
2017
2. Conway LJ, Levickis, P.A., Mensah, F., Wake, M., & Reilly S. What mother-
toddler talk tells us about language, social, emotional & behavioural development:
A prospective community-based study of slow-to-talk toddlers. Presentation to
School of Human Communication Sciences, Sheffield University, UK, 21 July 2017
3. Conway L.J., Levickis P., Smith J., Mensah F., Wake M., & Reilly S. Maternal
communicative behaviours and interaction quality as predictors of child language
development: A prospective community-based study of slow-to-talk toddlers.
Centre of Research Excellence in Child Language Conference, Melbourne,
Australia, 20-21 April 2017
4. Conway L.J., Levickis P., Smith J., Mensah F. & Reilly S. The contribution of
maternal directives to child language development: A prospective community-based
x study of slow-to-talk toddlers. Gold Coast Health and Medical Research
Conference, Queensland, Australia, 1-2 December 2016
5. Conway L.J. et al. Language outcomes at 7 years: early predictors and co-occurring
difficulties in literacy, social-emotional and behavioural difficulties and Quality of
th Life. 14P P International Association for the Study of Child Language, Lyon, France,
17-21 July 2017
6. Conway L.J., Reilly S., Eadie P., Wake M., Mensah F. & Pezic A. Child language
& comorbidities at 7 years: A population based study. Child Language Seminar,
University of Warwick, UK, 20-21 July 2015
7. Conway L.J., Sciberras E., Mensah F., Pezic A., & Reilly S. The association
between child language problems and social, emotional & behavioural difficulties
from 4-7 years: A population based study. 13th International Congress for the Study
of Child Language, Amsterdam, The Netherlands, 14-18 July 2014
Community Engagement;
• Perfect Pitch Competitor – Women in Science Parkville Precinct (WISPP) and
National Australia Bank Science Communication Competition, October 2016
• What do speech pathologists do? Presentation to four-year-old kindergarten children
and staff. Tulip Street Kinder, Melbourne, September 2016
• Invited guest lecturer for University of Melbourne Masters of Speech Pathology
Course. Social and biological determinants of communication as part of
Communication across the lifespan module. 17 May 2016
• MCRI IMPACT Blog Post. Mother and child interactive behaviours as predictors of
later language skills. May 2016
xi Scholarships and Grants;
I have been supported during my PhD through a number of scholarships and grants for which I am grateful and acknowledge here.
• An Australian Government Research Training Program Scholarship (Previously
known as an Australian Postgraduate Award): 2014-2017
• Melbourne Children’s Postgraduate Health Research Top-up Scholarship; 2013-
2016 ($5,000 per annum)
• Murdoch Children’s Research Institute Student Conference Support Scheme; 2017
($2,500)
• Henry and Rachel Ackman Travelling Scholarship; 2017 ($3,000)
• International Association for the Study of Child Language, Student Bursary; 2017
($1,495)
• University of Melbourne, Department of Paediatrics Travelling Scholarship; 2015
• International Association for the Study of Child Language, Student Bursary; 2014
($2,000)
xii Acknowledgements
My sincere gratitude is extended to an entire community of people who have encouraged, supported and guided me through my PhD. Firstly, thank you to my supervisors, Prof Sheena Reilly, Dr Fiona Mensah, and Dr Penny Levickis. Your steadfast support, enthusiasm and cheer through my “wibbles and wobbles” has been invaluable. Thank you to Prof Melissa Wake who permitted me to work on the LLL /
L4L datasets and provided guidance to me as part of the supervisory team at key stages.
To my advisory panel, Prof Katrina Williams, Prof Tricia Eadie and Prof Margot Prior, thank you for always being available to chat over coffee. To my mentors, Prof Sylvia
Metcalf and Prof Jan Nicholson, a big thank you for helping me prioritise the many
‘opportunities’ I unwisely said ‘yes’ to.
Thank you to the members of the Centre for Research Excellence in Child
Language, and the Early Language in Victoria Study team, for your warmth and encouragement: Prof Edith Bavin, Dr Lesley Bretherton, Dr Cristina McKean, Prof
James Law, A/Prof Obi Ukoumunne, Prof Sharon Goldfeld, Dr Emma Sciberras, Dr Liz
Westrupp, and Dr Katy Mueller. Special thanks my fellow CRE students, Shannon
Bennetts and Peter Carew, whose camaraderie kept me sane. And to Jodie Smith who made the coding process much more enjoyable than it might otherwise have been – thank you so much for all the hours you spent helping me with piloting and reliability testing.
Thank you to the Murdoch Children’s Research Institute and the Department of
Paediatrics at the University of Melbourne for support throughout my candidature, and provision of academic and travelling scholarships without which I would not have been
xiii able to do this. Thanks also to the Let’s Learn Language and Language for Learning investigator teams, and the many families who took part. What a gift those 200 video recordings were to me, and what a privilege to observe so many wonderful women of
Melbourne playing with their toddlers and giving them the best start in life.
I have been fortunate to be surrounded by many hardworking, intelligent, and inspiring colleagues at MCRI and collaborating institutions. I have been motivated to strive harder by the excellence of their endeavours. Heartfelt thanks go to my wonderful
MCRI friends who have kept me company through the ups and downs of PhD life: Amy
Watts, Eileen Cini, Petrea Cahir, Lauren Pigdon, Elaina Kefalianos, Fallon Cook, Anna
Duncan, Angela Morgan, Vera de Costa, and Kerry Ttofari-Eecen. A special thanks to my critical friend, Kylie Smith, who has weathered this storm with me, with whom I have celebrated and commiserated, and without whom I might not have got out alive.
To have undertaken a PhD at this stage of life was insanity, but having a great friend with whom to share the experience was the most extraordinary good luck.
And finally, thank you to my family for believing in me. To my mum, dad,
Claire and Mark who always knew I had it in me, and whose pride bolstered me on the grimmest of days. Thank you, dad, for correcting the final version of my thesis and sorry for being “hyphen-happy”. To my mother-in-law, Judy, whose help with the kids and household was immeasurable. And to my husband, Marc, to whom I am forever indebted for his unwavering love, support and pride.
I dedicate this thesis to Isla May and Liam Patrick. Yes, Mum has finally finished her PhD.
xiv Abbreviations and Glossary of Terms
ASD Autism Spectrum Disorders
CELF-P2 Clinical Evaluation of Language Fundamentals Preschool
2nd Edition
CBCL Child Behaviour Checklist
C-JE Coordinated Joint Engagement
DLD Developmental Language Disorder
ELVS The Early Language in Victoria Study
Expressive language The ability to encode ideas into language forms and
symbols
FC Fluency and Connectedness
Interaction Communicative exchange lasting three seconds minimum
JE Joint Engagement
L4L Language for Learning study
Lexicon Mental dictionary of words or the knowledge of
vocabulary
LLL Let’s Learn Language study
M Mean
MCDI-UKSF MacArthur-Bates Communicative Development
Inventory: UK Short Form
MCH Maternal and Child Health
MIBS Maternal and Interactive Behaviours Study (this PhD)
MRBS Maternal Responsive Behaviours Study
PLS-4 Preschool Language Scale – 4th edition
xv Pragmatics The appropriate use of language in context
Receptive language The ability to understand the meanings that others have
expressed using language
RCT Randomised controlled trial
SD Standard Deviation
SDQ Strengths and Difficulties Questionnaire
SEIFA Socio-Economic Indexes for Areas (SEIFA) Index of
Relative Social Disadvantage (IRSD)
SEB Social, emotional and behavioural
Secs Seconds
Semantics The meaning of words and sentences
S-JE Supported Joint Engagement
Slow-to-talk Scoring at or below 20th percentile on expressive
vocabulary at 18 months old
SSLM Sure Start Language Measure
Utterance A unit of speech bounded by silence
xvi Table of Contents
Abstract41T 41T ...... ii
Declaration41T 41T ...... v
Preface41T 41T ...... vi
Acknowledgements41T 41T ...... xiii
Abbreviations41T and Glossary of Terms 41T ...... xv
141T 41T Introduction41T and Thesis Summary41T ...... 1
1.141T 41T Research41T Overview41T ...... 2
1.241T 41T Research41T Context 41T ...... 5
1.341T 41T Thesis41T Overview41T ...... 6
241T 41T Literature41T Review and Rationale 41T ...... 8
2.141T 41T Overview41T 41T ...... 8
2.241T 41T Typical41T Language Development 41T ...... 9
2.341T 41T Social,41T Emotional, and Behavioural Development41T ...... 12
2.441T 41T Problems41T that Co-occur with DLD and SEB Difficulties 41T ...... 13
2.541T 41T The41T Association between DLD and SEB Difficulties...... 41T 14
2.641T 41T Risk41T Factors for DLD and SEB Difficulties41T ...... 18
2.741T 41T Parenting41T and Child Development 41T ...... 23
2.841T 41T 41TInterventions for DLD and SEB Difficulties41T ...... 33
2.941T 41T Similarities41T between DLD and SEB Interventions41T ...... 40
2.1041T 41T Rationale41T for PhD Study41T ...... 48
2.1141T 41T Summary41T of Literature Review41T ...... 48
xvii 341T 41T Maternal41T and Interactive Behaviours 41T ...... 50
3.141T 41T Overview41T 41T ...... 50
3.241T 41T Maternal41T Responsive Behaviours Study41T ...... 51
3.341T 41T 41TIdentifying Behaviours: Step 1 41T ...... 52
3.441T 41T 41TIdentifying Behaviours: Step 2 41T ...... 63
3.541T 41T Praise41T 41T ...... 65
3.641T 41T 41TIntrusive Directives 41T ...... 67
3.741T 41T Emotion41T and Mental State Talk41T ...... 70
3.841T 41T Fillers41T and Missed Opportunities 41T ...... 72
3.941T 41T Joint41T Engagement 41T ...... 74
3.1041T 41T Fluency41T and Connectedness 41T ...... 75
3.1141T 41T Summary41T of Maternal and Interactive behaviours41T ...... 76
3.1241T 41T Aims41T Overview41T ...... 76
441T 41T Methodology41T 41T ...... 78
4.141T 41T Overview41T 41T ...... 78
4.241T 41T Study41T Context 41T ...... 79
4.341T 41T Participants41T 41T ...... 87
4.441T 41T Coding41T Overview41T ...... 89
4.541T 41T Development41T of Scheme One: Maternal Behaviours 41T ...... 90
4.641T 41T Development41T of Scheme Two: Interactive Behaviours 41T ...... 98
4.741T 41T Reliability41T 41T ...... 105
4.841T 41T Measures41T 41T ...... 110
4.941T 41T Analysis41T 41T ...... 119
4.1041T 41T Summary41T 41T ...... 125
xviii 541T 41T Language41T and Social, Emotional and Behavioural Development 41T .. 127
5.141T 41T Overview41T 41T ...... 127
5.241T 41T PDF41T of Manuscript 41T ...... 128
5.341T 41T Summary41T 41T ...... 147
641T 41T Mother41T -child Interaction and Language Outcomes: PART ONE 41T ... 149
6.141T 41T Overview41T 41T ...... 149
6.241T 41T Author41T Accepted Manuscript 41T ...... 150
6.341T 41T Supplementary41T Results 41T ...... 199
6.441T 41T Summary41T 41T ...... 203
741T 41T Mother41T -child Interaction and Language Outcomes: PART TWO41T .. 204
7.141T 41T Overview41T 41T ...... 204
7.241T 41T Abstract41T 41T ...... 206
7.341T 41T Background41T 41T ...... 206
7.441T 41T Method41T 41T ...... 210
7.541T 41T Results41T 41T ...... 217
7.641T 41T Discussion41T 41T ...... 227
7.741T 41T Conclusions41T 41T ...... 233
7.841T 41T Summary41T 41T ...... 233
841T 41T Mother41T -child Interaction and Social, Emotional and Behavioural
Outcomes 41T ...... 234
8.141T 41T Overview41T 41T ...... 234
8.241T 41T 41TIntroduction 41T ...... 235
8.341T 41T Results41T 41T ...... 235
8.441T 41T Discussion41T 41T ...... 240
xix 8.541T 41T Summary41T 41T ...... 243
941T 41T Discussion41T and Conclusions...... 41T ...... 245
9.141T 41T Overview41T 41T ...... 245
9.241T 41T Summary41T and Interpretation of Key Findings41T ...... 246
9.341T 41T Strengths41T and Limitations 41T ...... 253
9.441T 41T Clinical41T Implications41T ...... 258
9.541T 41T Future41T Research41T ...... 260
9.641T 41T Conclusions41T 41T ...... 263
References41T 41T ...... 265
Appendices41T 41T ...... 317
xx List of Tables
Table41TU 1 Strategies Taught in Early Interventions for DLD and SEB Difficulties U41T ...... 41
Tabl41TU e 2 Existing Communicative Behaviour Coding Schemes ConsultedU41T ...... 54
Table41TU 3 Definitions of Maternal Responsive Communicative Behaviours U41T ...... 85
Table41TU 4 Comparison of MRBS Sample and MIBS Sample U41T ...... 88
Table41TU 5 Summary of Behaviours Piloted for Coding Scheme OneU41T ...... 94
Table41TU 6 Summary of Behaviours Piloted for Coding Scheme TwoU41T ...... 102
Table41TU 7 Final Maternal Behaviours Coding Scheme OneU41T ...... 106
Table41TU 8 Final Interactive Behaviours Coding Scheme Two U41T ...... 107
Table41TU 9 Reliability Ratings for Coding Schemes U41T ...... 109
Table41TU 10 Outline of Measures used in MIBS by Age and Study in which CollectedU41T .... 111
Table41TU 11 Association between Emotion and Mental State Talk and Language ScoresU41T 200
Table41TU 12 Sample Characteristics and Language ScoresU41T ...... 212
Table41TU 13 Joint Engagement States Coding Scheme...... U41T ...... 215
Table41TU 14 Definitions of Maternal Responsive Communicative Behaviours U41T ...... 217
Table41TU 15 Joint Engagement and Maternal Responsive Behaviours Summary U41T ...... 218
Table41TU 16 Correlation Matrix of Responsive Behaviours, Joint Engagement, and
Language ScoresU41T ...... 219
Table41TU 17 Associations between Joint Engagement State and Language Scores U41T ...... 221
Table41TU 18 S-JE Interaction Effects on Association between Responsive Behaviours and
Language U41T ...... 223
Table41TU 19 Simple Slope Statistics for Association between Maternal Behaviours and
Language Scores by S-JE U41T ...... 225
Table41TU 20 Association between Maternal and Interactive Behaviours and SEB ScoresU41T 238
Table41TU 21 Summary of Findings U41T ...... 249 xxi List of Figures
Figure41TU 1 Flowchart of participants and measures U41T ...... 80
Figure41TU 2 Flowchart of participants through LLL Survey and Trial U41T ...... 84
Figure41TU 3 Transformation of Data from Wide-form to Long-form U41T ...... 123
Figure41TU 4 Associations between Maternal Behaviours and Language Scores at 24 months
by S-JEU41T ...... 224
Supplementary41TU Figure 5 Expressive and Receptive Language Score Distributions U41T .... 226
xxii List of Appendices
Appendix A: LLL Parent/Guardian Information Statement and consent form……….318
Appendix B: LLL baseline questionnaire……………………………………………..323
Appendix C: LLL 18 months questionnaire ………………………………………….330
Appendix D: LLL Human Research Ethics Committee Approval certificate………...333
Appendix E: L4L Parent/Guardian Information Statement and consent form………..334
Appendix F: L4L Human Ethics Research Committee Approval certificate…………341
xxiii Third Party Copyright Material
The accepted manuscripts have been included in Chapters Five and Six. Both papers have been accepted by the International Journal of Language and
Communication Disorders (IJLCD). IJLCD has an embargo period of 12 months post online publication.
The latest version of the manuscript going through the review process at the
Journal of Child Language is included in Chapter Seven. This will be covered by an embargo period if accepted.
xxiv 1 Introduction and Thesis Summary
The following chapter provides an introduction to the research focus of this
PhD, a brief overview of the research context within which this thesis is situated, and an overview of the thesis structure.
1 1.1 Research Overview
Language difficulties are one of the most common developmental problems in childhood, affecting up to 20% of four-year-old children (Reilly et al., 2010). Also known as Developmental Language Disorder (DLD) (Bishop, Snowling, Thompson,
Greenhalgh, & CATALISE-Consortium, 2016), these difficulties are associated with both concurrent and long-term negative outcomes in a range of areas including education, employment, and social, emotional and behavioural (SEB) difficulties
(Dockrell, Lindsay, Roulstone, & Law, 2014; Law & Stringer, 2014). However, how to identify and treat these children in the early years is challenging. Biological and environmental risk factors identified to date are not strongly predictive of later language or SEB difficulties (Bayer et al., 2012; Reilly et al., 2010; Zubrick, Taylor, Rice, &
Slegers, 2007).
Research into children with either language or SEB difficulties has revealed high rates of co-morbidity (e.g. Beitchman, Brownlie, & Bao, 2014; Benner, Nelson, &
Epstein, 2002; Conti-Ramsden & Botting, 2008). Further, there is increasing evidence that the association between language abilities and SEB difficulties exists in clinical and population samples (Girard, Pingault, Doyle, Falissard, & Tremblay, 2016; Petersen et al., 2013; St Clair, Pickles, Durkin, & Conti-Ramsden, 2011). Research using routinely collected administrative population data found 3% of school entrants in Victoria,
Australia in 2011 experienced both language and SEB difficulties (Hughes, Sciberras, &
Goldfeld, 2016). A prospective population-based study found children with language difficulties had elevated levels of emotional and conduct difficulties across childhood until at least age 11-years (Yew & O'Kearney, 2015a, 2015b). However, it is unclear at
2 what age the association emerges, how the association might progress, and which language and SEB domains are affected.
Reasons for the association between DLD and SEB difficulties are unclear.
Some suggest SEB difficulties emerge first, and lead to language difficulties (Bagner,
Garcia, & Hill, 2016), whilst others suggest it is language difficulties that lead to SEB problems (Bercow, 2008; Salmon, O’Kearney, Reese, & Fortune, 2016). It is also possible that there may be an underlying vulnerability common to both areas. Many theories that attempt to explain the co-occurrence include a focus on the child and parent-child interaction (Hughes et al., 2016). For this reason, parenting (often mothering, or maternal) behaviours, have been explored in relation to language and SEB development.
Responsive parenting characterised by warm, sensitive behaviours, low psychological control, and appropriate behavioural control is associated with adaptive child development (Deater-Deckard & Petrill, 2004; Landry, Smith, & Swank, 2006;
O'Connor, 2002; Treyvaud et al., 2009). Hostile, intrusive or neglectful parenting in contrast is associated with poorer developmental outcomes, including internalising and externalising behaviour problems (Boyle et al., 2004). Subsequently, parent training or coaching is a common method for intervention for SEB difficulties in the preschool period (Salmon et al., 2016).
Parent communicative behaviours that are responsive, timely and contingent to the child are associated with positive language development (Bornstein, Tamis-
LeMonda, Hahn, & Haynes, 2008; Girolametto, Bonifacio, et al., 2002). Many are already targeted in parent-led language intervention programmes such as Hanen’s You
3 Make the Difference (Manolson, Ward, & Doddington, 1995) and the Heidelberg
Parent-Based Language Programme (Buschmann et al., 2009), but not all parenting behaviours have been tested with methodological rigour or in RCTs. Specific maternal responsive behaviours such as expansions, imitations and questions, predict better later language abilities (Levickis, Reilly, Girolametto, Ukoumunne, & Wake, 2014; Masur,
Flynn, & Eichorst, 2005). However, Levickis et al (2014) found no evidence of an association between supportive directives or responsive questions, which are targeted in interventions, and better language outcomes. Labels were also found to be associated with poorer language outcomes (Levickis et al. 2014). Given these unexpected findings, and the range of parenting behaviours targeted in interventions that have yet to be investigated either at all, or in a methodologically rigorous way, further research is required. If other parenting behaviours are found to be associated with language outcomes, they could be trialled in intervention studies for children with DLD.
Since language and SEB difficulties appear to co-occur, it is possible that parenting behaviours associated with one difficulty may also be associated with the other (Salmon et al., 2016; Tempel, Wagner, & McNeil, 2009). To date, there is limited research investigating whether interventions targeting one difficulty are also associated with the other (Bagner et al., 2016; Landry et al., 2006). A promising avenue of research may be to investigate whether parenting behaviours targeted in an intervention for one difficulty, for example responsive and intrusive behaviours, are also associated with the other difficulty. This information could contribute to the discussion about the role of intervention to improve language as a means to improve behaviour (Law & Stringer,
2014), and vice versa. For example, providing effective language intervention to children with hyperactivity is challenging. Identifying a parenting behaviour that
4 addresses both the language and the hyperactivity would be clinically applicable and useful.
This PhD study aimed to identify maternal and interactive behaviours associated with both language and SEB development. Using a large sample, representative of the community at recruitment, the findings have the potential to inform both language development theory and clinical practise, as well as to foster cross-disciplinary approaches to commonly co-occurring childhood conditions.
1.2 Research Context
To orientate the reader to the context in which this research study was located, a brief overview is provided. The study was nested within a cluster randomised controlled trial. Participants in the broader trial were recruited at age 12 months, and parents completed an expressive vocabulary screen at child age 18 months. Participants scoring at or below the 20th percentile were recruited into the trial, Let’s Learn Language, and a subsequent trial at 48 months, Language for Learning. Parents reported on their child’s
SEB difficulties at 24, 36 and 48 months. Mother and child were video-recorded during a 15-minute free-play session in their home when the child was 24 months. Children’s language was assessed at the same time point, and additionally at age 36 and 48 months.
In the current study, specific maternal and interactive behaviours were coded from the video recordings of the mother-child free-play sessions. These behaviours, described later, were analysed as potential predictors of child language and SEB outcomes at 24,
36 and 48 months. Detailed information about the procedures and measures will be provided later in the thesis.
5 1.3 Thesis Overview
In addition to this introductory chapter, this thesis comprises eight chapters.
Below is a brief outline of the contents of each.
Chapter Two presents the literature on parent-child interaction in the context of language and SEB development. The natural history of language and SEB development and difficulties is discussed, before exploring what is known about the association between these difficulties. The rationale for the thesis is presented at the end of this
Chapter. In Chapter Three, specific maternal and interactive behaviours that may contribute to both domains of development are introduced. Finally, the research aims and hypotheses are presented.
Chapter Four describes the study methodology, including participant recruitment, data collection, main predictor and outcome measures, and the analysis plans. It also describes the development of the two coding schemes, one for the maternal behaviours and one for the interactive behaviours. The process of developing the coding schemes was a major component of this PhD, from selecting appropriate behaviours, through to conducting reliability testing.
Chapters Five to Eight are the results chapters of the thesis. Chapters Five and
Six contain copies of two articles accepted for publication in the International Journal of Language and Communication Disorders. Chapter Seven comprises a manuscript ready for re-submission to the Journal of Child Language. This has undergone two rounds of peer-review, and a third revision has been invited. The version contained in the Chapter includes the changes recommended by the reviewers, and will be
6 resubmitted in November 2017. Chapter Eight presents and discusses the findings of the association between the maternal and interactive behaviours and SEB difficulties.
Chapter Nine presents a discussion of the main findings from this PhD, its strengths and limitations, and its clinical implications. Finally, it suggests avenues for future research.
7 2 Literature Review and Rationale
2.1 Overview
The literature review is presented across two chapters. Chapter Two describes early language and social, emotional and behavioural (SEB) development, and what is known about the co-occurrence of difficulties in these areas. Next, risk and protective factors for language and SEB difficulties are described, with a focus on the role of parenting. Similarities between parent-mediated intervention approaches for SEB and language difficulties are then considered, before examining the evidence for whether interventions for one difficulty might also address the other. Finally, the rationale for the PhD is presented. Chapter Three introduces the study within which this PhD is embedded. Specific maternal and interactive behaviours are described in relation to their association with language and SEB development. The aims and hypotheses are presented at the end of Chapter Three.
The following two sections lay the groundwork for the rationale of this thesis, introducing language development (section 2.2) before moving on to introduce SEB development (section 2.3).
8 2.2 Typical Language Development
It is impossible to overstate how important it is for children to acquire language skills. Learning how to use language is crucial for their educational, interpersonal and psychosocial development (C. J. Johnson, Beitchman, & Brownlie, 2010; Schoon,
Parsons, Rush, & Law, 2010). It is a major developmental milestone because it forms the foundation on which so many other competencies can develop (Zubrick, Taylor, &
Christensen, 2015). With expressive language skills, children can use words and gestures in sentences to participate in the world around them, and with receptive language skills they can understand and make sense this world.
2.2.1 Overview of Early Language Development
Language develops along a trajectory marked with milestones. These milestones are used by parents, health and education professionals to monitor children’s early development. During the pre-linguistic phase (0 to 12 months), infants communicate by crying, vocalising, sharing affect, and using nonverbal signals e.g. eye gaze, pointing, and showing. From around 9 months, they begin to use intentional communication to convey imperatives and declaratives, such as pointing and waving “bye-bye”. Infants utter their long-awaited first word around their first birthday, after which they learn five to 10 new words every week. As children’s comprehension skills increase, they begin to follow one- and two-step commands. Between 18 and 24 months, children’s range of intents and frequency of communication increases rapidly, from one communicative act per minute to up to seven (Paul & Shiffer, 1991; Wetherby, Cain, Yonclas, & Walker,
1988).
9 From around two years of age, children begin to combine words into sentences, and use increasingly complex grammar. They engage in symbolic play and start to use
‘decontextualised’ language to talk about objects and events outside of their immediate environment. Their conversational ability grows as they initiate and respond more frequently and appropriately, such that by age three they can take part in sustained conversations (Paul & Shiffer, 1991). Children learn how and when to approach someone, negotiate joint attention, share conversational turns, and repair misunderstandings (Rice, Sell, & Hadley, 1990). By age four, most children can use language to maintain interactions, empathise, and reason. Such social conversational skills shape their future behaviour and relationships (Scaramella & Leve, 2004), starting with their transition into school. Children failing to meet these milestones often come to the attention of health and education professionals and may be diagnosed with developmental language disorder (DLD).
2.2.2 Developmental Language Disorder (DLD)
Given the relatively short time in which language skills are acquired, it is unsurprising that language difficulties are one of the most prevalent developmental problems in early childhood (Norbury et al., 2016). Estimates of its prevalence vary between studies depending on how language difficulties are defined, the inclusion criteria used, and the age of the cohorts. For example, “Specific Language Impairment”
(SLI) was defined as low language skills in the absence of intellectual difficulties.
Tomblin et al. (1997) reported a prevalence rate of 7.0% for SLI in their sample of
American six-year-old children (n = 7217). In contrast, Reilly et al. (2010) defined “low language” as scoring ≥ 1.25 SD below the mean on expressive and/or receptive language assessments. They reported a prevalence rate of 20.6% for low language
10 among Australian four-year-old children (n = 1596). Once the SLI criteria were applied to this sample by excluding participants with non-verbal intelligence >1 SD below the mean, the rate decreased to 17.2%. Beitchman, Nair, Clegg, and Patel (1986) applied the term “language disorder” to their sample of Canadian five-year-old children (n =
1655), reporting 8.0% with the condition. More recently, Norbury et al. (2016) estimated the prevalence of “language disorder” in English seven-year-old children to be 7.58%.
A recent multinational and multidisciplinary Delphi consensus study of problems of language development recommended changing the approach to diagnosing language difficulties from a focus on formal assessment scores to a focus on the functional impacts of the difficulties (Bishop, Snowling, Thompson, Greenhalgh, &
CATALISE-2-Consortium, 2017; Bishop et al., 2016). The consortium proposed the term ‘Developmental Language Disorder’ (DLD) to refer to difficulties that cause functional impairment in everyday life and are associated with poor prognosis. The term excludes language difficulties associated with differentiating conditions in which the language disorder occurs as part of a more complex pattern of impairments (i.e. biomedical conditions including brain injury, certain neurodegenerative conditions, cerebral palsy, and oral language limitations associated with sensori-neural hearing loss). However, the consortium recommended that the term apply to language difficulties regardless of the presence of risk factors, other neurodevelopmental disorders (e.g. attention deficit hyperactivity disorder) or a mismatch between verbal and nonverbal ability (the latter being the distinction used for SLI). DLD will be used to refer to language difficulties throughout this thesis.
11 Having described the early developmental trajectory of typical language acquisition, and introduced DLD, the next section will describe the second outcome of interest in this thesis, SEB development. After introducing SEB difficulties, the chapter will bring together the two developmental areas by describing their co-occurrence
(section 2.5) and their shared risk factors (section 2.6).
2.3 Social, Emotional, and Behavioural Development
The development of social skills, emotional skills and appropriate behaviour occurs throughout childhood, continuing into adolescence. These skills are necessary for individual well-being and for full participation in the community. Social skills include learning how to make friends, approach people, and resolve social conflict and disagreement. Emotional skills include understanding and managing own and others’ emotions, coping with emotional situations, and empathising with others (Cross, 2011).
Finally, behavioural skills include paying attention, regulating own behaviour, and co- operating with others. Much like language development, SEB skills seem to develop effortlessly for most children. However, parents have an important role in modelling their use, and helping their children navigate the social world until their skills develop.
As for language development, a significant minority of children encounter difficulties with their SEB development. SEB difficulties are commonly defined as
‘internalising’ or ‘externalising’. In young children, internalising difficulties include generalised and separation anxiety (e.g. expressing worry over little things, being clingy to a parent), depression (e.g. being unhappy, sad, fearful), being withdrawn (e.g. refusing to play, avoiding eye contact, moving away), or having somatic complaints
(e.g. stomach aches, headaches, vomiting) (Bayer et al., 2006). Externalising difficulties
12 include being aggressive (e.g. screaming, being mean, destroying things) or being inattentive (e.g. problems sitting still, concentrating, finishing tasks).
2.4 Problems that Co-occur with DLD and SEB Difficulties
DLD and SEB difficulties in childhood both pose public health and education concerns (Bayer et al., 2012; Bishop et al., 2016; Law, Reilly, & Snow, 2013). Both have high prevalence, disrupt everyday family and school life, and are associated with an array of poorer outcomes that resonate into adulthood. DLD is associated with a cascade of negative outcomes from literacy difficulties, to disengagement from school, academic under-achievement, restricted employment opportunities, interpersonal problems, and, in some circumstances, to criminality (Beitchman et al., 2014; Conti-
Ramsden & Durkin, 2016; Dockrell et al., 2014; Law & Stringer, 2014; Snowling, Duff,
Nash, & Hulme, 2015). Childhood SEB difficulties are also associated with poorer academic, interpersonal and employment outcomes, as well as anti-social behaviour in adolescence and adulthood (Bayer, Sanson, & Hemphill, 2009; Brennan, Shaw,
Dishion, & Wilson, 2012; Halfon & Hochstein, 2002).
Unfortunately, a significant proportion of children experience co-occurring DLD and SEB difficulties. As mentioned earlier, routinely collected administrative data suggests 3% of children at school entry, or to one to two children per classroom, have co-occurring DLD and SEB difficulties (Hughes et al., 2016). These children face an accumulation of risk for poorer outcomes. Early identification and intervention either preventively or therapeutically is important to eliminate or alleviate the risks they face.
However, gaps exist in our understanding of how early language and SEB difficulties
13 are associated. These are explored in the following section, before discussing intervention approaches.
2.5 The Association between DLD and SEB Difficulties
Numerous studies have shown that children with DLD are more likely than those with typical language abilities to experience SEB difficulties (Beitchman et al.,
2014; Conti-Ramsden & Botting, 2008; Irwin, Carter, & Briggs-Gowan, 2002; Lindsay
& Dockrell, 2012; van Daal, Verhoeven, & van Balkom, 2007). A meta-analysis of eight prospective cohort studies of children with SLI (defined as having delayed language with non-verbal IQ in the normal range) confirmed that three to eight-year- olds with SLI were over two times more likely to have externalising behaviour problems and almost twice as likely to have internalising behaviour problems than their typically developing peers at follow-up (between two and 12 years later) (Yew &
O'Kearney, 2013). Worryingly, the behaviour problems experienced by the children with SLI were also likely to be in the clinical range, indicating severe problems requiring specialist help.
Similarly, a review of studies investigating children with SEB difficulties revealed that 71% were at risk of unrecognised language difficulties (Benner et al.,
2002). Language difficulties have also been reported in samples of seven to nine year old children with attention deficit and hyperactivity disorder (Helland, Biringer,
Helland, & Heimann, 2012), in eight to 16-year-old boys excluded from school, and at risk of exclusion from school due to behaviour problems (Clegg, Stackhouse, Finch,
Murphy, & Nicholls, 2009; Ripley & Yuill, 2005), and in seven to 14 year old children referred to child psychiatric services (Cohen et al., 1998).
14 There is some evidence that the association between DLD and SEB difficulties may differ depending on the particular difficulties the child is experiencing. For example, van Daal et al. (2007) found phonological problems in children with DLD were associated with both internalising and externalising problems, whilst semantic problems were associated with externalising problems only. Ripley and Yuill (2005) reported that children excluded from school due to behaviour problems were more likely to have expressive language difficulties than receptive difficulties, although this was not found in an older sample of boys at risk of exclusion (Clegg et al., 2009).
Population-based studies make an important contribution to the literature by examining the co-occurrence of language and SEB difficulties across the full distribution of both domains, rather than focusing on those with more severe and complex difficulties who often comprise clinical samples (Plomin, Price, Eley, Dale, &
Stevenson, 2002). In fact, these population-based studies have also found evidence for the association between language and SEB difficulties (e.g. McKean et al., 2017;
Whitehouse, Robinson, & Zubrick, 2011). Some also suggest the association might be determined by the specific difficulties experienced by the child (e.g. Clegg, Hollis,
Mawhood, & Rutter, 2005). Nevertheless, neither the clinical studies nor the population-based studies have yet revealed at what age the association first emerges, nor which domains of language or SEB functioning might be particularly at risk at young ages. This is important information for health professionals seeing young families, and for clinicians planning early intervention services.
15 2.5.1 Nature of the Association
There are different levels of explanation that attempt to explain the association between language and SEB difficulties. Overarching theories of language learning and
SEB development provide useful frameworks to explore the association. For example, the socio-pragmatic account of language development (e.g. Bruner, 1975; Tomasello,
Kruger & Ratner, 1993) proposes that language acquisition depends on social cognition and prosocial motivation. This means that to learn language, infants must develop an understanding of others as intentional beings, a motivation to understand these intentions and communicate their own, and an understanding of attentional states.
Likewise, theories of social cognitive development emphasise children’s language skills
(e.g. Astington & Baird, 2006; Carpendale and Lewis, 2006). For example, acquiring vocabulary about psychological states helps children understand and interpret their own and other people’s behaviours, feelings and beliefs, and to explain themselves and their actions to others (Dunn, Brown, Slomkowski, Tesla & Youngblade, 1991; Kostelnik,
Whiren, Soderman, & Gregory, 2006; Moore, Furrow, Chiasson and Patriquin, 1994).
These theories suggest that language and SEB development will be linked from infancy, meaning it may be difficult to separate the domains.
Several theories attempt to explain the nature of the association between language and behaviour at a level more proximal to the child’s language and SEB profile (see summaries by Dionne, Tremblay, Boivin, Laplante, & Perusse, 2003;
Salmon et al., 2016; Yew & O'Kearney, 2013). These are not necessarily mutually exclusive or incompatible with each other or the overarching theories, rather they represent levels of explanations. The main distinguishing feature between these perspectives is how they conceptualise the proposed direction of effect. Some argue that
16 SEB difficulties give rise to language difficulties by disrupting and limiting the quantity and quality of interactions because parents’ attention is occupied with managing their behaviour. This may reduce child exposure to rich linguistic input and impact language development (Bagner et al., 2016; Carpenter & Drabick, 2011).
Other researchers suggest that language difficulties lead to SEB difficulties, including withdrawal, frustration or aggression, because needs go unmet or intentions are misunderstood (Bercow, 2008; Cole, Armstrong, & Pemberton, 2010; Morgan,
Farkas, Hillemeier, Hammer, & Maczuga, 2015; Roben, Cole, & Armstrong, 2013;
Salmon et al., 2016). Since language is important for a range of developing competencies, any language difficulties may result in a cascade of negative effects, including SEB difficulties (Masten & Cicchetti, 2010). For example, DLD may result in learning difficulties, which affect children’s emerging self-concept and self-efficacy, leading to school disengagement, and social and mental health problems (Beitchman,
Brownlie, et al., 1996; Tomblin, Zhang, Buckwalter, & Catts, 2000). In contrast, children with good language skills may develop strong social skills at an earlier age through being able to engage in high quality interactions. Strong social skills may protect children from taking antisocial pathways (Dionne et al., 2003).
A different level of explanation again considers the possibility that the co- occurrence between language and SEB difficulties may be due to a shared biological
(e.g., low birthweight, low intelligence, genetics) or environmental risk (e.g., low socio- economic status, poor parenting skills) which create broad developmental vulnerabilities (Caulfield, Fischel, DeBaryshe, & Whitehurst, 1989; Dionne et al.,
2003). Low non-verbal intelligence, for example, may lead to a slower rate of language acquisition, and to delayed social competence which hinders SEB development (Dionne
17 et al., 2003; Dodge, 1990). Or, poor parenting may contribute to low quality verbal input, and poor disruptive behaviour management, resulting in the simultaneous development of language and SEB difficulties (Dionne et al., 2003).
Once established, the association between DLD and SEB difficulties may be punitively transactional as SEB difficulties may exacerbate language difficulties, which further exacerbate SEB difficulties and so on (Rutter & Lord, 1987). An important contribution to this debate has been made by Girard and colleagues who analysed language and SEB data collected concurrently and over time in two population-based studies in the UK and Canada. Their findings suggest a reciprocal rather than unidirectional association between language and SEB difficulties from 17-months to four years (Girard et al., 2016; Girard et al., 2014). The emergence of problems in either language or SEB development might negatively impact upon one another over time
(Girard et al., 2016). The consistency of the associations across time, they suggest, although small, might indicate that early problems might persist without intervention.
Examination of the association by type of language problem and type of SEB problem might clarify which difficulties parents and health professionals should be vigilant for.
2.6 Risk Factors for DLD and SEB Difficulties
Having described DLD and SEB difficulties and their association, this review will now explore factors which may facilitate the identification of children at risk of these difficulties and/or be used in interventions. Since both DLD and SEB difficulties represent public health concerns, extensive research efforts have been made to identify their risk factors. However, research in one area has been conducted largely independently of the other; DLD research is often disseminated within the more general
18 child development literature or the developmental disabilities literature, whilst SEB difficulties research is disseminated within the mental health literature. Integrating the knowledge gained in both literatures is likely to be beneficial for the significant proportion of children experiencing co-occurring difficulties and the health and education professionals working with them.
2.6.1 Risk Factors for DLD
There are multiple challenges to predicting children who are likely to experience
DLD. Epidemiological studies have revealed how early language development is marked by instability and fluctuations (e.g. Reilly et al., 2010). This means it is not possible to use early language scores alone to predict later language abilities at the level of the individual. For example, whilst between 10-19% of two-year-olds are ‘late talkers’ due to producing fewer than 50 words and/or no word combinations
(Desmarais, Sylvestre, Meyer, Bairati, & Rouleau, 2008; Fenson et al., 2007; Rescorla,
1989), between 62% and 80% have language skills within the normal range two to five years later (Bavin & Bretherton, 2013; Dale & Hayiou-Thomas, 2013). Furthermore, children who were not late to talk can go on to develop DLD, as highlighted by a
Norwegian study (n = 10,587) which found 6.5% of children fell into this category
(Zambrana, Pons, Eadie, & Ystrom, 2014).
Another approach to predicting DLD has been to identify risks among child, maternal and family characteristics that are known to be associated with child development (C. L. Taylor, Zubrick, & Rice, 2013). These could be used alongside measures of early language abilities to increase the predictive capability (Clegg, Law,
Rush, Peters, & Roulstone, 2015; H. D. Nelson, Nygren, Walker, & Panoscha, 2006;
19 Zubrick et al., 2007). By identifying children at risk, early, prevention or remediation approaches can be implemented. In 2006, a US Preventive Services Task Force identified 12 risk factors in a systematic review that could potentially be used for selective screening for speech and language delay in preschool. These were male gender, twin birth, preterm birth, low birth weight, minority status, low socio-economic status background, birth order, family size, childhood illness, a family history of speech and language difficulties, low parental education, and younger maternal age (H. D.
Nelson et al., 2006). A follow-up review also concluded that male gender, family history, low parental education, preterm birth, birth difficulties, and low birthweight were risk factors for speech and language difficulties (Wallace et al., 2015). Maternal mental health status has also been found to impact on language development (Noel,
Peterson, & Jesso, 2008; Stein et al., 2008).
Many of the risk factors identified by the Task Force have been considered in subsequent prospective cohort studies and have been found to operate with low to moderate and inconsistent predictive strength (e.g. Christensen, Zubrick, Lawrence,
Mitrou, & Taylor, 2014; Reilly et al., 2010; Zubrick et al., 2007). Prediction of expressive vocabulary delays at age two from predictors in pregnancy, birth and early development was poor (Zubrick et al., 2007), whilst prediction of four-year-old language skills from 10 of the 12 factors identified by H. D. Nelson et al. (2006) was moderate, explaining 21% of the variability (Reilly et al., 2010). This increased to 30% once two-year-old late talker status was considered, which although high from an epidemiological perspective, still leaves much of the variability unexplained. At older ages, the predictive strength of risk factors combined with earlier language scores appears to be low; Christensen et al. (2014) showed that of those children with low expressive vocabulary at age four who they predicted would have low expressive
20 vocabulary at age eight using a range of risk factors, only 25.8% actually did. These risk factors included many identified by the Task Force plus others including parenting consistency, child temperament, reading to child, and maternal work hours.
Herein lies the dilemma for health and education professionals in identifying children at risk of DLD. If identification occurs too early, there is a risk of capturing
‘late bloomers’ and missing children with later onset difficulties. However, if identification is left until later, crucial intervention time for those with early onset difficulties might be missed, and the language problems may become entrenched.
Screening for DLD is currently not recommended until methods of identifying children at risk, and methods for intervening are strengthened (McKean et al., 2017; Wallace et al., 2015). A recent review observed a wide range of reported sensitivity and specificity in studies of screening instruments, and found no single measure or age at which screening is optimal (Wallace et al., 2015). Furthermore, due to heterogeneity in studies examining different DLD treatment, the authors concluded it was not possible to identify specific factors associated with effective treatments.
Further work is therefore required to investigate other biological and environmental risk and protective factors for identification and treatment of DLD.
Given the close association between DLD and SEB difficulties, examining risk and protective factors for SEB difficulties and comparing them to those for DLD represents a possible avenue for investigation.
2.6.2 Risk Factors for SEB Difficulties
Unlike early language development, evidence suggests that internalising and externalising behaviours up to age five years tend to persist rather than be transient
21 (Bayer et al., 2012). Hence, children at risk of having persistent mental health symptoms may be identified in the early toddler years (Bayer et al., 2012). Risk factors for SEB difficulties include many of the factors also considered to contribute to language development. A large scale Australian cohort study (n = 500) examined the contribution made by the following factors to children’s internalising and externalising behaviour scores: male gender, birth order, child temperament, maternal age, marital status, minority status, parental education, SES, maternal mental health status, hours in childcare, family violence, substance misuse, parental expectations about child development, and harsh parenting (Bayer et al., 2012). The strongest predictors of internalising behaviours at age five years were harsh discipline, maternal stress, having no older siblings, single parenthood, and maternal substance abuse, explaining 22% of the variation. The strongest predictors for externalising behaviours were male sex, harsh discipline and maternal stress, explaining 24% of the variation (Bayer et al., 2012).
There is therefore an overlap between risk factors for both DLD and SEB difficulties.
2.6.3 Risk Factors Shared by DLD and SEB Difficulties
Drawing this together, it can be observed that boys are at increased risk for externalising behaviours (Bayer et al., 2012) and for DLD (Reilly et al., 2010; Zubrick et al., 2007). Children living in poverty are more likely to experience externalising behaviour problems (Bagner et al., 2016; Briggs-Gowan, Carter, Skuban, & Horwitz,
2001) and DLD (Fernald, Marchman, & Weisleder, 2013; Halle et al., 2009; Horwitz et al., 2003). Children of mothers experiencing stress and mental health problems have poorer language development (Noel et al., 2008; Stein et al., 2008), and more internalising and externalising problems (Bayer et al., 2012). Interestingly, for some factors the association with DLD and SEB difficulties is in different directions. For
22 example, having older siblings has been reported to be a risk factor for poorer language development (Hoff-Ginsberg, 1998), but a protective factor for internalising behaviour problems (Bayer et al., 2012).
A further factor that has been extensively investigated within the literature concerning SEB development, and to a lesser extent, language development, is parenting behaviours. This is because the parent-child relationship is considered pivotal to child health and wellbeing (O'Connor, 2002; Sameroff, 1998). Parenting behaviours have been found to mediate the associations between some of the risk factors discussed above and child outcomes. For example, there is evidence that parenting mediates the effects of poverty on externalising problems (Pachter, Auinger, Palmer, & Weitzman,
2006) and on language difficulties (Raviv, Kessenich, & Morrison, 2004).
This chapter will now examine how parenting is associated with child language and SEB development. It will then describe how this knowledge has been used to develop parent-mediated intervention approaches for children with DLD and for children with SEB difficulties.
2.7 Parenting and Child Development
The belief in the importance of the parent-child relationship to child development has its roots in social learning theory (Bandura, 1977) and social interactionism (Bohannon & Bonvillian, 1997; Bruner, 1978). These perspectives postulate that young children learn social behaviour in interactions with their caregivers/guardians/parents, ‘parents’ hereafter. In adaptive relationships, children learn that when they signal, their parents will interpret their signal and respond promptly to their need or desire. This motivates children to signal again, and reinforces
23 parental responsiveness. The result is reciprocal interactions that benefit and contribute to children’s socioemotional and language development (Kong & Carta, 2013; Landry,
Smith, Swank, & Guttentag, 2008; Mahoney & Perales, 2003). Parent-child interactions are therefore a key developmental pathway for child development (Guralnick, 2011,
2017).
The pattern of parent-child interaction is transactional (Sameroff & Chandler,
1975). Both parent and child contribute and respond to each other’s cues and developing competencies (Bornstein, Hendricks, Haynes, & Painter, 2007; Song, Spier,
& Tamis-Lemonda, 2014). Warm, sensitive, and responsive interactions that are low in psychological control but high in appropriate behaviour control are associated with optimal cognitive, socioemotional, academic and language development (Deater-
Deckard & Petrill, 2004; Landry et al., 2006; O'Connor, 2002; Pettit & Bates, 1989;
Treyvaud et al., 2009). Numerous studies have provided empirical support for this, and several are described in more detail below (e.g. Gardner, Ward, Burton, & Wilson,
2003; Hart & Risley, 1995; Landry, Miller-Loncar, Smith, & Swank, 2002; Magill-
Evans & Harrison, 2001; Masur et al., 2005; Song et al., 2014; Supplee, Shaw,
Hailstones, & Hartman, 2004).
2.7.1 Parenting Behaviours and Language Development
Parent-child interactions have been examined to identify risk and protective factors for language development (Bruner, 1986; Tomasello, 1992; Tomasello & Todd,
1983; D. C. Vigil, Hodges, & Klee, 2005). Communicatively responsive parents encourage their children’s language use by motivating them to engage in stimulating and rewarding conversations (Pungello, Iruka, Dotterer, Mills-Koonce, & Reznick,
24 2009). Responsive and assertive children provide more opportunities for parents to model language skills and provide linguistic feedback. Together, parent and child balance their turn-taking roles (Raver, 1996), adapting and evolving according to the child’s changing language ability (Conti-Ramsden & Friel-Patti, 1986; Tamis-LeMonda
& Bornstein, 1991).
Parents who are insensitive or unaware of their children’s needs or perspectives may be unable to adapt appropriately to their children’s evolving abilities. Unsupportive or intrusive behaviours might disrupt language learning by interrupting children’s attention, inhibiting their participation, or missing opportunities to extend their language. Consequently, the children may learn words at a slower rate, and initiate and respond less frequently. Their parents then have fewer opportunities to support verbal exchanges (van Balkom, Verhoeven, & van Weerdenburg, 2010). This maladaptive cycle may contribute to the children having poorer language outcomes.
Research efforts to test these hypotheses have examined variations in the quantity, quality, and diversity of parental linguistic input. Researchers have found differences in the interactions between mother-child dyads with and without children with DLD (e.g. Conti-Ramsden & Friel-Patti, 1986; Tamis-LeMonda & Bornstein,
1991). For example, mothers of toddlers with low expressive vocabulary reportedly use more initiations, prohibitions, and directives, and the toddlers initiate less frequently
(Paul & Shiffer, 1991; Rescorla & Fechnay, 1996; D. C. Vigil et al., 2005). It is not possible to determine causality, and it is possible that these differences are appropriate adaptations to children’s abilities (Hudry et al., 2013; Marfo, 1990). Children with DLD may be more challenging to engage with and perceived as less enjoyable to interact with compared to children with typical language skills (Irwin et al., 2002). These children
25 may have fewer or lower quality interactions with their caregivers, perpetuating their difficulties (La Paro, Justice, Skibbe, & Pianta, 2004).
As well as comparing interactions between groups of mother-child dyads, researchers have explored whether variations in parental linguistic behaviours are associated with children’s current or later language skills (e.g. Hart & Risley, 1995;
Rowe, 2012). In the renowned Hart and Risley (1995) study, audio recordings of parent- child interactions in the home were compared between 42 families split into three different SES groups: professional, working class and those receiving welfare. Children in welfare families heard on average 1,400 fewer words every hour than those in the professional group, were responded to less frequently by their parents, and received lower quality input in terms of different word types. Over the first three years of life, this discrepancy extrapolated to a thirty-million-word gap, correlating with poorer expressive vocabulary scores among the children in welfare families. Since expressive vocabulary skills at school entry are strongly associated with later literacy skills (e.g.
Dickinson & Porche, 2011) this finding was the catalyst for a range of public health programmes in the US aimed at increasing the number of words heard by children during their first three years.
Since Hart and Risley (1995), many other researchers have demonstrated how the quantity and quality of maternal linguistic input is important for language learning
(Cartmill et al., 2013; Hoff & Naigles, 2002; Huttenlocher, Haight, Bryk, Seltzer, &
Lyons, 1991; Landry et al., 2002; Masur et al., 2005; Rowe, 2012). Quality is usually measured by the number and diversity of word types and syntactic structures (e.g.
Huttenlocher, Waterfall, Vasilyeva, Vevea, & Hedges, 2010; Pan, Rowe, Singer, &
Snow, 2005). More nuanced investigation has revealed how specific types of quality
26 might be important to language acquisition at different points in child development. For example, Rowe (2012) examined the association between quantity and quality of parent input in 50 mother-child dyads at ages 18, 30 and 42 months and children’s expressive vocabulary skills one-year later. Whilst using diverse and rare vocabulary at 18 and 30 months explained later vocabulary ability, it wasn’t until 42 months that parental use of decontextualized language, that is, talk about objects/events outside the immediate environment, became important to the children’s vocabulary skills. This finding was pivotal because it revealed how children’s ages and language abilities are vital for them being able to access or benefit from the “more fine-grained aspects” of parental verbal input (Rowe, 2012, p. 1771).
Another approach to exploring how parenting behaviours may be associated with language development is to examine categories of communicative input. For example, responsive communicative behaviours, defined as prompt, contingent and appropriate, include following the child’s lead, recasting, and repeating the child’s last utterance (e.g. Bornstein & Tamis-LeMonda, 1989). These behaviours are believed to facilitate language learning by, for example, encouraging and reinforcing children’s involvement in interactions (Hoff & Naigles, 2002; Masur et al., 2005), and providing descriptions of objects within the child’s attention thereby maximising their cognitive capacity for language learning (Tamis-LeMonda, Bornstein, & Baumwell, 2001;
Tomasello & Farrar, 1986).
Many researchers have shown that responsive communicative behaviours are positively associated with early language development (e.g. Bornstein et al., 2008;
Girolametto, Bonafacio, et al., 2002; Levickis et al., 2014; Masur et al., 2005). Levickis et al. (2014) found expansions, imitations, and responsive questions were positively
27 associated with 24 and 36 month receptive and expressive language scores in a community-derived sample of children (n = 251). Expansions appeared to be an especially important parental strategy for enhancing language learning since their use predicted change in language scores from 24 to 36 months. Contrary to expectations, labels were associated with poorer language scores, and there was no evidence that interpretations or supportive directives were associated with language scores. Much like
Rowe’s (2012) finding that diverse vocabulary and decontextualized language use were associated with language abilities at different ages, Levickis et al. (2014) speculated that different responsive behaviours may facilitate language learning at different ages or developmental levels. For example, interpretations may be appropriate for children with severe difficulties (Girolametto, Bonifacio, et al., 2002; Girolametto, Weitzman, Wiigs,
& Pearce, 1999), whereas labelling and supportive directives may be appropriate for younger children or those with slower language acquisition (Akhtar, Dunham, &
Dunham, 1991; Masur et al., 2005).
The opposite of responsive parenting has also been investigated in relation to language development, that is, negative-intrusive parenting. This captures the degree to which parents interfere with children’s needs, interests or behaviours beyond their developmental requirements (Pungello et al., 2009). Types of negative intrusive behaviours include criticising, ignoring, prohibiting, unnecessarily restraining, and verbally controlling the child with repeated, unnecessary direction (Pungello et al.,
2009). This behaviour can undermine children’s autonomy and confidence, and is associated with poorer child outcomes across several development domains including social-cognitive development, executive functioning, and academic performance (e.g.
M. Carpenter, Nagell, & Tomasello, 1998; Culp, Hubbs-Tait, Culp, & Starost, 2000;
Landry, Smith, Swank, & Denson, 2000).
28 However, research findings are mixed regarding the association between negative-intrusive parenting and language acquisition. Some researchers found a negative association with language outcomes Keown, Woodward, and Field (2001), and others that the negative association was explained by differences in maternal sensitivity
Tamis-LeMonda, Shannon, Cabrera, and Lamb (2004). A finding by Pungello et al.
(2009) helped explain this ambiguity by revealing that negative-intrusive parenting at
12 and 24 months was associated differently with receptive and expressive language outcomes depending on the family context (i.e. race and SES).
A specific type of intrusive behaviour has received a great deal of research attention: directives (Akhtar et al., 1991; Levickis et al., 2014; Masur et al., 2005; Pine,
1992; Tamis-LeMonda, Baumwell, & Cristofaro, 2012; Tomasello & Farrar, 1986).
Originally viewed as a negative parenting behaviour associated with poorer language development (e.g. McDonald & Pien, 1982; Newport, Gleitman, & Gleitman, 1977), subsequent research revealed that not all types of directives are equivalent (Pine, 1992).
For example, Barnes, Gutfreund, Satterly, and Wells (1983) found directives were associated with gains in language skills. Others found directives that followed into the child’s focus of attention were associated with better language outcomes compared to those that led children’s attention, which were associated with poorer language (Akhtar et al., 1991; Masur et al., 2005; Tomasello & Farrar, 1986). Although Levickis et al.
(2014) went on to find no evidence that directives following children’s attention were associated with language outcomes at 24 or 36 months. “Directiveness” as a parenting style has also been investigated. N. Taylor, Donovan, Miles, and Leavitt (2009) reported that directiveness containing few prohibitions, low negative affect, and more guidance was associated with advanced language skills in toddlers. Therefore ambiguity still
29 exists around how negative and intrusive behaviours may be associated with language development.
Having reviewed the findings for how responsive and intrusive parenting behaviours may be associated with language outcomes, and concluded that the evidence is mixed, it became apparent that further investigation was required. The following section will outline how responsive and intrusive parenting behaviours may be associated with SEB development.
2.7.2 Parenting Behaviours and SEB Development
Parenting practices that have been examined in relation to optimal child development have been labelled in different ways. Positive parenting practices are thought to foster children’s socioemotional development by providing a secure base where children develop internal beliefs that they are valued, and safe, and from which they can explore their world (Cicchetti & Toth, 1998). Positive practices include using a positive tone of voice, talking with rather than to children, responding to their questions, and praising, kissing and caressing (Caldwell & Bradley, 2003). Responsive parenting is associated with child compliance (Roskam, Brassart, Loop, Mouton, & Schelstraete,
2015). ‘Warm-engaging’ practices are nurturing and receptive to children’s communications, and ‘autonomy-encouraging’ practices help children attain skills at a developmentally appropriate pace, to explore, reason, and make independent choices
(Bayer, Sanson, & Hemphill, 2006; Cicchetti & Toth, 1998).
In contrast, several parenting dimensions are associated with poorer SEB outcomes (Bayer et al., 2006; Pinquart, 2017; Shaw & Gross, 2008; Sitnick et al., 2015).
Parenting characterised by hostility and neglect can lead to parent-child interactions
30 which are distressing for children and teach them that relationships do not provide support or safety (Bayer et al., 2006; Michiels, Grietens, Onghena, & Kuppens, 2008).
Negative practices include behavioural control such as demanding obedience using yelling or hitting, and psychological control including sarcasm, criticism, and failing to acknowledge the child’s signals, as well as overtly rejecting the child (Cytryn &
McKnew, 1996; Mills & Rubin, 1998; Rubin & Mills, 1991). Over-involved/protective parenting is also detrimental to child development in different ways. Examples include trying to protect children from all difficulties in life, setting rules to prevent children from making mistakes, and describing challenging situations to children as scary or too hard (Bayer et al., 2006). It is thought to threaten the development of autonomy, self- belief and coping mechanisms and to encourage dependence (Parker, 1983). By providing children with too much support and instruction relative to their capabilities, it is a central form of early parenting control (Bayer et al., 2006).
Researchers interested in examining how parenting behaviours might be associated with SEB development have used similar approaches to those used by language researchers. Some studies have compared parenting between groups containing dyads with and without child SEB difficulties (e.g. Lougheed, Hollenstein,
Lichtwarck-Aschoff, & Granic, 2015; Sanders, Dadds, Johnston, & Cash, 1992). For example, compared to parents of typically developing children, parents of children with externalising difficulties use more coercive techniques, fewer initiations, behavioural controls, and mutual problem solving, and are less responsive (Gardner, 1994; Mills &
Rubin, 1998; G. R. Patterson, 1982; Sanders et al., 1992). This has been observed even in the absence of poor child behaviour (Gardner, 1994). Parents of children with internalising symptoms reportedly use more behavioural and psychological control,
31 including imperative commands for compliance and non-response (Mills & Rubin,
1998).
Other researchers have investigated whether parenting behaviours are associated with concurrent or later SEB difficulties (e.g. Girard et al., 2016; Pettit & Bates, 1989;
Shaw & Gross, 2008; Sitnick et al., 2015). Intrusive parenting behaviours in the early years have been shown to be associated with internalising problems (e.g. Barrett, Rapee,
Dadds, & Ryan, 1996; Mills & Rubin, 1998), and externalising behaviours (Pettit &
Bates, 1989; Shaw & Gross, 2008; Sitnick et al., 2015). A meta-analysis of over 1,400 studies revealed that harsh control and psychological control were the parenting practices most strongly associated with child externalising problem behaviours (J. M.
Vigil, 2009). Over-involved/protective parenting at two and four years was associated with more internalising difficulties at age four (Bayer et al., 2006) Recently, using the large scale Millennium Cohort Study in the UK, Girard et al. (2016) found that harsh parenting was positively associated with conduct problems at age four, whilst positive parenting was associated with fewer such problems.
It is not possible to determine causality with the observational designs used in these studies. The lack of child externalising behaviours during the interactions observed by Gardner (1994) might indicate that parents were using less responsive practices regardless of their child’s poor behaviour, perhaps reflecting the skills they had learnt through their own experiences of being parented. It is also plausible that their choice of parenting behaviours was stopping their child’s externalising behaviour before it became observable to researchers. Alternatively, the parents may have habituated to coercive parenting techniques due to the child’s history of poor behaviour.
32 Researchers have begun to conduct more nuanced and sophisticated investigations addressing subtleties in parent-child relationships that contribute to the associations between parenting behaviours and child outcomes, in a similar way to language researchers (e.g. Rowe, 2012). For example, Lougheed et al. (2015) compared interactions between mothers and children aged eight to 12 years with (n = 191) and without (n = 54) externalizing difficulties during emotionally challenging tasks.
Maternal use of supportive strategies including positive emotional directives (e.g. “You should feel proud of yourself”), validation (e.g. expressions of support, empathy or approval), and reappraisal (e.g. attempts to positively reframe an issue) were coded.
Children with externalizing difficulties were less likely to respond to their mothers’ supportiveness than children without externalizing difficulties. The authors speculated that this could be because children with externalising difficulties do not attend to their environment to the same extent as those without difficulties. So over time, parents may reduce their use of such strategies as they observe that they make no difference to their child’s mood.
Many theories that seek to explain the association between DLD and SEB difficulties focus on the role played by parent-child interaction (Carpenter & Drabick,
2011; Hartas, 2011; Hughes et al., 2016). The next section describes how aspects of parent-child interaction associated with both language and SEB development has been harnessed as a tool to improve child outcomes within the early intervention arena.
2.8 Interventions for DLD and SEB Difficulties
Early intervention is considered key to helping children with, or at risk of, difficulties in reaching their full potential (Guralnick, 2017). Many models aim to
33 modify or enhance parent-child interactions as a key mechanism for promoting child development (Lunkenheimer et al., 2008). However, the development of interventions for SEB difficulties has occurred independently of the development of language interventions. SEB interventions have explicitly targeted parenting styles, and are known as parenting programmes, parent training or parent education. Their goal is to improve warm, responsive parenting behaviours and to reduce coercive, harsh or other negative behaviours (e.g. Webster-Stratton & Reid, 2010).
Language interventions, on the other hand, are usually referred to as parent- mediated, parent-implemented, or responsive interaction interventions. Their goal has been to coach parents to use specific communicative behaviours that are responsive to their children’s utterances and will promote language learning, and to avoid disruptive behaviours that might hinder language learning. Interventions for both difficulties are often group-based, occurring weekly or fortnightly over a few months, sometimes with individual sessions as well. Integration of the SEB and language intervention literature might prove helpful for clinicians who are likely to see children experiencing co-morbid conditions, including both DLD and SEB difficulties. Furthermore, integration might lead to cross-fertilisation of ideas for intervention approaches (Brookman-Frazee,
Stahmer, Baker-Ericzen, & Tsai, 2006).
The following sections outline parenting approaches used in language interventions, and those used in SEB interventions. They draw this together by highlighting the similarities between the two, and the research that has begun to examine whether these interventions might benefit both areas of difficulty.
34 2.8.1 Interventions for DLD
Parental communicative behaviours are targeted in a range of early language interventions, including Enhanced Milieu Teaching (Kaiser, 1993), The Hanen Early
Language Parent Programme and adaptations (Girolametto, Greenberg, & Manolson,
1986; Manolson et al., 1995), the Heidelberg Parent-Based Language Programme
(Buschmann et al., 2009), and the Play and Learning Strategies programme (Wheeden
& Fewell, 1995). Proponents of these strategies believe that if parents are given the proper tools, encouragement and understanding of child development, they can become active participants in changing their child’s language learning environment (Leffel &
Suskind, 2013). The main parenting strategies targeted in the programmes include recasting or expanding upon children’s communication, increasing linguistic input quality, and balancing conversational turns (Roberts & Kaiser, 2011). The goal that is shared by these different interventions is to promote parents’ verbal responsiveness by teaching them how to react sensitively, contingently and at a level appropriate to the children’s developmental level, minimising directives and adult-directed talk (Kong &
Carta, 2013). In this way, they aim to improve children’s language development. Some interventions also aim to improve children’ emotional and cognitive development
(Girolametto, Pearce, & Weitzman, 1996; Mahoney & Perales, 2005).
Evidence suggests that the interventions are effective in teaching parents to modify their use of different communicative behaviours (Brassart & Schelstraete, 2015;
Kong & Carta, 2013; Roberts & Kaiser, 2015). A review of parent-child interaction interventions for children with or at risk of developmental delays found verbal responsiveness strategies (e.g. imitating, balancing turns, recasting/expanding, giving children opportunity to respond), and emotional responsiveness strategies (e.g.
35 acknowledging children’s signals, providing contingent responsiveness with warm and sensitive behaviours, showing positive affect) had moderate to large positive effect sizes on parenting behaviours (range 0.21 to 1.16) (Kong & Carta, 2013). A recent RCT of a short-term intensive intervention specifically for children with expressive and receptive language delays (aged 24 to 42 months) also provided support that mothers can be taught to use language facilitation strategies by using systematic caregiver instruction
(Roberts & Kaiser, 2015).
Evidence also suggests that the interventions are associated with changes to child language outcomes (Brassart & Schelstraete, 2015; Kim & Mahoney, 2005; Kong
& Carta, 2013; Roberts & Kaiser, 2011, 2015). A meta-analysis of 18 studies for children aged 18 to 60 months by Roberts and Kaiser (2011) concluded that they had a positive impact on children’s receptive and expressive language skills, although most reported small effect sizes immediately post intervention. In their later RCT, Roberts and Kaiser (2015) reported that the change in parent behaviours was associated with small to moderate changes in mainly receptive language skills immediately after the intervention. In their review, Kong and Carta (2013) identified seven parent- implemented intervention studies for toddlers with either language difficulties or who were at environmental or biological risk of developmental delays (Deutscher, Fewell, &
Gross, 2006; Dieterich, Landry, Smith, Swank, & Hebert, 2006; Girolametto, Pearce, &
Weitzman, 1997; Girolametto, Pearce, & Weitzman, 1995; Girolametto et al., 1996;
Landry et al., 2008; Moran, Pederson, & Krupka, 2005). There were mixed findings for improvements in child social communication (e.g. vocalisations, turn taking, intentional communication) (Girolametto et al., 1997; Girolametto et al., 1995, 1996; Landry et al.,
2008), and small to moderate improvements in emotional behaviours (positive and
36 negative affect) (Moran et al., 2005), and cognitive behaviours (e.g. complex play skills) (Deutscher et al., 2006; Dieterich et al., 2006).
Currently, the evidence for sustained intervention effects for changes to parent behaviour and child language outcomes are limited and weak. Many studies do not include follow-ups beyond the immediate post-intervention. It is also not clear which specific parenting behaviours are associated with changing child outcomes (Roberts &
Kaiser, 2011). Research is required to identify whether there are specific components of language interventions strategies that are key or most active in promoting language learning, and may be associated with language development over time.
2.8.2 Interventions for SEB Difficulties
Parent education approaches are a common method for intervention for SEB difficulties in the pre-school period (Furlong et al., 2012; Salmon et al., 2016). They include Parent Child Interaction Therapy (Zisser & Eyberg, 2010), Behavioural Child
Management Training (e.g. Sanders & Markie-Dadds, 1992), Parent Management
Training (Forgatch & Patterson, 2010), the Incredible Years (Webster-Stratton & Reid,
2010), Triple P (Nowak & Heinrichs, 2008), the Family Check-up (Dishion et al.,
2008), the Chicago Parent Program (Breitenstein et al., 2012), and the Solihull
Approach parenting group Understanding your Child’s Behaviour (Douglas, 2006).
These programmes have a broader remit than those for DLD described above. The SEB parent-training programmes share the aims of coaching parents about positive involvement in parent-child interactions, limit setting, effective communication, problem solving, and skill encouragement (Dougherty et al., 2015; Sitnick et al., 2015;
Webster-Stratton & Taylor, 2001). Parents are also taught to decrease negative
37 parenting behaviours such as coercion (e.g. Dishion, Patterson, & Kavanagh, 1992;
Martinez & Forgatch, 2001). For example, parents participating in the Parent-Child
Interaction Therapy are taught ‘Do-skills’: praise, reflect, imitate, describe, enjoy.
Parents taking part in the Family Check-up are offered three home visits, and are taught about positive parental involvement, positive reinforcement, interactive engagement and proactive structuring of the environment.
There are hundreds of randomised controlled trials evaluating parent-training programmes, with recent estimates as high as 200 studies for the age group two to nine years alone (Gardner & Leijten, 2017). Such interventions are considered to be the most effective treatment approach for mental health problems in the preschool period
(Dougherty et al., 2015; Salmon et al., 2016). Evidence supports their ability to change parental behaviours, improve parent-child interactions and child outcomes, at least in the short term (e.g. Barlow & Stewart-Brown, 2000; Lunkenheimer et al., 2008;
Webster-Stratton & Taylor, 2001). A review of 24 evaluations of Parent-Child
Interaction Therapy and Triple-P programmes concluded that they improved parenting by improving parental warmth and self-efficacy while decreasing hostility and parental stress, and most programmes reduced child behaviour problems from pre- to post- intervention (Thomas & Zimmer-Gembeck, 2007). Although very few studies followed up families beyond immediate post-treatment, those that did suggested positive effects up to three months post-intervention end, but conclusions could only be tentative about long term benefits.
A later review and meta-analysis evaluated the longitudinal effects of 25 preventive programs targeting externalising behaviour difficulties in children aged two to 19 years using parent coaching (Smedler, Hjern, Wiklund, Anttila, & Pettersson,
38 2015). Since the authors examined studies with follow-up periods at least six months post-intervention, their review contained some different studies from Thomas and
Zimmer-Gembeck (2007). Although their overall conclusion was that effects on externalising behaviours six to 12 months after intervention were small and inconsistent, there were a few exceptions, notably for programmes in the preschool years - Triple-P (Nowak & Heinrichs, 2008), the Incredible Years (Webster-Stratton &
Taylor, 2001), and the Family Check-up (Dishion et al., 2008; Gardner, Dishion, Shaw,
Burton, & Supplee, 2007). Triple-P reduced symptoms of externalizing problems in preschool children for at least 12 months, with small to medium effect sizes. For example, Heineichs et al. (2006) reported an effect size of -0.3 at 12 to 16 months post- intervention. Although the Incredible Years had inconsistent effect sizes on symptoms of externalizing problems six to eight months post-intervention, it did have a small effect size on behaviours 12 to 16 months later (-0.1 to -0.2) (e.g. Gross et al., 2009;
Webster-Stratton & Taylor, 2001). Finally, two large trials of the Family Check-up with preschool children showed reduced symptoms of externalizing behaviour of a medium effect size for at least 12 months post-intervention (Dishion et al., 2008; Gardner et al.,
2007). These are very promising findings which support the considerable work on the part of researchers, clinicians, and families taking part in these interventions.
In recognition of the considerable efforts involved in designing, running and participating in parent-training programmes, Kaminski, Valle, Filene, and Boyle (2008) conducted an innovative review of programmes for child behaviour problems with the aim of identifying components consistently associated with more successful programmes. They concluded that components enhancing the overall quality of mother- child interactions, e.g. positively attending to appropriate behaviour and following their lead during play, were the most active for improving child behaviour. Specific
39 intervention strategies that were best for improving parent behaviours were educating parents to use active listening (e.g. reflecting children’s utterances), avoid negative communication (e.g. sarcasm), and use emotion coaching. Those that were associated with better child behaviour were teaching time-out and the importance of parenting consistency. The authors suggest that interventions could focus on those strategies that were consistently associated with larger effects on child outcomes rather than focusing on strategies that have smaller effects.
In a similar vein, limited resources could be directed at strategies associated with co-occurring difficulties. Since DLD and SEB difficulties often co-occur, it is possible that strategies targeted for one domain could also be associated with the other.
However, the intervention models presented here have rarely focused on the co- occurrence between DLD and SEB difficulties. Instead researchers have focused on developing specific skills for either language or SEB interventions rather than developing a more integrated approach (Dionne et al., 2003). A promising avenue of research is to investigate whether there are interventions that affect both SEB and language outcomes, and which specific components, that is, parenting behaviours, are associated with both outcomes. These components could be integrated into interventions for either difficulty. In the next section this information is integrated and similarities in approaches are identified.
2.9 Similarities between DLD and SEB Interventions
In recent years there has been increasing interest in whether early interventions designed for SEB difficulties might also address language difficulties, and vice versa
(Salmon et al., 2016; Tempel et al., 2009). The reviews of current interventions by
40 Kong and Carta (2013) and by Kaminski et al. (2008) highlighted effective parenting strategies associated with improving language development and problem behaviours respectively (noting that Kong and Carta’s review included programmes for children with developmental delay). As shown in Table 1, there is overlap between these strategies. Both reviews identified the importance of parents following the child’s lead, reflecting/recasting the child’s utterances, and the role of emotion in interactions for interventions targeting difficulties in both developmental domains.
Table 1 Strategies Taught in Early Interventions for DLD and SEB Difficulties Strategies for behaviour problems Strategies for developmental delay, including DLD (Kong & Carta, 2013) (Kaminski et al., 2008) Examples of Programmes included: Programmes included: Hanen Incredible Years (Webster-Stratton, (Girolametto et al., 1996), PALS 2000), Triple P (Sanders et al., 2000) (Landry et al. 2008) Positive interaction: Increasing verbal responsiveness: • Positively attending to appropriate • Imitating behaviour • Balancing turns • Following their lead during play • Recasting/expanding • Providing linguistic mapping • Giving child opportunity to respond • Providing contingent consequences to child’s acts Following Child’s Lead Emotional interaction Emotionally responsive • Active listening, e.g. reflecting • Acknowledging child’s signals child’s utterances • Contingent responsiveness with • Avoiding negative communication warm and sensitive behaviours e.g. sarcasm • Showing positive affect • Using emotion coaching Teaching time out (effective limit setting) Parenting consistency
41
Tempel et al. (2009) undertook a similar exercise comparing approaches used in
Parent-Child Interaction Therapy for children with problem behaviours with those used by speech pathologists for children with DLD. They compared the four Parent-Child
Interaction Therapy strategies of 1) reflecting on children’s verbalisations, 2) imitating
(appropriate) child behaviours, 3) giving behavioural descriptions, and 4) giving information descriptions, to the following strategies used by speech pathologists:1) expanding, extending, recasting, build-ups and break-downs, 2) imitating child utterances, 3) using parallel talk, and 4) using self-talk, respectively. The similarities in approaches suggests that skills health professionals are currently using for one child domain could be transferred to another.
However, Kaminski et al. (2008) found that programmes for child behaviour problems which taught parents skills to foster language or literacy were only associated with small changes in parenting behaviours and were not associated with changes in child externalising behaviours. Those teaching parents to promote children’s social skills also reported significantly smaller effects on child externalising behaviour outcomes. Reese, Leyva, Sparks, and Grolnick (2010) cautioned that programmes tailored to promote language are more beneficial to parent–child language than programmes targeting other aspects of development.
Another approach is to consider adding appropriate extra modules to existing interventions to target the co-occurring child difficulty. For example, adding a psychological component in the treatment of children with language difficulties (e.g.
Toppelberg & Shapiro, 2000). Yew and O'Kearney (2013) suggested this could cover
42 strategies that help broad psychological abilities e.g. resilience, problem focused coping, and advice about responding to children to encourage prosocial behaviour.
Including a parenting behaviour that is not too dissimilar to ones already targeted to address the primary problem might be readily received and delivered by clinicians. The following section describes the few studies to date that have examined whether there is a cross-over effect of intervention studies.
A handful of studies have investigated whether parent-mediated intervention programmes targeting one difficulty might also address the related difficulty. Most have examined whether programmes for behaviour problems might also affect language development (Bagner et al., 2016; Garcia, Bagner, Pruden, & Nichols-Lopez, 2015;
Gridley, Hutchings, & Baker-Henningham, 2015; Lunkenheimer et al., 2008), although a few have investigated the effects of language interventions on behaviour problems
(Brassart & Schelstraete, 2015; Brassart, Schelstraete, & Roskam, 2017; Delaney &
Kaiser, 2001; Hancock, Kaiser, & Delaney, 2002; Wake et al., 2011).
Bagner et al. (2016) examined the effect of the Parent-Child Interaction
Therapy for behaviour problems (Eyberg, 2005) on infants’ expressive language production. Children scoring above the 75th percentile on an externalising behaviours screen were recruited into the study (n = 60, mean age 13.5 months). Reductions in externalising behaviours from pre- to post-intervention were found to mediate small improvements in expressive language scores three and six months post-intervention end.
Garcia et al. (2015) reported similar evidence from their clinical sample of 46 older children (mean age 45-months) who had externalising problems and were at risk of developmental delay. Lunkenheimer et al. (2008) reported evidence of modest (effect sizes of 0.02 to 0.03) ‘collateral benefits’ on children’s later language and self-
43 regulation skills from the Family Check-up. This was delivered at age two years and language and self-regulation skills were not targeted. The authors suggest that improvements in parenting may have resulted in more parent-child conversation and play, indirectly fostering these skills (Lunkenheimer et al., 2008). Taken together, these studies present some support for interventions for behaviour problems having very modest associations with language development.
Gridley et al. (2015) took a different approach and examined whether the
Incredible Years Parent-Toddler Programme enhanced parental responsive communication because it included sessions on how to verbally scaffold children’s language and emotional development. There was a positive trend for parents in the intervention group to use more encouraging language and praise than control parents used. However, there were no other observable changes in the quantity or quality of parental verbal input. Somewhat inexplicably, there was an increase in parent passivity among dyads in the intervention group compared to the control group. This is a disconcerting finding which requires investigation as it suggests that taking part in the intervention may have affected parents’ confidence in interacting with their children.
The researchers did not report whether there were changes in the children’s language skills so it is unknown whether this programme for SEB problems was associated with language outcomes.
However, a new version of the Incredible Years Programme is currently being trialled for children with autism spectrum disorder and language delay (Williams,
Hastings, Charles, Evans, & Hutchings, 2017). It employs the core Incredible Years components plus it targets communication, and social and adaptive skills. A pilot study comprising nine families (Hutchings, Pearson-Blunt, & Pasteur, 2016) reported
44 improvements in children’s emotional symptoms and peer problems, but no language outcomes were reported. The inclusion criteria described in Williams et al. (2017) is for children with a diagnosis of autism spectrum disorders only. Nevertheless, it will be interesting to note whether the programme results in changes to these children’s language skills.
Other researchers have examined whether language interventions might influence SEB outcomes. Wake et al. (2011) trialled a modified version of Hanen’s You
Make the Difference in a RCT of 301 children identified with low expressive vocabulary scores at 18-months. As well as finding no evidence for a difference in mean language scores between the intervention and control group, there was also no evidence of differences in parent-reported behaviour difficulties between the groups. In comparison, a much smaller study (n =16) by Brassart and Schelstraete (2015) provided preliminary support that a verbal responsive intervention could improve both parental responsiveness and child SEB difficulties in a sample of children at risk of externalising behaviour problems. A subsequent study of children with clinical levels of behaviour problems (rather than simply those at risk) (n = 21, mean age 49 months) also found improvements in parental responsiveness, parental self-efficacy (i.e. confidence in own parenting abilities), and in parent-reported child SEB difficulties. However, because researchers did not observe any changes in the children’s behaviour in their observational paradigm, they were cautious about drawing conclusions as to the effectiveness of the language intervention in addressing SEB difficulties. They proposed that whilst verbal responsive interventions might be sufficient for some children e.g. those with subclinical problems, they might be insufficient for those with clinical difficulties.
45 Few studies have attempted to target both SEB and language difficulties in one intervention for pre-school children. Law, Plunkett, and Stringer (2012) identified 19 studies of children aged five to 12 years in their systematic review of interventions for co-occurring DLD and SEB difficulties. All 19 studies reported improvements in children’s language and behaviour, but Law et al. (2012) graded the majority as of relatively low quality. In the pre-school years, studies are fewer. Delaney and Kaiser
(2001) trialled an intensive intervention (25-35 sessions) teaching parents to support their children’s communication and non-compliant behaviour (n = 4). Parental communicative responsiveness improved, as did their responses to child compliance and non-compliance. Children’s non-compliance also decreased. Hancock et al. (2002) replicated this intervention with a further five children with language delays and emergent behaviour problems. Again, parents were found to improve their responsiveness, and their reactions to non-compliance, and child language and behaviour improved. However, although parent behaviour changes were sustained six months post-intervention, the persistence of child language and behaviour changes was variable. Finally, Fanning (2007) developed the Success in Parenting Pre-schoolers
(SIP2) programme that was designed to coach low-income parents to use strategies to facilitate both language and prosocial behaviour. Parents in the intervention group (n
=14) reported increases in self-efficacy, and were observed to use more positive interaction and language facilitation techniques than control parents (n = 14). Children in the intervention arm also demonstrated significant improvements in receptive language scores, and in problem behaviours.
46 2.9.1 Summary of Similarities in Intervention Studies
There is evidence that parent-mediated interventions for DLD and SEB difficulties can result in changes to parent behaviours, and to child outcomes (Kaminski et al., 2008; Kong & Carta, 2013). The comparison of strategies used in DLD and SEB interventions has revealed similarities in approaches (e.g. Tempel et al., 2009). There is some preliminary evidence that interventions for one difficulty might be associated with improvements in the other (e.g. Lunkenheimer et al., 2008), although such studies are scarce (Law, Rush, & McBean, 2014). A small number of studies also suggests that interventions that target both difficulties may be effective in improving child language and SEB outcomes (Delaney & Kaiser, 2001; Fanning, 2007; Hancock et al., 2002) although their exploratory nature and small samples mean further investigation is necessary before drawing definitive conclusions. Currently it is not possible to determine which parenting behaviours targeted in the studies outlined above are the interventions’ active components (Kaminski et al., 2008; Roberts & Kaiser, 2011).
Many studies have not reported parental use of the targeted behaviours before and after the intervention, in addition to pre and post-intervention child language and SEB scores.
Hence it is not possible to conclude whether changes to child outcomes are due to changes in parental behaviours. Without a clearer understanding about precise parenting behaviours that might be associated with both language and SEB development, it will be difficult to progress efforts to offer effective interventions to children at risk of both
DLD and SEB difficulties. Going back to the drawing board to explore parenting behaviours used with young children during free-play, and their association with child language and SEB skills represents a feasible way forward. This approach might identify parenting behaviours that are key factors for child development in both language and SEB domains.
47 2.10 Rationale for PhD Study
Evidence from both research and clinical practise shows that language and SEB difficulties often co-occur in childhood. However, it is unclear when the association between language and SEB difficulties emerges, and which language and SEB domains are associated. Parenting behaviours are targeted in speech and language therapy and in psychotherapy interventions for preschool children with DLD and SEB difficulties respectively (Falkus et al., 2016; Salmon et al., 2016). However, to date, there has been minimal research into whether similar parenting behaviours may be associated with both language and SEB difficulties, despite the acknowledged association between the two. There is an absence of research investigating whether parent behaviours observed during naturalistic interactions between mothers and their children are associated with the child’s concurrent or later language and SEB skills. This has important research and clinical implications for identifying children at risk, and intervening before problems arise or become entrenched. The potential of early intervention for DLD and SEB difficulties requires a clear understanding of the age at which problems emerge, which difficulties co-occur, and whether there are specific parental behaviours that could be targeted in intervention to address both difficulties.
2.11 Summary of Literature Review
This Chapter has outlined the literature regarding the co-occurrence of DLD and
SEB difficulties in early childhood, and the current situation regarding parent-mediated interventions for both difficulties. The overall rationale for the thesis has been explained. The following Chapter will introduce the Maternal Responsive Behaviours
Study (MRBS, Levickis et al., 2014) within which this PhD is embedded. It will
48 introduce the parenting behaviours pertinent to the current study, before outlining the aims of the PhD.
49 3 Maternal and Interactive Behaviours
3.1 Overview
This Chapter will first outline the maternal responsive communicative behaviours examined as predictors of child language skills at 24 and 36 months in the
Maternal Responsive Behaviours Study (MRBS, Levickis et al., 2014). Second, the
Chapter will describe the process of identifying other parenting behaviours of interest that are hypothesised to be associated with child language and/or SEB difficulties.
Third, as a compliment to the manuscripts that comprise Chapters 6 and 7 of this thesis, a review of the literature regarding these parenting behaviours is provided. Finally, the aims of the PhD will be presented.
50 3.2 Maternal Responsive Behaviours Study
The MRBS was embedded within an intervention study for children identified as being slow-to-talk at 18 months old, Let’s Learn Language (LLL, Wake et al., 2011).
The MRBS examined whether six maternal responsive communicative behaviours often targeted in speech and language therapy were associated with concurrent and later language scores when observed during parent-child free-play at 24 months old Levickis et al. (2014). The procedures for the MRBS are fully described in the Methods Chapter
(section 4.2.2). Levickis et al. (2014) found evidence that greater use of expansions, imitations and responsive questions predicted higher receptive and expressive language scores at 24 and 36 months, after adjusting for biological and environmental confounders. However, labels were negatively associated with language scores at 24 and 36 months, and no evidence of an association was found for interpretations or supportive directives.
The aims of the MRBS were focused solely on specific maternal responsive behaviours hypothesised to be associated with language outcomes. However, proponents of the transactional model of parenting recognise that maternal behaviours are influenced by child effects (Bell & Harper, 1977), and suggest that mother and child therefore cannot be viewed as entirely independent of each other (Funamoto & Rinaldi,
2015; Sameroff, 2009). Taking this perspective, it is possible that by measuring child or interactional behaviours, important aspects of the mother-child relationship might be captured that would otherwise be missed by focusing solely on maternal behaviours
(Aksan, Kochanska, & Ortmann, 2006; Funamoto & Rinaldi, 2015). A more complete investigation would include examination of child and interactive behaviours in addition to the responsive maternal behaviours. This would also allow for examination of how
51 maternal and interactive behaviours in combination are associated with child outcomes
(Kaminski et al., 2008).
Furthermore, the MRBS focused solely on child language outcomes. Given the evidence for the association between child language and SEB difficulties, and the evidence for parenting behaviours being associated with both developmental areas, it is feasible that child, interactive and maternal behaviours might be associated with both language and SEB outcomes. This rationale is particularly strong for the maternal responsive behaviours coded in the MRBS, since it is widely acknowledged that parental responsiveness is positively associated with general child development (e.g. O'Connor,
2002).
This following section will describe the process of selecting which child, interactive and other maternal behaviours might be associated with language and SEB outcomes. The selection was restricted to behaviours that could be explored within the framework of the MRBS.
3.3 Identifying Behaviours: Step 1
The first step was a generative literature review examining the numerous existing taxonomies for coding maternal and child behaviours during interactions.
Preference was given to coding schemes designed to record counts of behaviour rather than ratings, in line with the methodology used by Levickis et al. (2014). As shown in
Table 2, the schemes considered included counts of different types of maternal speech acts, such as directives and praise (e.g. Bornstein et al., 2008; Whitehurst et al., 1988), those for coding initiations and responses in interactions (e.g. Andersen & Marinac,
2007; Rice et al., 1990), and those rating aspects of the interaction itself such as overall
52 sensitivity (e.g. Crawley & Spiker, 1983). The purpose of this first step was to gain an understanding of the range of behaviours that could be investigated for their association with language and SEB development. Because this PhD study utilised the data collected in the MRBS, it was necessary to develop a set of inclusion and exclusion criteria when considering the maternal, child and interactive behaviours of interest. The behaviours had to be used relatively frequently by mothers of children aged 18- to 48-months in a free-play scenario. The behaviours had to be coded in real-time (i.e. no transcription) which meant they had to be easily observable, require minimal decision-making from the coder, and be independent from each other. Transcribing the interactions was deemed unmanageable and outside the scope of a PhD study that already involved coding almost 200 videos twice. Further, the MRBS had already established that this method of coding behaviours was appropriate for coding a large number of parent-child interactions in child language research.
In Table 2 the behaviours of interest and coding schemes are summarised. Those behaviours which are the same or similar to those coded in the MRBS, and those that were not coded in the MRBS, are highlighted in separate columns. The final column contains a high-level commentary about the behaviours that were and were not considered for inclusion in this PhD study.
53 Table 2 Existing Communicative Behaviour Coding Schemes Consulted Researchers / Codes Coded in MRBS Not coded in MRBS Comment Scheme Whitehurst et al. • Directives • Directives: Supportive • Conversation Consider: (1988) directives • Labelling* • Yes/No Questions • Conversation/Answer: • Labelling: Labels as part of initiations • Conversation • Imitative directives and responses? • Wh- questions: • Yes/No Questions • Other Responsive questions • Directives: intrusive • Wh- questions* • Praise/confirmation • Repetition: Imitations • Imitative directives: • Imitative directives • Correction intrusive • Expansion: Expansions • Other • Compliance • Praise / confirmation
• Praise/confirmation • Criticism • Compliance • Correction • Answer Do not consider: • Repetition* • Yes/No question: difficult to code when • Compliance responsive qus have • Expansion* already been coded. Coder must think • Criticism quickly about type of • Answer question being asked. • Correction/Criticisms: unlikely to happen frequently in free-play scenario
54 Researchers / Codes Coded in MRBS Not coded in MRBS Comment Scheme Winsler, Diaz, Regulation/control • Implied directives: • Commands Consider: McCarthy, overlap with supportive • Commands • Negative corrections • Commands: intrusive Atencio, and directives Chabay (1999) • Implied directives • Conceptual questions • Implied commands • Directive questions: • Directive questions overlap with responsive • Broadcasting • Praise: consider questions combining into one • Instructions • Praise of the task category • Instructions: overlap with • Negative corrections • Praise of child’s supportive directives • Broadcasting, Self- character Teaching: evaluations, Maternal • Perceptual questions: • Praise of child’s plans: code as part of • Perceptual questions overlap with responsive actions maternal initiations and • Conceptual questions questions responses? • Unqualified praise • • Labelling Labelling: Labels Do not consider: • Self-evaluations
• Broadcasting • Negative corrections: • Maternal plans Praise: unlikely to happen frequently in free-play • Praise of the task • Conceptual questions: • Praise of character difficulty coding questions, as above. • Praise of actions • Unqualified praise Modelling: • Self-evaluations • Maternal plans
55 Researchers / Codes Coded in MRBS Not coded in MRBS Comment Scheme Ninio, Snow, Pan, A comprehensive • Directives and responses: • Speech elicitations Required transcription. and Rollins (1994) taxonomy of speech acts: overlap with supportive and responses Very detailed directives Inventory of • Directives and • Commitments and Consider: Communicative responses • Demands for clarification: responses • Directives and Acts overlap with supportive • Speech elicitations and • Declarations and responses: consider directives responses responses intrusive directives & • Questions and responses: child compliance • Commitments and • Markings and overlap with responsive responses responses • Speech elicitations and questions responses: consider as • • Declarations and Statements and part of intrusive responses responses directive’ • Markings and • Performance • Performance responses evaluations evaluations: consider as • Statements and • Demands for part of praise responses clarification • Vocalisations: consider • Questions and • Vocalisations as part of child responses initiations & responses
• Performance Do not consider: evaluations • Commitments, • Demands for Declarations, Markings, clarification Statements: requires transcription to • Vocalisations distinguish between these categories.
56 Researchers / Codes Coded in MRBS Not coded in MRBS Comment Scheme Mundy et al. Joint attention – lower and All behaviours in MRBS Higher and lower levels Requires structured (2003) higher levels occurred within joint of joint attention. assessment for 20 mins, and Early Social attention. That is, episodes of extensive coder training. Communication joint attention within the ten- It was therefore not Scales (ESCS) minute recordings were possible to use. identified, and codes applied only during these episodes. Consider: other elements of joint attention that can be coded in real time. Crawley and Ratings of maternal, • Directiveness: elements • Maternal sensitivity Sessions must be viewed Spiker (1983) dyadic & child behaviours. captured with supportive twice so decided not to use. • Child social directives Mother-Child Maternal interaction: responsivity Consider: Rating Scales • Elaborativeness: elements • Directiveness • Child interest • directive behaviours captured with expansions, • Sensitivity interpretations • Child positive affect • child communicativeness • Elaborativeness Child ratings included: • Social responsivity (degree positively responds to initiations) • Interest • Positive affect
57 Researchers / Codes Coded in MRBS Not coded in MRBS Comment Scheme Bornstein et al. • Affirmations • Imitations/expansions • Affirmations Consider: (2008) Imitations and • Imitations/expansions • Descriptions of an Affirmations: consider as a Expansions object, event or type of praise? • Descriptions of an • Questions: Responsive activity object, event or Descriptions: as part of questions activity, initiations/responses • Play prompts: • Questions Supportive directives • Play prompts • Exploratory prompts: • Exploratory prompts Supportive directives Mahoney, Powell, Maternal interactive style: • Directiveness: supportive • Overall ratings Needs extensive coder and Finger (1986) directives training so not suitable. • Directiveness • Affect/animation Maternal • Responsiveness: specific Consider specific aspects of • Responsiveness* Behaviour Rating responsive behaviours directiveness. Scale • Affect/animation counted rather than ratings
Dunst, Trivette, Rating scales Rating scales were dis- and Cross (1986) Parent-child play scale, regarded in this first step. Caregiver Styles of Interaction Scale
58 Researchers / Codes Coded in MRBS Not coded in MRBS Comment Scheme Andersen and Adult utterances directed None of these behaviours Adult utterances directed Required many decisions Marinac (2007) to the child coded as: were coded in MRBS to the child coded as: and judgements that would be difficult to make in real Observational • Response required • Response required framework for time. • • Response not coding interactions Response not required Consider: Child linguistic response required • Child compliance: to to those utterances: Child linguistic response directives to those utterances: • Compliance • Extrapolating the • Compliance • Valid coding for adult • Valid utterances, consider • Acknowledgement coding whether mother • Acknowledgement • Inappropriate does or does not • Inappropriate respond to her child’s • Ambiguous utterance • Ambiguous • No response • No response Farran and Kasari Rating scales Rating scales were dis- (1985) regarded in this first step. Parent-child interaction scale
59 Researchers / Codes Coded in MRBS Not coded in MRBS Comment Scheme Drummond, Paul, • Simple affect talk None of these behaviours Required transcription to Waugh, were coded in the MRBS code the six different types • Hammond, and Desire talk of talk. • Brownell (2014) Emotion explanations / Consider using one code to Emotion and elaborations capture all type of talk. mental state talk • Other internal state talk • Mental state talk • Empathy statements Rice et al. (1990) Example of codes: • Responsiveness: specific • Addressee To be used in real time but responsive behaviours using a 5-mins on, 5-mins Social Interactive • Addressee • Verbal interactive were coded off methodology. Very Coding System status (assertiveness • Verbal interactive involved and required much (SICS) • status (assertiveness Repeat: imitations and responsiveness) training for reliability. Captures pattern and responsiveness) • Initiation Consider: of initiations and • Initiation • responses Verbal response • Ignore: consider as • Repeat • One-word response poor quality input ‘ • Verbal response • Nonverbal response • Initiations, Verbal response and One-word • One-word response • Ignore response: consider for • Nonverbal response generating measure of assertiveness & • Ignore responsiveness.
60 Researchers / Codes Coded in MRBS Not coded in MRBS Comment Scheme Robinson and Parent behaviour: • Direct command and • Labelled praise Consider: Eyberg (1981) Indirect command: • Direct command • Unlabelled praise • Direct command: Dyadic Parent overlap with supportive intrusive • • Child Interaction Indirect command directives Positive physical • Indirect command Coding Scheme • Labelled praise • Negative physical (DPICS) • Praise: consider coding • Unlabelled praise • Descriptive question: • Descriptive statement in one category overlap with responsive • Positive physical • Acknowledgement question • Irrelevant verbalisation: • Negative physical • Irrelevant consider as low-quality verbalisation input • Descriptive question • Critical statement • Child Comply: consider • Descriptive statement with directives • Whine • Acknowledgement • Non-comply: consider • Yell • Irrelevant verbalisation with directives • Destructive • Critical statement • Child behaviour: Non-comply Do not consider: • • Whine Comply Child while, yell, destructive: unlikely to be • Yell observed frequently • Destructive • Non-comply • Comply
61 Researchers / Codes Coded in MRBS Not coded in MRBS Comment Scheme Bakeman and • Unengaged All behaviours coded in the • Unengaged Consider: Adamson (1984) MRBS occurred during • Onlooking • Onlooking • Unengaged Joint engagement episodes of joint attention. • • • coding Person/Object focused That is, episodes of joint Person/Object Onlooking attention within the 10- focused • Supported Joint • Person/Object focused minute recordings were Engagement • Supported Joint identified, and codes applied • Supported Joint Engagement • Coordinated Joint only during these episodes. Engagement Engagement • Coordinated Joint • Coordinated Joint Engagement • Symbol-infused Engagement • Symbol-infused Do not consider Symbol- infused: requires coder to watch the recordings twice to distinguish between joint engagement that is and is not symbol infused.
62
3.4 Identifying Behaviours: Step 2
After reviewing the existing coding schemes, behaviours were ruled out if they closely aligned with responsive communicative behaviours already coded; if they required detailed training to be used (e.g. Mundy et al.’s Early Social Communication
Scales); if they required transcription or more than one pass through the video recordings (e.g. Crawley and Spiker’s Mother-child Rating Scales); if they required coding judgements that would be difficult to make in real time (e.g. Andersen and
Marinac’s Observational Coding Scheme); or if they were unlikely to be observed during the free-play activity (e.g. the negative behaviours in Robinson and Eyberg’s
Dyadic Parent-Child Interaction Coding Scheme). The following behaviours were chosen for further investigation (relevant researchers and codes if applicable are listed in parentheses):
• Directives (‘play and exploratory prompts’ from Bornstein et al. 2008;
‘directive behaviours’ from Crawley & Spiker (1983); ‘directives and
responses’ from Ninio et al. (1994); ‘direct command and indirect command’
from Robinson and Eyberg (1981); ‘commands and implied commands’
from Winsler et al. (1999); ‘directives and imitative directives’ from
Whitehurst et al. (1988)).
• Compliance (Andersen & Marinac, 2007; Robinson & Eyberg, 1981;
Whitehurst et al., 1988)
• Praise (‘affirmations’ from Bornstein et al. (2008); ‘praise’ from Robinson
and Eyberg (1981); ‘performance evaluation’ from Ninio et al. (1994);
63
‘praise for task, child’s character, child’s actions, unqualified praise’ from
Winsler et al. (1999); ‘praise/confirmation’ from Whitehurst et al. (1988)).
• Emotion and mental state talk (Drummond et al. 2014).
• Joint engagement states (Bakeman and Adamson, 1984; Mundy et al. 2003).
• Maternal and child initiations and responses (Crawley & Spiker, 1983; Ninio
et al. 1994; Rice et al. 1990; Whitehurst et al. 1988; Winsler et al. 1999).
• Fillers and missed opportunities (‘irrelevant verbalisation’ from Robinson
and Eyberg (1981); ‘ignore’ from Rice et al. (1990); ‘response required’
from Andersen and Marinac (2007)).
More detailed literature searches of these behaviours were then undertaken using
Medline, psycINFO, Pubmed, and Web of Science. This was to refine the hypotheses, and create the definitions to pilot. Search terms included ‘parent child relation’, ‘mother child interaction’, ‘parent behaviour’, ‘reciprocity’, ‘intrusive’, ‘language development’,
‘language disorder’, ‘behaviour problems’, ‘externalising behaviour’, and ‘internalising behaviour’. At this point, the scope was broadened to include children with neuro- developmental disorders, including Autism Spectrum Disorder and Down’s Syndrome.
This was because researchers often compare mother-child interactions between dyads with and without a child with difficulties to identify parenting behaviours that might explain differences in child abilities.
The two-step process resulted in the development of two coding schemes. The piloting process and final schemes are described in detail in Chapter Four (sections 4.5 and 4.6). Difficulties encountered with coding initiations and responses during piloting resulted in three additional behaviours being considered: missed opportunities to
64
respond to the child, conversational ‘fillers’, and a measure of interaction quality called
Fluency and Connectedness (FC).
The relevant literature for each maternal and interactive behaviour is described below. Further detailed descriptions are contained in the papers that comprise Chapters
Six and Seven. In Chapter Six, the paper entitled ‘Maternal communicative behaviours and interaction quality as predictors of language development: Findings from a community-based study of slow-to-talk children’ included analysis of praise, missed opportunities, successful directives, unsuccessful directives, FC, and expansions, imitations, and responsive questions from the MRBS. In Chapter Seven, the paper entitled ‘The role of joint engagement in the development of language in a community- derived sample of slow-to-talk children’ included analysis of Supported Joint
Engagement (S-JE), Coordinated Joint Engagement (C-JE), expansions, imitations and responsive questions.
3.5 Praise
Praise is defined as verbal or non-verbal positive evaluations of a child’s efforts, global characteristics, and generic positive evaluations (Gaertner, Spinrad, & Eisenberg,
2008; Winsler et al., 1999). Many of the existing taxonomies of parental speech acts reviewed in section 3.3 contain praise (e.g. Whitehurst et al., 1988; Winsler et al.,
1999). Robinson and Eyberg (1981) distinguished between labelled and unlabelled praise, as according to the principles of social learning theory (Bandura, 1977). Labelled praise provides a greater learning opportunity to the child by clearly linking the verbal praise to the activity or characteristic that is being positively evaluated (e.g. Robinson &
Eyberg, 1981; Swenson et al., 2016).
65
More recently, researchers have distinguished between person and process praise
(e.g. Dweck, 2006; Gunderson et al., 2013). Person praise (e.g. “you’re so clever”) is thought to imply that children have fixed attributes. Pre-schoolers who believed in fixed attributes were reportedly less likely to resolve disagreements using positive strategies, and more likely to support using aggression (Giles & Heyman, 2003). In contrast, process praise (e.g. “you must have tried hard”) emphasises efforts, strategies or actions.
Dweck (2006) proposed that this fosters a “growth mindset”, a belief that success is due to effort or hard work that is associated with motivation, cognition and behaviour (e.g.
Blackwell, Trzesniewski, & Dweck, 2007; Dweck, 2006; Mueller, Rowe, &
Zuckerman, 2017). Children’s exposure to process praise within their homes at ages 14 to 38 months was associated with a growth mindset five years later (n = 53) (Gunderson et al., 2013).
Parental use of praise has been linked to many developmental domains in small- scale, experimental studies, including social competence (Brophy & Dunn, 2002; Grady
& Karraker, 2014; Shinohara et al., 2010), persistence and motivation (Henderlong &
Lepper, 2002), beliefs and behaviours (Cimpian, Arce, Markman, & Dweck, 2007;
Corpus & Lepper, 2007; Zentall & Morris, 2010) and attention (Gaertner et al., 2008).
This may be because praise reinforces desirable behaviours, or helps children learn that other people have beliefs and values (for a review see Henderlong & Lepper, 2002).
Praise may also contribute to a sensitive environment where children feel supported and encouraged to participate in high quality interactions where they can practise their language skills (Bigelow et al., 2010; Nozadi et al., 2013). However, to date, the association between praise and language development has not been examined
(Henderlong & Lepper, 2002).
66
Nonetheless, parents are taught to use praise in many interventions for DLD and
SEB difficulties, e.g. the Parent-Child Interaction Therapy (Nixon, Sweeney, Erickson,
& Touyz, 2003), the Chicago Parent Program (Gross et al., 2009) and in speech and language parent-child interaction therapy (Falkus et al., 2016). Recently, Bagner et al.
(2016) and Garcia et al. (2015) found the Parent-Child Interaction Therapy had indirect effects on expressive language skills. Although it is plausible that an increase in praise was associated with these effects, parental use of behaviours was not reported.
Breitenstein et al. (2012) did report on praise however, among parents taking part in one of two prevention trials of the Chicago Parenting Program for pre-schoolers: (n = 291
African American families in one study, and n = 213 Latino families in the other study).
Whilst a short-term increase in praise was observed pre- to immediately post- intervention, this was not maintained six or 12-months post-intervention. This was despite improvements in the children’s externalising behaviours over this time, suggesting parents were not using praise to reinforce their children’s good behaviour.
In summary, parental use of praise has been linked to a range of developmental areas, and is taught in interventions for DLD and SEB difficulties. However, it is unclear whether naturally occurring praise is associated with language development and with early SEB development, and this warrants further investigation.
3.6 Intrusive Directives
Parental use of directives was included in six of the existing taxonomies examined (Bornstein et al., 2008; Crawley & Spiker, 1983; Ninio et al., 1994; Robinson
& Eyberg, 1981; Whitehurst et al., 1988; Winsler et al., 1999). The generative review identified three types of directives that had been examined in mother-child interaction.
67
Firstly, whether the directive was direct or indirect (Rescorla, Bascome, Lampard, &
Feeny, 2001; Rescorla & Fechnay, 1996); secondly, whether the directive followed or led the child’s focus of attention (Akhtar et al., 1991); and thirdly, whether the directive commanded changes to the child’s behaviour (behavioural directives) or attention
(attentional directives) (Masur et al., 2005; Masur, Flynn, & Lloyd, 2013; Pine, 1992;
Tamis-LeMonda et al., 2012). These three types of directives will be discussed below in turn.
The distinction between direct and indirect commands has only been examined within small groups of mother-child dyads with and without children with DLD or SEB difficulties (Rescorla et al., 2001; Rescorla & Fechnay, 1996; Winsler, 2009). Their use has yet to be tested in relation to language or SEB skills across the distribution of abilities. It is possible that indirect directives may be associated with poorer language skills than direct ones since children with low language abilities may have more difficulty understanding the command. However, indirect commands could also be interpreted as offering more choice to the child, and therefore be associated with better
SEB outcomes (Hebert, Swank, Smith, & Landry, 2004).
The second type of directive considered was supportive versus intrusive directives. Supportive directives follow the child’s focus of interest (Levickis et al.,
2014). Akhtar et al. (1991) suggested that asking children to do something on which they are already focused fosters their understanding of words, enabling them to respond appropriately, and promoting further interaction. However, Akhtar’s positive finding was not replicated in the MRBS, perhaps due to the older age of the MRBS sample
(Levickis et al., 2014). Intrusive directives on the other hand, lead the child’s attention, and are thought to deplete their attention and cognitive systems, leaving fewer resources 68
for language learning (Akhtar et al., 1991; Pine, 1992; Rescorla & Fechnay, 1996;
Tomasello & Farrar, 1986). This might be particularly problematic for slow language learners (Vuksanovic & Bjekic, 2013). However, the evidence is inconclusive. Masur et al. (2005) found that intrusive directives were negatively associated with expressive vocabulary abilities between the ages 11 to 21 months (n = 20). However, M. Carpenter et al. (1998) found no consistent negative association using a broader definition of intrusive utterances and language between the ages nine to 15 months (n = 25). The literature would benefit from an investigation of the association between intrusive directives and language outcomes in a larger scale, population-derived sample.
The third type of directive identified by this review was between directives that command a change to the child’s behaviour (behavioural directives) and those that command a change to attention (attentional directives) (Masur et al., 2005; Masur et al.,
2013; Pine, 1992; Tamis-LeMonda et al., 2012). Pine (1992) suggested these two directives should be treated as separate entities because they were not significantly correlated with each other (r = -0.10). Further, only attentional directives were associated with poorer maternal responsiveness (r = -0.79), which Pine et al (1992) suggested indicated maternal insensitivity to the child’s interests. Akhtar et al. (1991) also found intrusive attentional directives were negatively correlated with expressive vocabulary (r = -0.65), whilst there was no evidence that intrusive behavioural directives were correlated with language outcomes (r = -0.04).
To the author’s knowledge, the only type of directive investigated within a large-scale community-derived sample is supportive directives (Levickis et al., 2014).
Even though interventions advise parents against using an intrusive directive style (see sections 2.8), evidence for their association with either language or SEB outcomes is 69
limited to small scale, non-representative samples. Further investigation would help quantify the role directives play in these developmental domains, and the importance of targeting them in interventions. The piloting and final definition of directives used in this study is described in detail in Chapter Four, section 4.5. Further information about directives can also be found in the papers presented in Chapters Six and Seven.
3.7 Emotion and Mental State Talk
Emotion and mental state talk (EMST) is the use of words that describe, label or explain emotions, desire, mental states or internal states (Drummond et al., 2014).
Studies investigating the association between EMST and child development have been conducted across the early years, with babies (e.g. Meins, Fernyhough, Fradley, &
Tuckey, 2001; Spinrad & Stifter, 2006), toddlers aged between 18 and 36 months (e.g.
Brownell, Svetlova, Anderson, Nichols, & Drummond, 2013; Drummond et al., 2014;
Laible, 2004) and pre-schoolers aged between three and five years (e.g. Dunn, Brown,
& Beardsall, 1991; Martin & Green, 2005; Ruffman, Slade, Devitt, & Crowe, 2006).
These studies have revealed that maternal use of EMST is associated with children’s developing social competence in early childhood, which encompasses SEB abilities
(Carpendale & Lewis, 2004, 2006; de Rosnay & Hughes, 2006; Ruffman et al., 2006).
Labelling and explaining emotions (Brownell et al., 2013; Denham, Zoller, &
Couchoud, 1994; Drummond et al., 2014; Garner, Jones, Gaddy, & Rennie, 1997), and using cognitive-state and other internal-state words (Laible, 2004; Meins et al., 2001;
Ruffman, Slade, & Crowe, 2002) are positively associated with developing prosocial behaviour, empathy, and socio-cognitive understanding in toddlers and preschool children. Eliciting EMST (e.g. “How is dolly feeling?”) versus producing EMST (e.g.
“the dolly is happy”) is thought to make different demands on children’s understanding 70
(Martin & Green, 2005; Ninio, 1980, 1983), and to have differential effects on child outcomes (Drummond et al., 2014).
Maternal use of EMST is thought to socialise children to the world of emotions, by giving them the knowledge and vocabulary to understand, describe and explain their own internal states (Carpendale & Lewis, 2006; Eisenberg, Cumberland, & Spinrad,
1998). Similarly, it is hypothesised to help young children understand and interpret other people’s behaviour, feelings and beliefs (Dunn et al., 1991; Jenkins, Turrell,
Kogushi, Lollis, & Ross, 2003; Moore, Furrow, Chiasson, & Patriquin, 1994). This knowledge enables children to manage and regulate their emotions, behaviour, and mental states (Astington & Baird, 2005; Carpendale & Lewis, 2006; Eisenberg et al.,
2000; K. Nelson, 2007; Spinrad & Stifter, 2006).
Language difficulties might negatively affect the acquisition of this knowledge, leading to problems with self-regulation and subsequently to problem behaviours
(Salmon et al., 2016). Equally, good language skills might facilitate this development, as children acquire the symbolic means to communicate and understand wants and needs sooner. Recently, evidence has emerged that pre-schoolers with DLD may indeed have delayed emotion recognition (Rieffe & Wiefferink, 2017). Furthermore, DLD and impaired emotion recognition, known as alexithymia, have recently been found to co- occur in the youth justice population, although the authors suggest that this is likely to be a co-morbidity rather than a causal relationship (Snow, Woodward, Mathis, &
Powell, 2016). Nevertheless, this co-occurrence is a promising area of investigation.
An emerging body of work is investigating the role played by parents’ “emotion socialisation behaviours” (Eisenberg et al., 1998) in the development of SEB difficulties
71
in childhood and adolescence. These include parental reactions to, and discussion about, emotions, and emotion coaching (Klimes-Dougan, 2007). A recent meta-analysis reported small effect sizes for their concurrent (r = -0.08) and longitudinal (r = -0.11) associations with conduct problems across childhood and adolescence, concluding they could inform parent-mediated interventions for conduct problems (A. M. Johnson,
Hawes, Eisenberg, Kohlhoff, & Dudeney, 2017). Related work also suggests that elaborative reminiscing, that is, detailed, emotional, and collaborative discussion about the past, between parents and children aged 18 months to five years fosters children’s language and socio-emotional development by encouraging the use of challenging language, putting thoughts and feelings into words, and helping memory (Salmon &
Reese, 2016). Further investigation is required to determine whether spontaneously occurring EMST during naturalistic interactions is associated with early language or
SEB outcomes.
3.8 Fillers and Missed Opportunities
As mentioned in section 3.4, the difficulties in establishing reliability for child initiations and responses precluded these behaviours being included. Instead, maternal missed opportunities to respond and conversational fillers (e.g. ‘hmmm’, ‘uh huh’) were included to represent no input and low-level input respectively, in contrast to contingent, responsive input.
There is little directly relevant literature for these two behaviours, especially for
‘fillers’. Robinson and Eyberg (1981) coded ‘irrelevant verbalisation’ which could be considered a ‘filler’. Rice et al. (1990) coded maternal ‘ignore’ which could be considered a type of missed opportunity, although ‘ignore’ suggests a deliberate act,
72
whilst missed opportunities includes maternal lack of awareness. High frequencies of these behaviours might indicate a passive interactive style, where the mother observes more than participates (Gridley et al., 2015). This interactive style might have a negative impact on children’s language learning and self-esteem, as exposure to rich linguistic input and positive reinforcement might be low. There has been some speculation that non-responsiveness may be part of over-protective parenting, where the parent is involved but not responsive to the child’s cues or needs. This might undermine the child’s developing sense of self-efficacy (Mills & Rubin, 1998). One study reported that mothers of primary school-aged children with internalising symptoms (n = 23) failed to respond to their children’s signals more frequently that mothers of typically developing children (Mills & Rubin, 1998). However, it is currently unknown whether there is any difference in non-response across the distribution of SEB or language abilities.
Having described the maternal behaviours chosen, the following two sections describe the two types of interactive behaviour selected: Joint Engagement and Fluency and Connectedness (Adamson, Bakeman, Deckner, & Nelson, 2012; Hirsh-Pasek et al.,
2015). These two interactive behaviours were examined as potential predictors of language and SEB outcomes, but also as potential modifiers of the effects of the maternal behaviours. Both behaviours were part of a battery of items designed to rate separate but related aspects of parent-child interaction (Adamson et al., 2012). The joint engagement ratings characterise the amount and quality of two different forms of joint engagement, supported and coordinated. The Fluency and Connectedness rating captures the way the shared topic is sustained and embedded in a fluent, connected
73
conversation by both parent and child. Later, Hirsh-Pasek et al (2015) conceptualised these items as ratings of interaction quality. Both are discussed below.
3.9 Joint Engagement
Joint engagement is where mother and child are engaged in joint attention for a period of at least three seconds, that is, they show active attention and sustained interest in a shared object or activity, and are aware of each other doing so (Adamson &
Bakeman, 1991; Moore & Dunham, 1995; Tomasello & Todd, 1983). Joint engagement skills develop between nine and 15 months: during interactions children begin to alternate their gaze from person to object and back again, and to point, show and give
(Adamson, Bakeman, & Deckner, 2004; Beuker, Rommelse, Donders, & Buitelaar,
2013; M. Carpenter et al., 1998; Mundy et al., 2007).
Joint engagement is thought to create a shared referential framework (M.
Carpenter et al., 1998) that helps children make the correct word-referent connections
(Bruner, 1975; Tomasello & Akhtar, 1995; Yu & Ballard, 2007), and provides a scaffold for effective interaction. Joint engagement skills are positively associated with language and social competence skills (Farrant, Devine, Maybery, & Fletcher, 2012;
Kasari, Gulsrud, Freeman, Paparella, & Hellemann, 2012; Tomasello & Todd, 1983).
Difficulties with establishing or maintaining joint engagement might contribute to problems with language development and interaction (Mundy, Kasari, Sigman, &
Ruskin, 1995). Indeed, children with autism spectrum disorders and late talkers have been shown to have difficulties with joint engagement compared with typically developing children (S. Y. Patterson, Elder, Gulsrud, & Kasari, 2014; Paul & Shiffer,
1991; Wetherby, Yonclas, & Bryan, 1989).
74
Bakeman and Adamson (1984) developed a comprehensive coding scheme to examine whether the way in which parent and child coordinate their attention during interactions contributes to language learning (Adamson & Bakeman, 1991; Adamson et al., 2004; Adamson, Bakeman, Deckner, & Romski, 2009; Bakeman & Adamson,
1984). The codes describe children’s engagement states during interactions, from unengaged through to being engaged in supported or co-ordinated joint engagement (S-
JE and C-JE respectively). Codes also exist to indicate whether the child attends to words and symbols during joint engagement (known as symbol-infused joint engagement), evidenced by speaking or following instructions. In their study of typically developing 18 month olds (n = 56), symbol-infused S-JE, but not symbol- infused C-JE, was positively associated with expressive and receptive vocabulary at 30 months (Adamson et al., 2004; Adamson et al., 2009). It is unclear whether S-JE and C-
JE are associated with later language regardless of symbol-infusion, or whether they are associated with SEB development. Joint Engagement is discussed in detail in the paper comprising Chapter Seven.
3.10 Fluency and Connectedness
The final interactive behaviour selected was Fluency and Connectedness
(Adamson et al., 2012). This is a measure of the flow and cohesion between mother and child. It has been conceptualised as a measure of interaction quality, along with Joint
Engagement, to which both mother and child contribute (Hirsh-Pasek et al., 2015).
Evidence suggests that Fluency and Connectedness is positively associated with later expressive language scores in low-income children (Hirsh-Pasek et al., 2015), and mediates the impact of adversity on early communication skills (Smith et al., 2017).
Whether Fluency and Connectedness is directly associated with language and SEB 75
development has yet to be investigated in a large-scale community-derived sample.
Furthermore, whether interaction quality modifies the effects of maternal behaviours on child outcomes is unknown. Fluency and Connectedness is discussed in detail in the paper comprising Chapter Six.
3.11 Summary of Maternal and Interactive behaviours
This Chapter has described the context in which this study will be undertaken, and how it relates to and builds on the MRBS (Levickis et al., 2014). The procedures for identifying maternal and interactive behaviours of interest were described, as well as a brief overview of the literature to compliment later chapters. The following section presents the overall aims of the PhD.
3.12 Aims Overview
In a community-based sample of slow-to-talk toddlers I aimed to explore the contribution of maternal, child and interactive communicative behaviours to child language and social, emotional and behavioural development.
3.12.1 Aim One
To examine to what extent language skills and SEB adjustment are associated in a population-based prospective study of slow-to-talk toddlers. Specifically, I aimed to examine whether:
1. (a) receptive and expressive language and internalising and externalising
behaviours are associated at 24, 36 and 48 months, and (b) the form of these
associations (i.e. linear vs. non-linear);
76
2. the associations vary across the different ages;
3. the associations may be explained by biological or environmental confounders.
3.12.2 Aim Two
To examine whether:
(a) maternal and (b) interactive communicative behaviours observed at child age
24 months were associated with child receptive and expressive language at 24,
36 and 48 months;
(c) maternal responsive communicative behaviours modified the effect of
maternal intrusive communicative behaviours; and
(d) the cross-sectional associations between maternal communicative behaviours
and language scores were modified by interactive communicative behaviour
ratings.
3.12.3 Aim Three
To determine the extent to which maternal and interactive communicative behaviours observed at child age 24 months were associated with child internalising and externalising behaviour scores at 24, 36 and 48 months.
77
4 Methodology
4.1 Overview
This Chapter describes the study methodology for the thesis. There is overlap with the manuscripts that form Chapters 4, 5 and 6. However, more detail is given here about the study context, the Maternal Responsive Behaviours Study (MRBS) on which this thesis builds (Levickis et al., 2014), and the development of the two coding schemes, which formed a major component of this project.
78
4.2 Study Context
This PhD study was nested within two existing studies, Let’s Learn Language
(LLL) and Language for Learning (L4L). An overview of each is provided below, and further details can be found in Wake et al. (2011) and Wake et al. (2012). This PhD builds upon a sub-study of LLL, the Maternal Responsive Behaviours Study (MRBS,
Levickis et al., 2014), for which an overview is also provided. The procedures used in the MRBS are described because there is overlapping methodology between the MRBS and this PhD. Figure 1 describes the flow of participants through all the stages.
4.2.1 Let’s Learn Language
The first study was a cluster randomized-controlled trial (RCT) based in a population-level survey, Let’s Learn Language (LLL) (NHMRC #384491). Three of
Melbourne’s 31 local government areas were selected for recruitment, representing lower (Frankston), middle (Banyule) and upper (Kingston) socio-economic status neighbourhoods, as measured by the Australian census-derived Socio-Economic
Indexes for Areas (SEIFA) Index of Relative Social Disadvantage (IRSD) (ABS, 2001).
Potential participants (mothers and infants) were identified by maternal and child health
(MCH) nurses from Council records of births between May and December 2006. MCH nurses told parents about the study when they attended their child’s 12 month old visit between May and December 2007. The study team sent a Parent Information Statement, consent form and baseline questionnaire to interested families (see Appendices A and
B). The majority of the 1,451 participants eligible for recruitment completed the baseline questionnaire and consented to entry into the trial if they were subsequently eligible 1,217 (83.9%).
79
Figure 1 Flowchart of participants and measures Adapted from Levickis et al., 2014
Notes: LLL Let’s Learn Language; L4L Language for Learning; SSLM Sure Start Language th Measure; PLS-4 Preschool Language Scale 4P P Edition; CBCL Child Behaviour Checklist; CELF-P2 Clinical Evaluation of Language Fundamentals Preschool Edition; SDQ Strengths and Difficulties Questionnaire; ASD Autism Spectrum Disorder; 6x responsive behaviours were expansions, imitations, responsive questions, labels, supportive directives, and interpretations
80
Exclusion criteria were as follows: known developmental delay, Down’s syndrome, Cerebral Palsy, other serious intellectual or physical disability, major medical condition, suspected autism spectrum disorder, and parents with insufficient
English to complete written questionnaires (written at grade 6 reading level, typically age 11-12 years).
Six months later, at age 18 months, parents were sent a questionnaire containing a screening survey for expressive vocabulary skills, the 100-word Sure Start Language
Measure (SSLM) (Roy, Kersley, & Law, 2005) which 93.5% (n = 1,138) of parents completed. Children who scored at or below the 20th percentile based on population norms for the screen (n = 301, 26.4%) were invited to participate in the LLL trial. This equated to speaking either no or very few words (the mean number of words spoken by the 301 participants was 5.8, SD 3.1). The study-specific term of ‘slow-to-talk’ was coined to describe the participants. This selection criterion was used to cast a wide net in an effort to capture more children at risk of later difficulties. They were allocated to either the intervention (n =158) or control arm (usual care) (n = 143) of the RCT depending on which MCH centre they attended. The centres, or clusters, had previously been randomly allocated to the intervention or control arm based on which local government area they were in and the number of eligible participants that attended them. This is described in more detail by Wake et al. (2011). LLL received ethics approval from the Royal Children’s Hospital Human Research Ethics Committee (see certificate in Appendix D).
The LLL intervention was a modified version of The Hanen Centre’s “You
Make the Difference” programme, a group-based parent-toddler language promotion programme delivered over six weekly sessions (Manolson et al., 1995). Findings 81
showed that there was little evidence of differences in language or behaviour outcomes between the intervention and control groups (Wake et al., 2011). Therefore, the control group and intervention group participants were pooled together in one group for this thesis, and all adjusted analyses controlled for intervention group status.
4.2.1.1 Relevant LLL Measures
Demographic information was collected in the baseline questionnaire at 12 months. At 24 and 36 months parents reported on their child’s social, emotional and behavioural (SEB) difficulties in the Child Behaviour Checklist (CBCL, Achenbach &
Rescorla, 2000). Face-to-face language assessments were conducted in their homes when the children were 24 and 36 months using the Preschool Language Scale, 4th
Edition (PLS-4, Zimmerman, Steiner, & Pond, 2002). These measures are described in section 4.8.
4.2.2 Maternal Responsive Behaviours Study
At 24 months, the LLL participants (n = 301) were invited to participate in a sub- study called the Maternal Responsive Behaviours Study (MRBS) (Levickis et al. 2014).
The purpose of this study was to collect and code video recordings of mothers and their two-year-old children during free-play to investigate the contribution of maternal responsive behaviours to child language outcomes. 251 (83.4%) mothers consented to participate (see Figure 2).
4.2.2.1 Observational Procedure
A research assistant (RA) visited the families in their homes, blind to their LLL intervention group status. Before starting the video-recording, the RA set up a play space, usually in the sitting room or kitchen, by removing as many distractions as 82
possible e.g. child’s own toys. The RA placed two bags of toys on the floor a short distance apart, one containing a farmyard and accessories, and the other a doll and accessories. The RA placed the doll and the barnyard beside each bag, and left the accessories in the bags for the child to explore themselves if they wanted to. The RA then used the following script to introduce and explain the task to the mothers:
“Now I will videotape you and your child during free-play for 15 minutes. The purpose of the video is to observe how your child communicates with a familiar person.
I have set up a nurturing activity with a doll and related toys and a farm activity with a barn and animals. Your child may go to whichever activity he or she is interested in and may switch from one activity to another at any time. Please play with your child as you normally would. If your child stops playing or protests, I will stop the camera.”
If the child was uncooperative or became upset, the appointment was re- scheduled. If the child preferred to play with his or her own toys, the mother was told that this was okay. The interaction was video-recorded for 15 minutes using a Hitachi
DZ-GX5060SW DVD camcorder mounted on a tripod, and recorded onto Sony DVD
DS DVD-RW discs. The RA set up the camera so that the mother, child and toys could be seen in the frame. Recording started once the mother and child were ready.
Recordings were downloaded onto a DVD and backed up to an external hard-drive.
83
Figure 2 Flowchart of participants through LLL Survey and Trial
Reproduced from Wake et al. (2011)
84
4.2.2.2 Maternal Responsive Behaviours Study Measures
As described in the Literature Review (Chapter Three, section 3.2), Levickis et al. (2014) coded maternal expansions, imitations, responsive questions, labels, supportive directives and interpretations. These behaviours are defined in Table 3.
Table 3 Definitions of Maternal Responsive Communicative Behaviours Behaviour Definition Example Expansion Mother repeats one or all of the child’s C: “Ball”, M: “It’s a preceding words and adds to the child’s red ball” preceding verbalisation Imitation Mother repeats the child’s preceding C: “Ball”, M: “Ball” vocalisation or verbalisation exactly or with a reduction in the words. Responsive Mother asks a “wh” question (e.g. “what”, Child is playing with question “when”, “who”), which is immediate and horse dependent on the child’s preceding act. M: “What’s that?” Supportive Mother directs or commands the child to Child is playing with directive verbalise or do an action in relation to what doll the child is focused on. M: “Feed the doll” Label Mother labels an object or action, which is the Child is playing with a focus of the child, with the label in the final toy horse position of the carrier phase. M: “That’s a horse” Interpretation Mother responds to the child’s vocalisation or C: “Bi Fir Ti” verbalisation with one or several words to M: “She’s a bit interpret the child’s intended message using thirsty” the context as cue
Note: Behaviours coded in the Maternal Responsive Behaviours Study (Levickis et al. 2014)
All behaviours were contingent on the child, occurred within episodes of joint attention during the ten-minute recording of mother-child interaction, and followed into the child’s focus of attention. Expansions and imitations are dependent on a child verbalising or vocalising, which means that a child has to be at a certain stage of development for the mother to respond, whereas the child does not have to verbalise or vocalise for the other behaviours. The mother is responding to her child’s focus of 85
attention rather than the child’s language, but this suggests that mother is responding to child’s level of development, i.e. child with better language, mum more likely to use expansions/imitations as more opportunity).Relevant to analysis in this thesis are expansions, imitations and responsive questions
4.2.3 Language for Learning
When the LLL participants were aged 48 months, all the families who had consented to take part in LLL when their children were 12 months old (n = 1,217) were invited into a new study called Language for Learning (L4L) (NHMRC #60740) (Wake et al., 2012). This included the 251 children who took part in the MRBS (see Figure 1).
See Appendix E for parent/guardian information statement and consent form, and
Appendix F for the ethics approval certificate from the Royal Children’s Hospital
Human Research Ethics Committee. L4L was an RCT of a year-long population-based intervention targeting four-year-old children with language delay. The first step was to administer a formal language assessment to all children on their fourth birthday.
Children scoring more than 1.25 SD below the normative mean were recruited into either the intervention or control (usual care) arm of the L4L study. Of the 251 families who took part in the MRBS, 204 (81.3%) participated in the four-year-old assessment for the L4L trial.
4.2.3.1 Relevant L4L Measures
At 48 months, parents were sent a questionnaire which included the Strengths and Difficulties Questionnaire (SDQ) (Goodman, 2001). A face-to-face language assessment was conducted in the children’s homes using the Clinical Evaluation of
86
Language Fundamentals - Preschool 2nd Edition (CELF-P2) (Semel, Wiig, & Secord,
2006). These measures are described in section 4.8.
4.2.4 This PhD: Maternal and Interactive Behaviours Study
This PhD brings together data from participants who took part in three inter- connected studies: LLL (from 12 to 36 months), the MRBS (at 24 months), and L4L (at
48 months). The language and SEB difficulties measures were collected as part of LLL and L4L. The video-recordings of mother-child interaction and the maternal responsive behaviours codes were collected as part of the MRBS. This thesis builds upon the findings of this latter study by examining other maternal and interactive behaviours. The participants, coding, and measures will be described in the following sections.
4.3 Participants
Data were available for 204/251 (81.3%) families who took part in the MRBS.
202 consented to their data being used in other studies and therefore were eligible to be included in this thesis (see Figure 1). Five participants were excluded from the analysis: four were diagnosed with autism spectrum disorder when they were aged three or four years of age, and one video recording was discarded because the grandmother switched places with the mother part way through. The final sample therefore comprised 197 participants, which, as shown in Table 4, did not differ substantially from the 251 for whom we had video-recordings except for maternal age (1.3 years older in this sample).
87
Table 4 Comparison of MRBS Sample and MIBS Sample Characteristic, MRBS MIBS p M (SD) or n (%) n = 251 n = 197 Gender, male 129 (51.4) 103 (52.3) .302 Birth order, first born 96 (38.3) 73 (37.1) .374 Birthweight 3400.1 (624.3) 3380.3 (620.2) .738 SEIFA score of disadvantage 1027 (48.0) 1026.5 (53.3) .917 Parental education Not completed high school 54 (21.7) 38 (19.4) .509 Completed high school 86 (34.1) 65 (33.2) .395 Tertiary education 111 (44.2) 93 (47.5) .294 Maternal age (years) 33.0 (4.5) 34.3 (4.5) .003 Treatment status - 129 (51.4) 100 (50.76) .326 intervention Receptive Language 24m 88.9 (14.4) 90.6 (14.0) .212 36m 96.9 (16.3) 98.4 (15.2) .337 Expressive language 24m 90.5 (12.0) 91.8 (11.9) .256 36m 99.6 (15.3) 101.1 (14.5) .310 a Responsive behaviours (rpm P )P coded in MRBS Expansions 0.6 (0.6) 0.6 (0.6) 1.0 Imitations 0.5 (0.5) 0.5 (0.5) 1.0 Responsive questions 0.7 (0.5) 0.7 (0.5) 1.0 Labels 1.2 (0.7) 1.2 (0.7) 1.0 Supportive directives 0.6 (0.4) 0.6 (0.5) 1.0 Interpretations 0.5 (0.5) 0.6 (0.4) .023
Note: MRBS Maternal Responsive Behaviours Study (Levickis et al., 2014); MIBS a Maternal and Interactive Behaviours Study (this PhD study); P rpmP = Rate per minute
88
4.4 Coding Overview
The development of two coding schemes, one for maternal behaviours and one for interactive behaviours, was a major component of this thesis. The rationale for the maternal and interactive behaviours was presented in Chapter Three. Extensive piloting was conducted to test the ease of coding, refine the definitions, ensure the behaviours could be coded in a replicable way, and to check the behaviours were mutually exclusive. The piloting of both Schemes is described in detail below. It was conducted by the author and another PhD candidate, Jodie Smith, using five videos of mother- toddler interactions. After completing each of the pilot videos, any discrepancies were discussed and resolved before progressing to the next recording. Agreement between the coders needed to reach 80% before the scheme was considered acceptable, i.e. definitions were clear and replicable, to finish the piloting stage.
Once piloting was complete, participant IDs were organised in a random order generated in Excel. The sample was then batched into groups of approximately 50 participants. Coding was completed in batches, and one in ten IDs in each batch was randomly selected in Excel for reliability coding, i.e. approximately five videos per batch. Intra and inter-rater reliability were calculated before progressing to the next batch (described below). This protected against drift and ensured any necessary re- coding was manageable. Coding took place for four hours maximum a day to help maintain coding fidelity.
89
4.5 Development of Scheme One: Maternal Behaviours
The Maternal Behaviours Coding Scheme One was applied using specialist observation software called Observer ®XT. The middle ten minutes of each video were coded, as per the coding protocol developed during the MRBS (Levickis et al., 2014).
Maternal praise, missed opportunities, successful directives, unsuccessful directives, fillers and emotion and mental state talk (EMST) were the codes included in the final
Scheme One. However, as described in section 4.5.2 below, other behaviours were also piloted before the final behaviours were chosen. Each video took between 15 and 20 minutes to code. This next section describes the procedure for using Observer ®XT.
4.5.1 Observer ®XT Coding Procedure
Observer ®XT is professional software designed for use with observational data
(Noldus, 2008). Levickis (2013) conducted a comprehensive review of available software and found Observer ®XT to be the most suitable. It allows videos to be uploaded, watched in real time, and behaviours to be coded with the press of a button using continuous recording. This means recording the occurrence of each behaviour at the time of its occurrence within the video. Up to six behaviours can be coded at one time with adequate reliability.
Once a video was uploaded into Observer ®XT, the coder entered the participant’s ID number, their own initials and the day’s date. Other useful information was also entered, in this case, who the caregiver in the video was, and whether a sibling was present. The coder disregarded the first 2.5 mins of each recording as ‘warm-up time’ between the mother and child. The final 2.5 mins was discarded as ‘winding
90
down’ time. Coding therefore started after 2.5 mins had elapsed, and each behaviour was coded as it occurred until the software automatically stopped after 10 minutes. In a small number of cases (<10), the recording ended before the10 minutes elapsed. In these cases, coding continued until the recording stopped or the mother was observed speaking to the research assistant about ending the session. Observer ®XT recorded the total number of each behaviour and the duration of the recording, from which it calculated the rate per minute for each behaviour for each participant. This information was exported into Excel, and then into Stata for analysis.
4.5.2 Piloting Scheme One: Maternal Behaviours
The original plan for Coding Scheme One was to code five maternal behaviours
(direct directives, indirect directives, praise, emotion and mental state talk production and elicitation) and two child behaviours (verbal and nonverbal communicative acts, and compliance with all directives). The piloting process was complex and time- consuming, and is described below. A summary of all the behaviours piloted for Coding
Scheme One, relevant notes and modifications made to their definitions is presented in
Table 5.
During piloting, it became immediately apparent that it was difficult to focus simultaneously on child and maternal behaviours using continuous coding. Furthermore, it was not possible to accurately record the children’s non-verbal communication because they were often out of frame or had their backs to the camera. For these reasons, maternal behaviours became the focus of Scheme One. The child codes were replaced with two additional maternal behaviours – fillers and missed opportunities.
Emotion and mental state talk (EMST) elicitation did not occur with enough frequency
91
to warrant a separate code from EMST production, so the two EMST codes were collapsed into one.
Piloting also revealed difficulties with coding directives that were direct (e.g.
“get the red one”, “look at this”) and directives that were indirect (e.g., “How about you try the red one?”). It was difficult to distinguish between ‘direct’ directives and
‘supportive’ directives that had been previously coded in the MRBS (see definition in
Table 3). Similarly, it was difficult to distinguish between ‘indirect’ directives and previously coded ‘responsive questions’ (Levickis et al., 2014). This was because the
‘direct’ and ‘indirect’ definitions did not distinguish between whether the mother was following into or leading away from the child’s attention, meaning there was overlap in the codes. This was problematic because the codes needed to be mutually exclusive and clearly distinguishable for reliability purposes. Despite refining the definitions of direct and indirect directives to only include those that led away from the child’s focus of attention, the distinction remained challenging. The coder had to decide whether a directive had been given, whether it contained an imperative verb (i.e. was direct or indirect), and whether it followed or lead the child’s attention. These decisions were too onerous to make in close to real time as is necessary when using continuous coding.
Therefore, these definitions were discarded, and the literature revisited to select a different type of directive hypothesised to be associated with language outcomes.
The second type of directive piloted distinguished between ‘intrusive attentional directives’ which commanded a shift in visual attention, and ‘intrusive behavioural directives’ which commanded a change in behaviour. Again, it was challenging to quickly distinguish between these two types of directives when attempting to code in real time. Mothers often used multiple directives in quick succession. This meant it was 92
difficult to code them reproducibly, leading to low agreement between coders and even within coders. For a second time, it was necessary to return to the literature and select a different type of directive. It was imperative to choose a definition that was simple to apply during continuous coding, and which was clearly distinct from maternal behaviours that had been coded as part of the MRBS.
The final definitions for directives, successful and unsuccessful intrusive directives, were objective, clear, and mutually exclusive. For a directive to be coded, a mother had to command the child to look at, say, or do something outside of the child’s immediate focus of attention, and the directive had to contain an imperative verb (with the exception of calling child’s name). For example, a child is playing with a doll, and the mother says, “Look in that bag!”. To incorporate an element of the child’s contribution, the directive was coded as successful if the child attempted to comply with it, or unsuccessful if the child made no attempt to comply.
93
Table 5 Summary of Behaviours Piloted for Coding Scheme One Behaviour Decision Definition Piloting notes Direct directive Exclude A command containing a imperative verb. e.g. “Get This definition overlapped with the Supportive the red one!”, “Look at this!” Directives definition used in the MRBS which were commands in relation to what the child was currently focused on. Indirect directive Exclude A ‘softer’ command that does not contain an This definition overlapped with the Responsive imperative e.g. “Let’s look at the farm.” Questions definition used in the MRBS which were questions in relation to what the child was currently focused on. Intrusive Exclude Mother commands or directs attentional focus away Mothers often used attentional and behavioural attentional from object or action the child is currently focused directives in quick succession, which was hard to directive on and toward something new or different e.g. calls code in real time. child’s name, “Look over there!” Intrusive Exclude Mother commands or directs the child to an behavioural object/activity different from child’s current focus directive with the intention of restraining or restricting behaviours. e.g. “Get the bottle out of the bag!” Child compliance Exclude Compliance - Child attempts to follow directive (any Refined the definition to clarify that compliance or non-compliance type) within two turns. could be any attempt, whether correct or incorrect. with directive Non-compliance - Child does not attempt to follow Was hard to simultaneously distinguish between directive (any type) within two turns different types of directives and also watch for the child’s compliance without pausing the video, rewinding and re-watching.
94
Behaviour Decision Definition Piloting notes Successful and Include Utterance that aims to re-direct child away from the Combined child compliance with intrusive directive Unsuccessful object or activity on which they are currently visually into one code. Directive focused to something new or different. Is a type of intrusive directive. Successful: when child attempts to follow. e.g. C playing with truck. M, “Look in the bag!” C, looks in bag. Unsuccessful: when child does not attempt to follow. e.g. M, “Look in the bag!” C, no attempt to look. Praise Include Mother uses verbal or nonverbal positive evaluations Refined coding rules and inclusions and exclusions. of child’s efforts, global characteristics, or generic Count each praise utterance, even one after the other. positive evaluations. Include: Examples “Yeah!” with an upwards inflection like a cheer in Child’s efforts: “Good job!”, “Great listening”, “You response to action/utterance. tried hard” Exclamations in response to behaviour e.g. “wow!” Child’s action: “Nice cuddling!” Do not include: Child’s global characteristics: “Good girl/boy!” Comments about child’s action unless accompanied Generic positive evaluations: “Good!”, “Awesome!” by positive evaluation e.g. “You fed the doll” vs Non-verbal praise (behavioural indices of praise, “You fed the doll nicely”. Gaertner et al 2008): clapping in response to child’s “Yeah” or “Yes” as an answer, agreement or actions, high five, thumbs up, pat on the back confirmation; “There you go!” Encouragement e.g. “You can do it!” (as per Henderlong and Lepper, 2002), or encouraging nod
95
Behaviour Decision Definition Piloting notes Emotion and Include Mother uses talk about desire, emotions, mental Refined inclusions and exclusions. Count each mental state talk: states, other internal states, or empathy statements. emotion or mental state word e.g. I think the baby is production (EMST) e.g. “Baby’s sad/hungry/upset”, “I think…” sad = 2 vs I think so = 1 Emotion labels and explanations: Nouns, verbs, The most common are: Like, think, know, need, want adjectives, or adverbs naming emotional feelings and Include: utterances that explain or elaborate on the reason…of “I’m not sure”, “I believe” an emotion Mother speaking on behalf of toy e.g. “How are you? Desires: References to wanting or needing something I’m thirsty” on behalf of toy dog concrete. Include questions to child e.g. “Do you “What’s wrong?” want a drink?” “Do you need the toilet?” Expression of relief e.g. “Phew” Other mental states: References to the past, or to Apologies e.g. “I’m sorry”, “sorry!” thinking, knowing, wondering, believing. Include “It’s sore”, “It hurts” questions e.g. “What do you think?” “Would you like Do not include: me to help you get it?” “You don’t like it”. Mental state terms that serve a conversation function Other internal states: References to other internal and do not refer to the act of mentalising e.g. “I states that are not affect- or mental-state related e.g. dunno”, “like” physiological states “Here you go” or “There you go” Empathy statements: Statements or emotion-related Repetition C: “Happy” M: “Happy” sounds that promote empathy with a character’s Verbs to do with emotion e.g. crying, laughing emotion “Like” when used as comparison e.g. “It’s like the window in Play School” “Be gentle “or “Be nice with baby” Statements about child’s personality “You’re noisy / cheeky / crazy / naughty / silly”. “She’s asleep”, or “She’s sleeping”
96
Behaviour Decision Definition Piloting notes Emotion and Exclude Mother asks the child to label or explain emotion or Occurred too infrequently to warrant individual code. mental state talk: mental state terms. e.g. “Why is baby sad?” Merge with EMST production to form EMST code. elicitation Child Exclude Verbal or non-verbal social communicative acts It was difficult to code mother and child behaviours communicative addressed to mother: gestures, imitations, simultaneously. behaviour vocalisations that are clearly communicative in Hard to code non-verbal behaviours when child nature, words, sentences. facing away from camera or moving around. Fillers Include Low level input to child which counts as mother’s Often mother used a ‘filler’ just before a verbal turn. utterance. For example, “Hmmm?...What do you e.g. “Hmmm?”, “Uh huh”. think?” This doesn’t really count as low level input. Hence, the filler must be standalone, with no other maternal utterance from mother within 3 secs. Missed Include Mother does not respond in any way to child’s Enter code on the fourth second. opportunities utterance within three secs.
97
4.6 Development of Scheme Two: Interactive Behaviours
Specialist observation software was unnecessary for coding the interactive behaviours in Coding Scheme Two. Instead, video recordings were observed in
Windows Media Player and ratings entered in an Excel spreadsheet. The final coding scheme comprised ratings for Supported Joint Engagement (S-JE), Coordinated Joint
Engagement (C-JE), and Fluency and Connectedness (FC). The rationale for these behaviours was presented in Chapter Three (section 3.9 and 3.10). After advice from one of the developers of the engagement codes and her colleague (Professor Adamson and Dr Suma at Georgia State University), it was decided that using the middle five minutes of each video-recording was sufficient to be representative of each mother- child interaction. Each video took approximately seven minutes to code. Professor
Adamson and Dr Suma generously answered questions and assisted with coding queries during this process.
4.6.1 Piloting Interactive Behaviours Coding Scheme Two
The initial plan for Scheme Two was to use mutually exclusive and exhaustive moment-by-moment observational codes to code the child’s engagement states during the middle 10 minutes of the recordings (Adamson et al., 2004; Adamson et al., 2009;
Bakeman & Adamson, 1984). The intention was to use the time-stamps attached to the maternal behaviours coded during the MRBS to map them onto the child’s engagement state. This was to determine whether the engagement state at the exact time when the maternal behaviour was used modified its association with child outcomes. See Table 6 with the summary of the codes piloted for the Interactive Behaviours coded in Coding
Scheme Two.
98
Piloting revealed the engagement scheme was too difficult to be used by a single coder, using one camera angle, and with the participants moving in and out of frame. It required the child’s engagement state for the entire duration of the recording to be coded, including when they were unengaged or solely focused on their mother or the object. This was difficult to do when the child’s back was to the camera or they were moving around. Fortunately, the moment-by-moment codes had recently been converted into a battery of rating items with which there were well correlated (engagement items r
= 0.66 to r = 0.77) (Adamson et al., 2012). This meant that each three second interval of time did not need to be assigned a code, rather, an appropriate duration for each engagement state sufficed, and it was not necessary to code all the different states of engagement. The decision was taken to pilot two joint engagement items from the battery (Supported and Coordinated Joint Engagement, or S-JE and C-JE), and an item that rated the quality of the mother-child interaction, called Fluency and Connectedness
(FC). The following section will describe the piloting for the joint engagement items and then the FC item.
The rationale for coding joint engagement was described in Chapter Three
(section 3.8). A state of joint engagement occurred when the mother and child were actively focused on the same object or event, at the same time for at least three seconds, and were aware of each other doing so. S-JE was when the child was focused almost exclusively on the shared object or event (but beyond just listening), and C-JE was when the child was coordinating his or her attention from the object or event to the mother.
The procedure for coding S-JE and C-JE was developed during piloting, using the technical manual from Adamson et al. (2012) and through consultation with 99
Professor Adamson. The key task was to decide whether mother and child were actively involved with the same object or activity, and whether this engagement was coordinated or supported, usually distinguished by eye contact between the child and mother. On a piece of paper, a vertical line was drawn to represent the five-minute timeline of the recording. To the left of the line, the time the mother and child entered and exited the engagement state was noted. When the engagement state was C-JE, a ‘C’ was written next to it. The rest was S-JE by default. To the right of the line, episodes of matched affect (e.g. mother and child smiling at each other), high energy, intensity, concentration and conversation progression were marked, with pluses (+) indicating high quality.
The approximate amount of time (in seconds) of each state was tallied to estimate the total time in S-JE and C-JE. This estimate was used to anchor the rating on a seven-point Likert scale. During piloting, an approximate duration in seconds was assigned to each of the seven scale points, based on the five minutes (or 300 seconds) duration of each recording. A rating of four represented approximately one third of the time in joint engagement (Adamson et al., 2012). Hence, rating four was assigned a duration of approximately 90 to 100 seconds. Finally, the quality of the engagement state was judged by using the notes written to the right of the line. If the engagement state was considered high quality, the rating was increased by one scale point. The coding rules were expanded during piloting, and inclusions and exclusions added. These are detailed in Table 6.
The third item piloted was Fluency and Connectedness (FC). This measure captured the flow and cohesion of the conversation between the mother and child, which could comprise verbal and non-verbal turns. The procedure for coding FC is describe within the published paper presented in Chapter Six. The piloting process resulted in the 100
development of examples of behaviours evident under each of the seven ratings (see
Table 6) to expand upon the anchors provided by Adamson et al. (2012). A rating of one indicated that no conversation had been established. A rating of four indicated that although the conversation was dominated by one partner (usually by the mother), the child did initiate at least twice, there were at least three turn taking episodes, and mother and child were mostly on the same topic. To be rated seven, the mother and child had to be engaged in fluid and sustained conversation. The turns had to flow beyond the child simply showing or giving something to the mother, and the mother acknowledging this, they had to be smooth, build from the last turn, and both mother and child had to extend the play.
101
Table 6 Summary of Behaviours Piloted for Coding Scheme Two
MomentU -by-moment engagement coding scheme (adapted from Adamson et al., 2004) General rules: Assign an engagement code to every moment of the video-recording; each state must last a minimum of three seconds; partner must respond within three seconds to count as Joint Engagement; record all bids for joint attention by selecting ‘mother initiator’ or ‘child initiator’; select whether it is Supported or Coordinated Joint Engagement at the end of the episode Behaviour Decision Definition Piloting notes Unengaged Exclude Child is uninvolved with toys or mother When child had back to camera or was off screen it was difficult to observe engagement state every 3 secs. Person/object Exclude Child is focused solely on toy or mother As for ‘unengaged’. focused Joint Switch to Mother and child actively involved with the same Difficult to determine an initiation during continuous engagement rating object or activity at the same time, and are aware of coding. each other doing so. Do not include ‘housekeeping’ or background narration of child’s activity which does not influence child’s play e.g. “You’re putting the horse in barn”. Supported Switch to Child does not overtly acknowledge mother. Occurs more frequently than Coordinated. rating If mother puts a relevant toy in front of child, who then incorporates into the game without acknowledging the mother, code as Supported e.g. child playing with toy horse, mother puts toy apple in front of horse, child proceeds to pretend horse is eating the apple. Coordinated Switch to Child overtly acknowledges mother e.g. eye contact Include eye contact, and talking directly to mother. rating If mother asks a question and child replies without looking, code as Coordinated.
102
RatingU Items coding scheme for Joint Engagement and Fluency and Connectedness (adapted from Adamson et al., 2012) General rules: Rate overall Supported and Coordinated Joint Engagement, and flow and cohesion of conversation in middle five minutes of video-recording; note time joint engagement episode starts and ends; note periods of high quality interaction; rate on seven-point Likert scale. Behaviour Decision Definition Piloting notes Supported Include Child is focusing almost exclusively on the shared Assigned time ranges in seconds to each rating to help Joint object or event (but beyond just listening). anchor rating. Engagement Notes from technical manual: Note down start and end time of each episode and tally at (S-JE) Mother’s participation influences child’s end. This gave estimate of total time in C-JE and S-JE activity/experience with object/event but child during the 5 mins. does not acknowledge this involvement. A command could be the start of a new episode but not a Mother is often actively manipulating object or marker. If followed command with sustained attention, making statements/commands that alter child’s code including command. actions. If child takes an object from mother and moves on without Continuation of S-JE: mother may manipulate a looking at mother, do not code anything. If takes child take toy, then sit back and watch whilst child continues the object and mother comments or child uses toy whilst to play with toy, then mother moves back in. mother actively watches, then code JE state from when child took toy. Higher level: child may appear energised; Lower level: child may appear bored by the shared activity
103
Behaviour Decision Definition Piloting notes Coordinated Include Child is coordinating attention from object or Mostly evident through glances at mother, also asking Joint event to the caregiver. questions, or for mother to do something, or verbally Engagement acknowledging mother without looking at her e.g. “Your Notes from technical manual: (C-JE) turn”. Child acknowledges mother: commonly repeated glances to face. Mother acknowledgement If mother asks question and child verbally answers with includes laughing. Mother’s handling of object meaningful response but and does look at mother code as may be minimal as child is actively balancing C-JE. attention between object & exchange. High quality: shared affect e.g. smiling at each other; Low quality: child may glance at mother due to command for visual attention. Fluency and Include Characterises the flow of conversation. Anchors Assigned descriptions of typical interaction at each rating connectedness for the two end points and the mid-point: to add to the descriptions in original coding scheme. (FC) 1: No conversation is established 1: Child off screen; Mother commenting to RA about child 4: Conversation lacks smoothness, appears to be 2: Some interaction, mother observer; child follows own largely dominated by one partner agenda; frequent not responding; child can initiate with mother responding 7: Fluid and balanced conversation that is often 3: Negotiate turns; child initiating at times with mother sustained responding 4: Evidence of at least 2 intervals of turn-taking; child initiates 3+ times; mostly on the same topic. 5: Turn-taking closer to 50/50; or 2+ striking exchanges; almost always on the same topic. 6: Turns flow beyond child showing/giving and mother acknowledging - must be smoother. Turns need to build from the last turn; mother and child both extend play. Can’t go off-screen to be counted as 6 or 7
104
4.7 Reliability
The final Maternal Behaviours Coding Scheme One is detailed in Table 7. The final Interactive Behaviours Coding Scheme Two is detailed in Table 8. Reliability was calculated on 10% of the sample. Good reliability increases confidence that the behaviours in the coding scheme could be generalised to other samples, and applied by other researchers to other studies (Levickis et al., 2014). Intra-rater reliability checks for the degree of agreement among repeated coding of the same observation made at two different times by the same coder. Inter-rater reliability checks for agreement between two different coders who code the same participants, blind to each other’s selections.
The codes from the two different coding events are compared within Observer ®XT using their reliability calculation function. This compares the codes and their time stamps between the two events, allowing a small window for reaction time variation (in this case three seconds). The codes from both events can then be viewed side-by-side, in the order they were coded, with all disagreements highlighted by Observer ®XT.
Overall percent agreement for intra and inter-rater reliability was calculated by
Observer ®XT, and percent agreement for each maternal behaviour was calculated by dividing the number of agreements by the number of agreements plus the number of disagreements. Intra and inter-rater agreement was >80% for all behaviours. Both inter- and intra-rater reliability was quantified using intra-class correlation (ICC), with ICC ranging from 0.98 to 0.99 for Coding Scheme One (see Table 9).
105
Table 7 Final Maternal Behaviours Coding Scheme One Behaviour Definition Example Successful and Utterance that aims to re-direct child away from the object or activity C: Playing with truck, M: “Look in the bag!” Unsuccessful Directive on which they are currently visually focused to something new or different. This is a type of intrusive directive. (Akhtar et al., 1991; Successful: child looks in bag Shimpi & Huttenlocher, Successful: child shifts visual focus toward the goal. Unsuccessful: child does not change focus 2007) Unsuccessful: child does not shift visual focus. Praise Verbal or non-verbal positive evaluations of child’s efforts, global “Good job!” (Gaertner et al., 2008; characteristics, or generic positive evaluations. “You’re right!” Winsler et al., 1999) Missed opportunity Mother does not respond in any way to a child’s utterance after 3 C: “Twuck”, M: No response secs Fillers Low level input to child which counts as mother’s turn. Must be M: “What’s he doing?” C: No response standalone – no other utterance from mother within 3 secs. M: hmm? Or M: “Uh huh” Emotion and mental state Utterances about desire, emotions, mental states, other internal “Baby needs her bottle” talk (EMST) states, or empathy statements. “Do you want/ like juice?” (Drummond et al., 2014)
106
Table 8 Final Interactive Behaviours Coding Scheme Two
Behaviour Definition (modified from Adamson et al. (2012) and Hirsh-Pasek et al. (2015))
Joint engagement (JE) Caregiver and child are actively attending to the same object or event with sustained interest for a minimum of 3 seconds. Active attending is usually evidenced by manipulation of the object, or being actively focused on the shared activity while one partner manipulates the object. Defined as either supported or coordinated (see below). Supported (S-JE) Child is focusing almost exclusively on the shared object or event (but beyond just listening). Caregiver’s participation influences the child’s activity/experience with object/event but the child does not acknowledge this involvement. The caregiver is often actively manipulating object or making statements/commands that alter child’s actions. Coordinated (C-JE) Child is coordinating attention from object or event to the caregiver. Child is acknowledging the caregiver, often with repeated glances to their face. The caregiver’s level of activity directly on the object may be minimal. Rating scale Describes both the quantity and quality of the child’s time in joint engagement. Note down: seconds in each engagement state, and tally to produce total estimated time in S-JE and in C-JE; episodes of matched affect, energy and intensity/concentration; After anchoring the interaction of the scale based on duration, adjust up a scale point if high quality based on notes of matched affect etc. Level 1= 2= 3= 4= 5= 6= 7= Approx. duration (secs) 0 15 30 90 120 150 >200secs Rating definition No episodes 1/3 time in moderate Frequently in of JE quality/briefly in rich and varied highly striking JE episodes. manner
107
Behaviour Definition (modified from Adamson et al. (2012) and Hirsh-Pasek et al. (2015))
Fluency and Characterises the flow of the conversation. Connectedness (FC) Rating scale Observe all verbal and non-verbal turns.
Level 1= 2= 3= 4= 5= 6= 7=
Rating definition No Interaction Instances of Conversation Shared topic Extension of Fluid and conversation established but child lacks throughout. interaction and balanced established. child initiations smoothness, Both partners play by both conversation frequently not reciprocated appears to be engaged in partners. that is often responding. by mother. largely relatively sustained. dominated by equal turn- one partner. taking.
108
Despite attempts at re-coding, and conducting intra-rater reliability checks again, the percent agreement for fillers continued to be below 80%. This, in combination with low inter-rater agreement, resulted in the decision to discontinue coding fillers.
Interactive Behaviours Coding Scheme Two reliability was determined using weighted and unweighted kappas. The kappa coefficient provides a measure of agreement over and above that which would occur by chance. A perfect agreement results is a kappa co-efficient of 1.0, no agreement results in a kappa of 0, and worse than chance agreements are indicated by negative coefficients (Landis & Koch, 1977).
Weighted kappas were also calculated, whereby parameterisation is the choice of the user (Cohen, 1968). In this instance, agreement was considered 1.0 if ratings were within one point of each other on the seven point rating scale as considered acceptable by the developers of the scale (Adamson et al., 2012; also see Bakeman & Quera,
2011). All weighted kappas were 1.0 for Coding Scheme Two.
Table 9 Reliability Ratings for Coding Schemes a Reliability P Behaviour Inter-rater Intra-rater b Successful directives (rpm P )P 0.99 0.99 Unsuccessful directives (rpm) 0.98 1.0 Missed-opportunities (rpm) 0.99 0.99 Praise (rpm) 1.0 1.0 c Fluency and Connectedness rating (FC) (rating scale P )P 0.64 0.86 c Supported Joint Engagement (S-JE) (rating scale P )P 0.47 0.77 c Coordinated Joint Engagement (C-JE) (rating scale P )P 0.66 0.74
a Note. P P Intra-class correlation random effects for maternal behaviours; Cohen’s kappa’s b c (Cohen, 1960) for S-JE, C-JE and FC; P P rpm = rate per minute; P RatedP on a 7-point scale
109
4.8 Measures
This research was embedded within two larger community-based studies (LLL and L4L) and built upon the MRBS, so the methods used were restricted by the pre- existing content of these studies. Aside from the two coding schemes described above, the author needed to choose within the variables which were available, rather than planning de-novo which measures would be used. The main measures used at each stage of data collection can be seen in Table 10. That is, in the original LLL study
(between 12 and 36 months), the embedded MRBS (at 24 months), and the follow-up
L4L study (at 48 months). The following section summarises key reliability and validity figures for the measures to illustrate their high quality, and justify the decision to proceed with including them in this thesis.
110
Table 10 Outline of Measures used in MIBS by Age and Study in which Collected Measure Age in months – Main Study 12 - LLL 18 - LLL 24 - LLL 36 - LLL 48 - L4L Demographics Questionnaire Language SSLM PLS-4 PLS-4 CELF-P2 SEB Difficulties CBCL CBCL SDQ Mother-child Interaction 15 mins free-play a Maternal Responsive Behaviours P Expansions, Imitations, Labels Interpretations, Supportive Directives, Responsive Questions b Maternal Behaviours P Successful and, Unsuccessful Directives, Praise, Missed opportunities, Emotion and mental state talk b Interactive Behaviours P Supported Joint Engagement, Coordinated Joint Engagement, Fluency and Connectedness
Notes: LLL: Let’s Learn Language Study; L4L: Language for Learning Study; SSLM: Sure Start Language Measure; PLS-4: Preschool Language Scale 4th Edition; CELF-P2: Clinical Evaluation of Language Fundamentals 2nd edition; CBCL: Child Behaviour Checklist; SDQ: Strengths and Difficulties a Questionnaire; EMST: Emotion and Mental State Talk; JE: Joint engagement; FC: Fluency and Connectedness. P BehavioursP coded in Maternal Responsive b Behaviours Study (MRBS); P BehavioursP coded in Maternal and Interactive Behaviours Study (MIBS)
111
4.8.1 Language Measures
Participants’ expressive vocabulary skills were measured at 18 months using the
Sure Start Language Measure (SSLM) (Roy et al., 2005). At 24 and 36 months the
Preschool Language Scale 4th Edition (PLS-4) (Zimmerman et al., 2002) was administered in the home, and at 48 months the Clinical Evaluation of Language
Fundamentals Preschool Second Edition (CELF-P2) (Semel et al., 2006) was used.
Both expressive and receptive language abilities were assessed because children can experience difficulties in either or both domains (Rhoades, 2003). It is also possible that these difficulties may be differentially associated with later language abilities
(Watt, Wetherby, & Shumway, 2006), and with SEB difficulties (Beitchman, Brownlie, et al., 1996). Furthermore, the association between language and SEB development may not be detected if a total language score was utilised. For example, receptive language difficulties may be indicative of broader underlying cognitive difficulties and therefore be more likely to be related to SEB adjustment (Toppelberg & Shapiro, 2000).
4.8.1.1 Sure Start Language Measure
The SSLM is a checklist of 100 words that parents complete to indicate which words their child is saying, and a single question asking whether the child combines words yet (see Appendix C). It is based on the original American short form of the
MacArthur-Bates Communication Development Inventory (CDI) (Fenson et al., 1994).
This was adapted to the UK context by altering ten words (Dale, Dionne, Eley, &
Plomin, 2000; Roy et al., 2005), and again to the Australian context by changing two further words (with the authors permission): ‘cream cracker’ became ‘cracker’, and
112
settee/sofa became settee/sofa/couch. The SSLM results in a cumulative, continuous raw score from 0 to 100 from which age and gender-based percentiles are derived.
The CDI has high reliability and validity (Can, Ginsburg-Block, Golinkoff, &
Hirsh-Pasek, 2013; Corkum & Dunham, 1996; Fenson et al., 2000). Separate validation checks and standardisation of the SSLM within the UK revealed it also has high reliability and validity (Feldman et al., 2005; Klee et al., 1998; Roy et al., 2005). A standardisation study containing 1303 participants in the UK (Roy et al., 2005) reported intra class correlation for total vocabulary scores of 1.0 and test-retest reliability of 0.99.
4.8.1.2 Preschool Language Scale 4th Edition
The PLS-4 is a directly-assessed measure for children from birth to age six years. It yields norm-referenced scores (mean (M) = 100, standard deviation (SD) = 15) for its two subscales, auditory comprehension (receptive language) and expressive communication (expressive language). The auditory comprehension subscale includes items tapping into attention, understanding of gestures, vocabulary, concepts, grammatical markers, syntactic structures, narratives, and phonological awareness. The
Expressive Communication subscale includes items assessing a child’s vocal development; use of gestures, ability to name objects, express concepts (e.g. descriptors, quantity), use prepositions, grammatical markers, and syntactic structures appropriately; and narrative production as well as phonological awareness (as described by Hurwitz &
Watson, 2016).
The PLS-4 has been shown to have acceptable reliability and validity in its standardisation study containing 2,400 US children (Zimmerman et al., 2002). The
Cronbach alpha coefficients for internal consistency ranged from 0.66 to 0.96, the
113
correlation coefficients for test-retest reliability for the expressive and receptive subscales ranged from 0.82 to 0.95, and the percent agreement for inter-rater reliability was 99%.
4.8.1.3 Clinical Evaluation of Language Fundamentals – Preschool Edition
The CELF-P2 is also a directly-assessed measure for children aged three to 6;11 years. It yields norm-referenced scores (M = 100, SD = 15) for receptive language
(sentence structure, concepts and following directions, and basic concepts subtests) and expressive language (word structure, expressive vocabulary and recalling sentences subtests). Test-retest reliability figures from a sample of 120 children are available for the US edition (Semel et al., 2006). Subtest correlation coefficients for the age range four to 4;11 years ranged from 0.79 to 0.95. Eadie et al. (2014) concluded that the measure has good to adequate reliability indexed by the reported standard error of measurement values that estimate the amount of error to be expected in test scores.
4.8.2 Social, Emotional and Behavioural Measures
Parents completed SEB measures as part of the questionnaires at ages 24, 36 and
48 months. The Child Behaviour Checklist for Ages 1.5-5 (CBCL) was used at 24 and
36 months (Achenbach & Rescorla, 2000). The Strengths and Difficulties Questionnaire
(two to four years) (SDQ) (Goodman, 2001) was used at 48 months.
4.8.2.1 Child Behaviour Checklist
The CBCL preschool form covers the age range 18 months to five years. It obtains parents’ ratings on 99 emotional, behavioural and social problems using a three- point scale (Not True (as far as you are aware); Somewhat or Sometimes True; Very
True or Often True), and an open-ended problem item. Summed ratings yield a total 114
score and seven syndrome scales (emotionally reactive, anxious/depressed, somatic complaints, withdrawn, sleep problems, attention problems, and aggressive behaviours)
(Achenbach & Rescorla, 2000). Two subscale scores are calculated from the syndrome scores: internalising and externalising behaviour problems. The internalising problem score is derived from the emotionally reactive, anxious/depressed, somatic complaints and withdrawn syndromes (possible range: 0-72). The externalising problem score is derived from the Aggressive behaviour and Attention problems scores (possible range:
0-48). Use of these summary internalising and externalising scores are recommended for use with ‘low risk’ community samples to most accurately discriminate clinical
disorder (http://www.aseba.org/preschool.html41TU ).U41T CBCL raw scores were used in all the analyses for this thesis rather than T-scores because the wide age-range of the norming sample (one to five years) might have reduced sensitivity to variation in the yearly age- bands (Bayer, Hiscock, Ukoumunne, Price, & Wake, 2008). Test-retest reliability scores were calculated from ratings of 1,728 pre-school aged children in the US (Achenbach &
Rescorla, 2000). Cronbach’s alpha coefficients for total score was 0.95, and for domains ranged from 0.66 to 0.92. Inter-rater reliability for parents was 0.61 (as reported in
Pontoppidan, Niss, Pejtersen, Julian, & Vaever, 2017). Clinical use of the CBCL 1.5-5 to assess child functioning in the seven syndrome areas has been supported by an international study testing 19,106 children aged 18 to 71 months in 23 societies in Asia,
Australasia, Europe, the Middle East and South America (Ivanova et al., 2010).
4.8.2.2 Strengths and Difficulties Questionnaire
The SDQ is a widely used brief behavioural questionnaire which screens for positive and negative psychological attributes in children aged three to 16 years. It was derived from the longer Rutter Questionnaire (Rutter, 1967) plus items on
115
concentration, peer relations and social competence based on DSM-IV criteria for mental disorders. Parents rate 25 statements covering four difficulty domains:
Hyperactivity/Inattention, Conduct Problems, Peer Problems, and Emotional
Symptoms; and a strength domain: Prosocial Behaviour. Each subscale has five items that informants rate on a three-point scale (Not true; Somewhat true; Certainly true).
The four problem domain scores are summed to give a Total Difficulties Score which indicates the severity and content of the difficulties (Stone, Otten, Engels, Vermulst, &
Janssens, 2010). The Hyperactivity/Inattention and Conduct Problem subscales are summed to give an Externalising Behaviour score (possible range 0-20), and the Peer
Problems and Emotional Symptoms subscales are summed to give an Internalising
Behaviour score (possible range 0-20). Higher scores indicate more problem behaviours. The summary scores for internalising and externalising behaviours were used in this thesis, modelled as continuous variables. For detailed scoring information
see http://www.sdqinfo.com41T .41T
The SDQ has been shown to have strong psychometric properties in a review of
48 studies (n = 131,223) (Stone et al., 2010). Internal consistency was high for total difficulties (r = 0.80), and moderate for the other subscales (range r = 0.53 to 0.76).
Test-retest reliability correlation coefficients for the Total Difficulty score were .76 and ranged from 0.65 to 0.71 for the problem behaviour and prosocial subscales. This indicates these subscales tended to be less reliable over time than the summary score.
Moderate to strong internal reliability (coefficient alphas range 0.59 to 0.80) has been reported across all SDQ subscales. Regarding concurrent validity, the SDQ had a moderate to high correlation with other measures of general psychopathology, most
116
relevant to this thesis being the CBCL (r = 0.76 between SDQ total difficulties and
CBCL total score).
4.8.3 Demographic Information Collected in the Study
Information on a range of demographic information was collected in the 12 month questionnaire (Appendix B). Co-variates to be included in the analyses as potential confounders were identified from these variables collected as part of LLL.
Selection was based on the literature describing risk and protective factors of either language and/or SEB difficulties. There are numerous child, family and environmental factors hypothesised to contribute to child language and SEB development (e.g. Clegg et al., 2015; Morgan et al., 2015; Zubrick et al., 2007). By adjusting for these factors, their potential for confounding the statistical estimates of associations between the maternal and interactive behaviours and language or SEB outcomes was reduced, and confidence in the associations existing independently of other factors increased (see
Kirkwood & Sterne, 2013 for a discussion on confounding).
Biological co-variates were: child gender, birth-order (first or later born), gestation (born at or over 37 weeks gestation, or earlier than 37 weeks), and birth weight (in grams, as a proxy measure of biological risk). Male gender has repeatedly been shown to be a risk factor for DLD (H. D. Nelson et al., 2006; Reilly et al., 2010) and for externalising behaviour difficulties (Bayer et al., 2012; Dubois-Comtois, Moss,
Cyr, & Pascuzzo, 2013). There is some evidence that first born children receive more verbal input from their parents than later born children and have greater language abilities (Hoff-Ginsberg, 1998), and that having no older siblings may be a risk factor for internalising behaviour problems up to age five (Bayer et al., 2012). Pregnancy and
117
birth difficulties may contribute to behavioural and cognitive difficulties. (e.g. Hack,
Klein, & Taylor, 1995), and associations have been reported between prematurity and low birth weight and child language skills (Stolt, Haataja, Lapinleimu, & Lehtonen,
2009) and parenting styles (Treyvaud et al., 2009). It is also feasible that mother-child interactions might differ in quality and quantity of maternal verbal behaviours, depending on whether the mother is playing with a boy or girl, or with a child who faced medical risk in infancy (e.g. low birthweight) (Treyvaud et al., 2009).
Environmental covariates were: SES, parental education, maternal age, household employment (at least one parent employed), and non-English-speaking background
(defined as a language other than English spoken in the home for more than 10 hours a week). The analyses were adjusted for socio-economic status (SES) using the
‘Socioeconomic Indices for Area’ (SEIFA) Index of Relative Social Disadvantage
(IRSD) (ABS, 2001). This is a neighbourhood score of disadvantage which the
Australian Bureau of Statistics calculates using census data. SEIFA scores apply to a postal area based on information about household education, employment, occupation, income, and composition. The mean score for the Australian population is 1000 and the standard deviation is 100. Lower scores indicate greater social disadvantage, which might result in less cognitively stimulating and more stressful environments that constrain child development (McLoyd, 1998). Low SES has consistently been shown to be associated with poorer language outcomes (H. D. Nelson et al., 2006; Pungello et al.,
2009; Reilly et al., 2010), poorer behaviour outcomes (Kiernan & Huerta, 2008), and lower maternal verbal input (Cartmill et al., 2013; Hart & Risley, 1995; Hoff, 2003;
Hoff & Naigles, 2002).
118
Finally, because maternal education and age are linked to parenting style
(Warren & Brady, 2007), language outcomes (Hoff, 2003; H. D. Nelson et al., 2006), and SEB outcomes (Anselmi, Piccinini, Barros, & Lopes, 2004; van Zeijl et al., 2006), they were included as co-variates. Maternal education was included as a categorical variable (did not complete high school / completed high school / completed tertiary/postgraduate degree or equivalent). Maternal age was included as age in years when the child was 12 months old.
4.9 Analysis
The statistical techniques used in this thesis were regression analyses. In most analyses, language and SEB scores were examined as continuous variables, standardised to z-scores (M = 0, SD = 1) to facilitate comparability between the different measures. Subsequently, most of the regression coefficients reported in the results chapters represent effect sizes, Cohen's d, which reflect the standardised difference between two means (Cohen, 1992).
The analyses using maternal behaviours as predictors used rate per minute as the unit of analysis rather than total frequency. This controlled for differences in recording durations. S-JE, C-JE and FC were modelled as continuous variables, from one to seven.
Children’s performance on the key measures of language and SEB skills was examined by plotting histograms to observe the variability and distribution of the data.
Next, to determine how many participants had low language scores at each age, cut- points commonly used in research (scoring ≥ 1.25 SD below the mean for standard
119
scores on expressive and/or receptive language e.g. Reilly et al. 2010) were applied to the language scores. For SEB scores, it was not possible to calculate the proportion scoring in the borderline or clinical range at 24 and 36 months old. This was because the
CBCL T-scores from which the borderline and clinical cut points were calculated were generated from a sample comprising a wide age range (from one to five years).
Converting the CBCL raw scores to T-scores in our data could have resulted in age- sensitive changes being concealed. Therefore, we retained the CBCL raw scores.
However, we did apply the SDQ borderline cut-points to the 48 month old SDQ results to give an indication of difficulties in our sample. Cut points for the version of the SDQ used in our sample (for ages 18 months to four years) were not available. Therefore, we used cut-points for the version of the SDQ for children aged four years and older to provide an indication of those experiencing difficulties at age four years in this sample.
Prior to using the regression techniques, tests for the assumption of linearity were conducted. These used likelihood ratio tests to compare regression models fitting the predictors as continuous and categorical (i.e. allowing for non-linearity of effects) to the nested model fitting a continuous term only (i.e. assuming a linear trend across the categories). Evidence of non-linearity was provided if these p-values were <0.05.
The analyses for some sections of this PhD involved multiple comparisons.
Therefore, all models were interpreted with caution and by looking for consistent patterns. This approach was chosen in preference to using formal correction such as the
Bonferroni method as this is too conservative, increasing the likelihood of type II errors
(i.e. rejecting true positive associations because the p-value is too stringent) (Perneger,
1998). Analyses were conducted with Stata 13.0 (StataCorp LP, 2013).
120
4.9.1 Confounders
As discussed in section 4.8.3 potential confounders were grouped into biological and environmental exposures. For Aim One (the association between language and SEB difficulties), child gender, gestation, birth weight, birth order, SEIFA index of relative social disadvantage, maternal education, household paid employment, non-English speaking background, and maternal age were included, as well as RCT treatment group
(intervention or control arm).
For Aims Two and Three, a larger number of comparisons were undertaken. The data were not pooled across ages as they were in Aim One, where both the predictor
(language score) and outcome (SEB score) were measured at each age. Therefore, to decrease the number of co-variates in each analysis for Aims Two and Three, key co- variates were selected for inclusion in the adjusted regression models. Household employment was excluded because there were only 2/197 participants living in households without either parent in paid employment. Gestation was excluded in favour of retaining birth weight as a measure of biological risk. Finally, non-English speaking background was excluded because the majority of participants did not live in a household where a language other than English was spoken more than 10 hours a week
(178/197).
The analyses are described in more detail in Chapters Five to Seven where the analysis sections are included in the accepted or under peer-review manuscripts.
However, a brief overview is given below for each of the aims.
121
4.9.2 Associations between Language and SEB Outcomes
Aim One was concerned with examining the association between language and
SEB scores in a sample of slow-to-talk toddlers. The language and SEB scores were converted to z-scores for comparability purposes because both measures changed at 48 months to different but comparable tools (see section 4.8). Initially, four linear regression models were fitted to the language and SEB measures at 24, 36 and 48 months separately. This was to address the question of whether receptive or expressive language scores were associated with internalising or externalising problem scores at
24, 36 and 48 months old. However, this resulted in 12 unadjusted statistical comparisons, and a further 12 fully-adjusted models. Since a high number of comparisons can increase the chance of spurious findings, the data were reshaped to represent repeated measures over time longitudinally. This means that rather than organising the data into one row per participant with three sets of outcomes, the data were organised into three rows per participant with one set of outcomes per row and an indicator variable specifying which age the outcomes refer to (see Figure 3).
The data were then analysed using repeated measures regression, with up to three repeated observations per participant, one for each of the ages (see Vittinghoff,
McCulloch, Glidden, & Shiboski, 2008). This reduced the number of comparisons and increased the precision of the estimates through increased observations per comparison.
To account for the within-participant clustering, and non-independence of the repeated measures within participants, robust standard errors were used (using the vce (cluster clustvar) option in Stata) (Huber, 1967; White, 1980). The analyses were adjusted for potential confounders outlined in section 4.9.1.
122
Figure 3 Transformation of Data from Wide-form to Long-form Wide form ID Language SEB Language SEB Language SEB score 24m score 24m score 36m score 36m score 48m score 48m 0011 81 5 87 6 82 5 0012 100 3 108 3 102 4
Long form ID Language SEB Score Indicator Score 1 = 24m 2 = 36m 3 = 48m 0011 81 5 1 0011 87 6 2 0011 82 5 3 0012 100 3 1 0012 108 3 2 0012 102 4 3
To explore the form/shape of the associations, Wald tests were used to compare regression models containing linear, quadratic and cubic terms to determine which model fitted the language and SEB data best. Finally, an indicator variable for age was added as an interaction term to test whether the models differed by child age. Wald tests compared findings from the models including and excluding the interaction terms.
These results are presented in Chapter Five.
4.9.3 Maternal and Interactive Behaviours and Language Outcomes
Aim Two was concerned with examining how maternal and interactive behaviours were associated with language outcomes. Univariate and multivariate linear 123
regression models were fitted to determine whether there was an association between the maternal and interactive communicative behaviours and language scores at 24, 36 and 48 months old. The analyses were adjusted for potential confounders (child gender, birth order, birth weight, SEIFA index of relative social disadvantage, maternal education, maternal age, intervention group). The analyses of language scores at 36 and
48 months were additionally adjusted for 24 month old language scores to determine whether the child’s language skills at the time of the recording explained the variability in later language scores over and above the maternal and interactive behaviours
(Adamson et al., 2009).
Pearson’s correlation coefficients were calculated to examine the associations between all maternal behaviours coded as part of this thesis (i.e. praise, emotion and mental state talk, missed opportunities, fillers, successful directives and unsuccessful directives), and those coded in the MRBS and found to be associated with better language scores (i.e. expansions, imitations, responsive questions). One intrusive behaviour and one responsive behaviour, which were weakly correlated with each other, were selected to include in a model together to understand how co-occurring maternal behaviours were associated with concurrent language outcomes.
To examine whether the association between maternal behaviours and language outcomes was modified by Fluency and Connectedness or Joint Engagement state, interaction analysis was used. An interaction term for ‘Fluency and Connectedness rating by maternal behaviour’ was added to the cross-sectional regression model of interest. Interaction for ‘Joint Engagement by maternal behaviour’ was similarly examined. Likelihood ratio tests compared models including and excluding the interaction term. Further interaction analysis was undertaken using simple slopes 124
(UCLA, 2015, Statistical Consulting Group) in order to examine interaction effects that were not evident by the standard interaction analysis (Kirkwood & Sterne, 2013). These analyses fitted the predicted regression line for the association between the maternal behaviour and language outcome at each level of Joint Engagement or Fluency and
Connectedness. These results are presented in Chapters Six and Seven.
4.9.4 Maternal and Interactive Behaviours and SEB Outcomes
Aim Three was concerned with whether there was an association between maternal and interactive behaviours and internalising and externalising behaviour scores. As for Aim Two, this aim was addressed using univariate and multivariate linear regression analysis. The analyses were adjusted for potential confounders (child gender, birth order, birth weight, SEIFA index of relative social disadvantage, maternal education, maternal age, intervention group). No further analysis took place for the internalising and externalising behaviour scores following the null findings from this step. These results are presented in Chapter Eight.
4.10 Summary
This Chapter has described the methodology used to conduct this PhD study.
Measures of language skills, SEB difficulties, and demographic information were used from two existing population-based studies, Let’s Learn Language and Language for
Learning. Video recordings of mothers and their toddlers were used from the Maternal
Responsive Behaviours Study. Two observational coding schemes were developed.
Coding Scheme One was used to code maternal behaviours (praise, emotion and mental state talk, missed opportunities, fillers, successful directives, and unsuccessful
125
directives). Coding Scheme Two was used to code interactive behaviours (Supported
Joint Engagement, Coordinated Joint Engagement, and Fluency and Connectedness).
Together, these data were used to address three aims; (1) to examine whether language and SEB difficulties were associated at 24, 36- and 48-months; (2) to examine to what extent maternal and interactive behaviours were associated with language scores at these ages; and (3) to examine to what extent maternal and interactive behaviours were associated with SEB outcomes at these ages.
Chapter Five presents the findings relating to Aim One in the form of a manuscript accepted for publication in the International Journal of Child Language and
Communication Disorders and entitled, ‘The association between receptive and expressive language and internalising and externalising behaviours: a community based sample of slow-to-talk toddlers’. Chapters Six and Seven present the findings relating to
Aim Two. Chapter Six considers praise, missed opportunities, successful and unsuccessful directives, and Fluency and Connectedness, and is in the form of a manuscript also accepted for publication in the International Journal of Child Language and Communication Disorders. It is entitled, ‘Maternal communicative behaviours and interaction quality as predictors of language development: Findings from a community- based study of slow-to-talk children’. A supplementary section of Chapter Six considers emotion and mental state talk. Chapter Seven comprises a manuscript considering
Coordinated and Supported Joint Engagement, which has been invited to undergo a third-round of reviews by the Journal of Child Language. Chapter Eight presents the findings for Aim Three.
126
5 Language and Social, Emotional and
Behavioural Development
5.1 Overview
This chapter addressed Aim One of the thesis, to explore the association between language skills and social, emotional and behavioural difficulties. The Chapter was accepted for publication in a peer-reviewed journal in March 2017, the
International Journal of Language and Communication Disorders. The pdf version of the paper is included below.
The citation is:
Conway LJ, Levickis P, Mensah F, McKean C, Smith KA, and Reilly S (2017).
Associations between expressive and receptive language and internalising and externalising behaviours: Findings from a community-based slow-to-talk sample.
International Journal of Language and Communication Disorders, 52 (6): 839-853. doi:
10.1111/1460-6984.12320
127
5.2 PDF of Manuscript
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
Supplementary Tables
Supplementary Table 1: Expressive Language and Internalising Behaviour: Regression Models Fitted
Model Language modelled as: β [95% CI] p Model Language modelled as: β [95% CI] p
1. Linear -.11 [-.22, -.01] .03 4. Linear -.13 [-.28, -.02] .10
2.a Linear -.10 [-.28, .08] .27 Cubic .005 [-.02, .03] .67
Categorical_2 -.15 [-.46, .16] .34 5. Linear -.13 [-.28, .03] .10
Categorical_3 -.19 [-.58, .21] .35 Quadratic .03 [-.01, .08] .14
Categorical_4 -.02 [-.57, .52] .94 Cubic .004 [-.02, .03] .74
Categorical_constant .06 [-.25, .38] .70 6. Quadratic N/A
3. Linear -.12 [-.22, -.01] .03 Cubic N/A
Quadratic .03 [-.01, .08] .13
Models: 1. Linear, 2. Linear & non-linear, 3.Linear & quadratic, 4. Linear & cubic, 5. Linear, quadratic & cubic, 6. Quadratic & cubic; a Wald – Linearity test F(3, 199) = 1.06, p=.37
Supplementary Table 2: Expressive Language and Externalising Behaviours: Regression Models Fitted
Model Language modelled as: β [95% CI] p Model Language modelled as: β [95% CI] p
1. Linear -.02 [-.14, .10] .74 4. Linear -.04 [-.11, .19] .63
2.a Linear -.17 [-.39, .05] .14 Cubic -.02 [-.05, .01] .26
Categorical_2 .12 [-.16, .40] .39 5. Linear .04 [-.11, .18] .63
Categorical_3 .10 [-.25, .46] .56 Quadratic .05 [-.01, .11] .08
Categorical_4 .47 [-.05, .99] .08 Cubic -.02 [-.04, .004] .11
Categorical_constant -.20 [-.47, .08] .16 6. Quadratic N/A
3. Linear -.02 [-.14, .09] .68 Cubic N/A
Quadratic .05 [-.01, .11] .12
Models: 1. Linear, 2. Linear & non-linear, 3.Linear & quadratic, 4. Linear & cubic, 5. Linear, quadratic & cubic, 6. Quadratic & cubic; aWald – Linearity test F(3, 199) = 1.99, p=.12
144
Supplementary Table 3: Receptive Language and Internalising Behaviours: Regression Models Fitted
Model Language modelled as: β [95% CI] p Model Language modelled as: β [95% CI] p
1. Linear -.19 [-.31, -.08] .001 4. Linear -.15 [-.31, .001] .051
2.a Linear -.16 [-.43, .11] .24 Cubic -.01 [-.07, .04] .63
Categorical_2 -.07 [-.35, .22] .65 5. Linear -.17 [-.32, -.01] .03
Categorical_3 -.08 [-.49, .34] .72 Quadratic .04 [-.03, .11] .25
Categorical_4 -.10 [-.70, .51] .75 Cubic -.01 [-.06, .04] .76
Categorical_constant -.04 [-.25, .32] .81 6. Quadratic N/A
3. Linear -.19 [-.29, -.08] .001 Cubic N/A
Quadratic .04 [-.04, .13] .27
Models: 1. Linear, 2. Linear & non-linear, 3.Linear & quadratic, 4. Linear & cubic, 5. Linear, quadratic & cubic, 6. Quadratic & cubic; a Wald – Linearity test F(3, 199) = .07, p=.98
Supplementary Table 4: Receptive language and Externalising Behaviours: Regression Models Fitted
Model Language modelled as: β [95% CI] p Model Language modelled as: β [95% CI] p
1. Linear -.14 [-.26, -.02] .02 4. Linear -.01 [-.16, .14] .90
2.a Linear -.29 [-.56, -.01] .04 Cubic -.05 [-.09, -.01] .02
Categorical_2 .02 [-.30, .34] .90 5. Linear -.03 [-.18, .12] .67
Categorical_3 .09 [-.38, .56] .71 Quadratic .07 [.003, .14 ] .04
Categorical_4 .44 [-.23, 1.1] .20 Cubic -.04 [-.07, .0003] .05
Categorical_constant -.16 [-.51, .19] .38 6. Quadratic .07 [.002, .14] .04
3. Linear -.13 [-.24, -.02] .02 Cubic -.04 [-.07, -.02] .002
Quadratic .09 [.01, .17] .03
Models: 1. Linear, 2. Linear & non-linear, 3.Linear & quadratic, 4. Linear & cubic, 5. Linear, quadratic & cubic, 6. Quadratic & cubic; aWald – Linearity test F(3, 199) = 1.75, p=.16
145
Supplementary Table 5: Variability in language and SEB associations by child age: Tests of interaction by wave Language SEB Combined Wald test across domain a adjustment b Interaction term β [95% CI] p interaction terms Expressive Internalising Wave2*Expressive -.03 [-.18, .12] .72 F (2, 199) = .08, p= .92
Wave3*Expressive -.005 [-.20, .19] .96
Expressive Externalising Wave2*Expressive -.12 [-.27, .03] .12 F (2, 199) = 1.36, p= .26
Wave3*Expressive -.11 [-.28, .07] .23
Receptive Internalising Wave2*Receptive -.03 [-.19, .13] .72 F (2, 199) = .19, p= .83
Wave3*Receptive .03 [-.16, .23] .72
Receptive Externalising Quadratic Wave2*Receptive -.09 [-.25, .08] .29 F (4, 199) = .93, p=.45
Wave3*Receptive -.05 [-.22, .13] .60
Cubic Wave2*Receptive -.03 [-.09, .04] .41
Wave3*Receptive -.03 [-.10, .04] .41
Note: Reference group is Wave 1, 24-months. Wave2 is 36-months; Wave3 is 48-months. CI = confidence interval, SEB = Social, emotional and behavioural adjustment score, R2 = percentage. a PLS-4 at 24- & 36-months, CELF-P2 at 48-months; b CBCL at 24- & 36-months, SDQ at 48-months
146
5.3 Summary
The findings presented in this Chapter have revealed that receptive language abilities were associated with internalising and externalising problem behaviours at 24,
36 and 48 months. Whilst previous studies using samples of children with diagnosed language or SEB difficulties, and some population-based studies had reported this association, crucial gaps existed in our understanding of the age the association emerges, and which specific difficulties are associated. It is now clear that language and
SEB difficulties are associated concurrently from as early as 24 months. Further, between 24 and 48 months, it is children with receptive language difficulties who are most at risk of SEB difficulties, whilst the association with expressive language is not yet evident. Since receptive language difficulties are more difficult to identify than expressive difficulties, recognition of their co-occurrence with externalising behaviour problems at these early ages among health and education professionals might help increase timely referral to services.
Future research using a typically developing sample will further improve our understanding of the non-linear association between receptive language and internalising behaviours. The wide confidence intervals around the estimates for high receptive language scores limited conclusions being drawn about the protective role of good receptive language skills. Although histograms revealed a normal distribution of language scores at 36 and 48 months, mean language scores were slightly lower than community norms. A typically developing sample might be expected to have a distribution slightly to the right of this current sample, with more participants scoring in the higher range. This would enable more precise estimates of the association at
147
higher receptive language scores. Equally, as stated in the paper, the association at lower receptive language scores might be greater in a typically developing sample than found here. This is because 18 month old children with receptive language difficulties and typical expressive vocabulary were not included in the current sample meaning children with receptive language difficulties may have been under-represented.
These findings also contribute to a broader understanding of the nature of the association between DLD and SEB difficulties. As discussed in the Literature Review
Chapter (section 2.5.1), one theory proposes that a broad underlying vulnerability may predispose some children to a range of developmental difficulties. The finding that poor receptive language abilities were associated with poor SEB difficulties as early as 24 months offer some support to this theory.
Whilst improving ways to identify children at risk of problems is essential, another piece of the puzzle is designing effective intervention strategies for these children. Given the co-occurrence of language and SEB difficulties, it would be worthwhile to examine whether effective interventions can be developed which promote both language development and reduce behaviour problems. Interventions for children with language difficulties and for children with SEB difficulties both commonly target parent-child interactions. The following Chapters examine whether aspects of these interactions are associated with language outcomes and with SEB outcomes. If commonalities are found, this could form an initial step in recognising parenting behaviours that might effectively promote language and SEB development.
148
6 Mother-child Interaction and Language
Outcomes: PART ONE
6.1 Overview
The following Chapter explores the associations between maternal and interactive behaviours and child language outcomes, thereby addressing aspects of Aim
Two. This paper has been accepted for publication by the International Journal of
Language and Communication Disorders. The paper examined the following behaviours from the maternal behaviours Coding Scheme One: maternal praise, missed opportunities, and successful and unsuccessful directives; and Fluency and
Connectedness (FC) from the interactive behaviours Coding Scheme Two.
The citation is:
Conway LJ, Levickis, P.A., Smith, J., Mensah, F., Wake, M., & Reilly S. Maternal communicative behaviours and interaction quality as predictors of language development: Findings from a community-based study of slow-to-talk toddlers.
International Journal of Language and Communication Disorders.
A final behaviour coded in Coding Scheme One was maternal emotion and mental state talk (EMST). This was not included in the publication because EMST has its own respective and growing literature, which although related to the literature considered in the paper, is not directly applicable. Hence, the findings for EMST are presented separately in the supplementary section below (section 6.3). 149
6.2 Author Accepted Manuscript
Title: Maternal communicative behaviours and interaction quality as predictors of language development: Findings from a community-based study of slow-to-talk children
Running head: Communicative behaviours and interaction quality
Keywords: Mother-child interaction, Language development, Parenting, Longitudinal
Study
150
Abstract
Background. Identifying risk and protective factors for language development informs interventions for children with developmental language disorder (DLD).
Maternal responsive and intrusive communicative behaviours are associated with language development. Mother-child interaction quality may influence how children use these behaviours in language-learning.
Aims. We aimed to identify (1) communicative behaviours and interaction quality associated with language outcomes, (2) whether the association between a maternal intrusive behaviour (directive) and child language scores changed alongside a maternal responsive behaviour (expansion), and (3) whether interaction quality modified these associations.
Methods & Procedures. Language skills were assessed at 24-, 36-, and 48- months in 197 community-recruited children who were slow-to-talk at 18-months.
Mothers and 24-month-olds were video-recorded playing at home. Maternal praise, missed-opportunities, and successful and unsuccessful directives (i.e. whether followed by child) were coded during a ten-minute segment. Interaction quality was rated using a seven-point Fluency and Connectedness (FC) scale, during a five-minute segment.
Linear regressions examined associations between these behaviours/rating and language scores. Interaction analysis and simple slopes explored effect modification by FC.
Outcomes & Results. There was no evidence that missed-opportunities or praise were associated with language scores. Higher rates of successful directives in the unadjusted model, and unsuccessful directives in the adjusted model were associated
151
with lower 24-month-old receptive language scores (e.g. unsuccessful directives effect size (ES) = -0.41). The association between unsuccessful directives and receptive language was weaker when adjusting for co-occurring expansions (ES = -0.34). Both types of directives were associated with poorer receptive and expressive language scores in adjusted models at 36- and 48-months (e.g. unsuccessful directive and 48- month receptive language, ES = -0.66). FC was positively associated with 24-, 36- and
48-month language scores in adjusted models (e.g. receptive language at 24-months, ES
= 0.21, at 48-months, ES = 0.18). Interaction analysis showed the negative association between successful directives and 24-month receptive language existed primarily in poorly-connected dyads with low FC levels.
Conclusions & Implications. These findings illustrated the effects of the combined interaction between different maternal communicative behaviours and features of the interaction itself on child language development, and the need to consider both in research and practice. Whilst more intrusive directives were associated with poorer language scores, this association attenuated when adjusting for co-occurring responsive expansions, and the association was strongest for children in lower quality interactions. This work may inform clinical practice, by helping clinicians target the most appropriate communicative behaviours for specific mother-child dyads.
152
What This Paper Adds
What is Already Known on this Subject?
Developmental language disorder (DLD) is a common childhood disorder, associated with educational, health and social difficulties.
Responsive and intrusive maternal communicative behaviours contribute to children’s language development, are modifiable, and therefore can be targeted in interventions.
The quality of the mother-child interaction forms the foundation on which language skills develop, and is co-constructed by the mother and the child.
What this Study Adds
Maternal directives were associated with poorer concurrent and later language in slow- to-talk children. Associations were not evident for missed-opportunities or praise.
Maternal expansions, known to facilitate language development, were protective against the negative association between directives and language outcomes.
The quality of mother-child interaction and child language abilities were positively associated.
The negative association between directives and language outcomes was primarily observed when mothers and children were poorly connected to each other.
Clinical Implications of this Study
The potential to assist clinicians in targeting the specific maternal communicative behaviours that may be most appropriate for specific mother-child dyads.
The importance of promoting high quality interactions as well as teaching specific communicative behaviours in interventions.
153
Introduction
Language proficiency is a foundation skill for optimal child development.
However, between 7% and 20% of children in the pre-school and early school years experience language difficulties (Norbury et al. 2016; Reilly et al. 2010).
Developmental Language Disorder (DLD) is associated with negative outcomes including poor literacy academic achievement, and psychosocial problems that can resonate into adulthood (e.g. Beitchman et al. 2014). Thus, considerable effort is being expended to understand how to help children at risk.
Language develops through a complex interplay between biology and the environment (Bishop et al. 2016). However, biological and environmental factors associated with early language outcomes (e.g. Morgan et al. 2015) only explain up to
20.9% of the variability in 48-month-old language outcomes (Reilly et al. 2010). This suggests additional factors must be at play. Parent-child interactions represent a rich area for investigation. Studies from the typical and atypical child development literature suggest early interactions are critical to shaping and enhancing language development
(Guralnick et al. 2008; Hart and Risley, 1995; Roberts and Kaiser, 2011; Rowe, 2012).
Investigations into the quantity and quality of parental input during parent-child interactions (Hart and Risley, 1995; Huttenlocher et al. 2010; Rowe, 2012) have supported the social interactionist perspective that children learn language during parent-child interactions (Bruner, 1983). Hearing fewer total and fewer diverse words leads to compromised expressive vocabulary at school-entry, which can have lasting impacts on children’s language, literacy and academic trajectories (Rowe, 2012).
154
Specific maternal and interactive communicative behaviours that have been investigated in relation to language learning will be discussed below.
Maternal Communicative Behaviours
Maternal communicative behaviours characterised as being responsive to the child, such as expansions and imitations, have been found to be associated with better language outcomes in population-based samples (e.g. Levickis et al. 2014; Masur et al.
2005). Subsequently, these are often the focus of parent-implemented language interventions (Roberts and Kaiser, 2011). Different aspects of maternal input matter at different ages and language abilities, so maternal use of a behaviour beyond an optimal age might indicate a mother-child dyad requiring help. Rowe (2012) revealed how the quantity of maternal words at 18-months, diverse and sophisticated vocabulary at 30- months, and decontextualized language at 48-months were associated with subsequent language skills.
Whilst it is generally accepted that maternal communicative behaviours influence child language, we will explore behaviours for which evidence of the nature of these associations is limited. This information could inform language interventions which target modification of parental behaviours. The behaviours considered were praise, missed-opportunities, and directives.
Praise
Maternal praise might contribute to a warm environment conducive to language- learning, yet to our knowledge, the associations with child language outcomes have not been investigated. A recent RCT of an intervention to improve language outcomes in 155
children with conduct problems (n = 60, aged 12 to 15 months) targeted parental behaviours, including praise (Bagner et al. 2016). Children receiving the intervention had better expressive language skills six-months post intervention compared with children receiving standard care. However, the study did not examine whether this improvement was mediated via improved parental behaviours, therefore we cannot conclude that parental praise was associated with better language scores.
Missed Opportunities
Since social feedback within interactions is important for early word-learning
(Bedford et al. 2013), children’s language development may be negatively impacted if their parents frequently miss opportunities to respond to them. Children may fail to see how their communication relates to parental responses, or how interactions comprise turn-taking sequences. Frequent missed-opportunities may ultimately dampen children’s motivation to interact. Reasons for missed opportunities include children with poor language skills using less salient communicative signals, and parents who are less engaged not noticing their children’s communication attempts.
Directives
Maternal directives have been studied for decades, but evidence for their association with language development remains inconclusive. Mothers of children at risk of DLD (e.g. with early expressive language delay (e.g. Paul and Shiffer, 1991); neurodevelopmental disorders (e.g. Blacher et al. 2013; Crawley and Spiker, 1983); and low birthweight, (e.g. Landry et al. 1997) are reportedly more directive than mothers of typically developing children. However, whether the directives contribute to the
156
language difficulties, or whether they are appropriate parental adaptations to the children’s characteristics is uncertain (Hudry et al. 2013; Marfo, 1990).
Directives have been found to be associated with poorer language outcomes (e.g.
Masur et al. 2005), better language outcomes (e.g. Akhtar et al. 1991), and to have no association with language (e.g. Tomasello and Todd, 1983). This is likely due to the heterogeneity in participant ages and definitions of directives used between studies, and the failure to consider the interactional context in which directives were used (Marfo,
1990). A renewed investigation is warranted to clarify their role.
Distinctions have been made between types of directives, specifically between
‘following’ or ‘supportive’ directives, and ‘leading’ or ‘intrusive’ directives (Akhtar et al. 1991). Supportive directives, which follow into the child’s focus of attention, were not found to be associated with language outcomes in this sample (Levickis et al.,
2014). However, intrusive directives, which lead children away from their focus of attention, have yet to be investigated in this sample. These might be associated with poorer language outcomes because they deplete children’s immature attention and cognitive systems and disrupt the flow of the interaction (Akhtar et al. 1991; Landry et al. 1997; Masur et al. 2005; Tomasello and Todd, 1983). Directives which children follow may be less detrimental to language learning than those not followed because children can map their parents’ words on to their new focus, as has been found with labels (Shimpi and Huttenlocher, 2007). Frequent unsuccessful directives might be indicative of a child with, or at risk of DLD, or a parent needing help to support their child.
157
The investigation into directives and language development could be further enriched by considering their association alongside co-occurring parental responsive behaviours. Attachment theorists argue that responsivity and directiveness are orthogonal characteristics, suggesting that directives should not affect the positive association between responsive behaviours and language or vice versa (Marfo, 1990).
This has received partial support from studies using global parenting ratings (Crawley and Spiker, 1983; Pungello et al. 2009), but has yet to be tested using specific parenting behaviours. If the association between directives and language outcomes is altered by responsive behaviours this will have implications for the content and focus of parent- implemented interventions.
Mother-Child Interactive Behaviours
The elements of parent-child interaction that are important to language development are much broader than maternal communicative behaviours alone.
Parenting is bidirectional, as parents and children respond to each other’s cues, signals and competencies (e.g. Guralnick et al. 2008). Bidirectionality is central to the transactional model of parenting, which explains that parent and child cannot be viewed as entirely independent of each other (Funamoto and Rinaldi, 2015; Sameroff, 2009).
Depending on their propensity and skills, parent and child may unconsciously create and maintain an environment either more or less conducive to language-learning
(Alston and James‐Roberts, 2005). Responsive parents may motivate children to engage in stimulating exchanges (Pungello et al. 2009), and responsive children may increase parental response opportunities. In contrast, intrusive parents might disrupt language- learning by distracting or inhibiting children, and children with ambiguous or infrequent communication might provide fewer response opportunities. In this way, parents and 158
children co-construct the foundation on which the children’s language skills are built
(Hirsh-Pasek et al. 2015).
Fluency and Connectedness
A more comprehensive investigation of how mother-child interactions are associated with language development would benefit from including a measure of the interactive dyad itself. Feasibly, such a measure might capture an aspect of the interaction missed by individual behaviours (Funamoto and Rinaldi, 2015). We selected
‘Fluency and Connectedness’ (FC), a measure of interaction quality that captures the flow and cohesion between mother and child, and how well they use verbal and non- verbal acts to stay on topic, orchestrate and sustain turn-taking (Adamson et al. 2012).
The benefit of using FC over a global parent-child interaction measure is that it has strong theoretical links to language development, rather than to general child development (Mahoney et al. 1996). Indeed, a study of low-income mothers and their two-year-olds (n = 60) found FC was more strongly associated with expressive language skills one year later than either the number of words spoken by the mother, or a global rating of parental sensitivity. This suggests FC captures aspects of language facilitation within mother-child interactions (Hirsh-Pasek et al. 2015).
We hypothesised that FC might be associated with language abilities, and may also modify the effect of maternal behaviours on language outcomes. This would have implications for maximising the impact of interventions, for example, by informing how best to weave maternal input “into the fabric of early caregiver-child interactions”
(Hirsh-Pasek et al. 2015, p. 1072).
159
Purpose of the Current Study
The current study used a community-derived sample, identified as slow-to-talk at 18-months-old, to investigate the association between language development and maternal missed-opportunities, praise, and directives, and mother-child interaction quality. This sample is of clinical interest due to its hypothesised risk of DLD.
Specifically, the study aimed to determine,
1. whether missed-opportunities, praise, successful and unsuccessful maternal
directives, and interaction quality (FC) at 24-months-old were associated with
expressive and receptive language scores at 24-, 36- and 48-months-old;
2. whether the association between maternal directives and language development
changed alongside a co-occurring maternal responsive behaviour; and
3. whether the cross-sectional associations between maternal communicative
behaviours and language scores were modified by interaction quality (FC).
Methods
Participants
This study is nested within Let’s Learn Language (LLL, a cluster randomized- controlled trial (NHMRC #384491) within a population-based survey (Wake et al.,
2011). Parents of 12-month-olds attending their well-child check-up in three of
Melbourne’s 31 local government areas were invited to participate. 1,217 completed a baseline questionnaire. Exclusion criteria were developmental delay, major medical condition, suspected autism spectrum disorder (ASD), and parents with insufficient
160
English to complete questionnaires at grade 6 (11-12 years-old) reading level. Ethics approval was received from the Royal Children’s Hospital Human Research and Ethics
Committee (#26028).
At 18-months-old, 93.5% (n = 1,138) completed a parent-reported expressive vocabulary screen (Roy et al. 2005). Children scoring ≤ 20th percentile based on the population norms (n = 301, 26.4%) were eligible for the trial of a low-intensity parent- toddler language promotion programme. At 48-months-old, participants were invited into a subsequent language intervention trial, Language for Learning (L4L) (NHMRC
#60740) (Goldfeld et al., 2012). Ethical approval was from the Royal Children’s
Hospital Human Research Ethics Committee (#30011). Because there were no evident differences in language outcomes between the intervention and control groups at ages
24- and 36-months, the sample was pooled for the current study (Wake et al., 2011).
Informed consent was received from 202 families for this study. Four children were excluded because they were diagnosed with ASD at 3-4 years-old, and one was excluded because the grandmother participated in the free-play session.
Measures
Language skills were assessed at home at 24-, 36- and 48-months-old. The
Preschool Language Scale (PLS-4) (Zimmerman et al. 2002), a directly-assessed language measure for children from birth to 6-years, was used at 24- and 36-months- old. It yields norm-referenced scores for auditory comprehension (receptive language) and expressive communication. The Clinical Evaluation of Language Fundamentals
Preschool Edition (CELF-P2) (Semel et al. 2006) was used at 48-months-old. It yields
161
norm-referenced scores for receptive (sentence structure, concepts and following direction, and basic concepts subtests) and expressive language (word structure, expressive vocabulary and recalling sentences subtests). The PLS-4 and CELF-P2 standard scores are each mean (M) 100, and standard deviation (SD) 15. The measure changed at 48-months when participants joined a subsequent trial, L4L (Goldfeld et al.,
2012), to harmonise measures between the studies. See study protocols for methodology details (Wake et al., 2011; Wake et al., 2012).
During the 24-month-old visit, mother and child were video-recorded playing together for 15-minutes. The mother was asked to play with her child as she normally would, using a doll and a barnyard. The coding scheme of maternal communicative behaviours and interaction quality are shown in Table 1.
Coding used Observer® XT software (Noldus, 2008). Inter-rater reliability was conducted on 10% of the sample by an independent rater (third author), blind to the primary rater’s coding. The primary rater (first author) re-rated 10% of the sample to calculate intra-rater reliability. Reliability was determined by calculating intra-class correlation random effects models for each behaviour, and ranged from .994 to .999 for intra-rater and .987 to .997 for inter-rater reliability (see Table 3).
The first and last 2.5 minutes of each sample were discarded as ‘warming-up’ and ‘winding-down’ time. The primary rater logged each behaviour observed in the middle 10-minutes. Observer® XT calculated the total number and rate per minute for each behaviour. Maternal responsive communicative behaviours (expansions, imitations, responsive questions, supportive directives and labels) were coded in an earlier study (see Table 1) (Levickis et al., 2014).
162
The dyad’s Fluency and Connectedness was rated on a 7-point scale using the middle 5-minutes; 1 represented ‘no conversation established’, 4 represented
‘conversation lacks smoothness, appears to be dominated by one partner’, 7 represented
‘fluent and balanced conversation that is often sustained’ (Hirsh-Pasek et al., 2015) (see
Table 2). To assist in selecting the appropriate rating, the rater noted episodes of turn- taking, child initiations, shared topic, and conversation progression. For example, an interaction rated 4 typically comprised two episodes of turn-taking with a long pause between them, a small number of child initiations, and missed response opportunities by mother and/or child. An interaction rated 7 comprised several episodes of smooth turn- taking which both mother and child contributed to equally, and conversation progression e.g. cuddling doll progressed to feeding doll.
Inter- and intra-rater reliability for FC was calculated for 10% of the sample (by first and third authors), using kappa to control for chance agreement (Cohen, 1960).
Intra- and inter-rater kappas were .86 and .64, respectively in unweighted tests, with all discrepancies within one point of each other, as considered acceptable by the scale developers (Adamson et al. 2012).
Potential confounders of language development identified a priori, gender, birthweight, birth-order, parental education, maternal age, and neighbourhood disadvantage measured by the Socioeconomic Indices for Area (SEIFA) disadvantage score (ABS, 2001), were collected in questionnaires completed between child ages 12- and 48-months-old.
163
Table 1 Coding Scheme for Maternal Communicative Behaviours Directive Utterance that aims to re-direct child away from the object or activity C: Playing with truck, M: “Look in the (Adapted from Akhtar et on which they are currently visually focused to something new or bag!” al. 1991; Shimpi and different Successful: child looks in bag Successful: child shifts visual focus toward the goal. Huttenlocher, 2007) Unsuccessful: child does not change focus Unsuccessful: child does not shift visual focus. Praise Verbal or non-verbal positive evaluations of child’s efforts, global “Good job!”, “You’re so clever!” (Adapted from Gaertner characteristics, or generic positive evaluations. et al. 2008; Winsler et al. 1999) Missed opportunity Mother does not respond in any way to a child’s utterance after 3 secs C: “Twuck”, M: No response Expansion Mother repeats one or all the child’s preceding words and adds to the C: “Ball”, M: “It’s a red ball” (Levickis et al., 2014) child’s preceding verbalisation Imitation Mother repeats the child’s preceding vocalisation or verbalisation C: “Ball”, M: “Ball” (Levickis et al., 2014) exactly or with a reduction in the words. Responsive question Mother asks a “wh” question (e.g. “what”, “when”, “who”), which is Child is playing with horse (Levickis et al., 2014) immediate and dependent on the child’s preceding act. M: “What’s that?” Supportive directive Mother directs or commands the child to verbalise or do an action in Child is playing with doll (Levickis et al., 2014) relation to what the child is focused on. M: “Feed the doll” Label Mother labels an object or action, which is the focus of the child, with Child is playing with a toy horse (Levickis et al., 2014) the label in the final position of the carrier phase. M: “That’s a horse”
164
Table 2 Coding Scheme for Fluency and Connectedness Fluency and Characterises the flow of conversation Connectedness 1= 2= 3= 4= 5= 6= 7= Modified from No Interaction Instances of child Conversation Shared topic Extension of Fluid and Adamson et al. conversation established but initiations lacks smoothness, throughout. Both interaction and balanced (2012) and Hirsh- established. child reciprocated by appears to be partners engaged in play by both conversation that Pasek et al. frequently not mother. largely dominated relatively equal partners. is often (2015). responding. by one partner. turn-taking. sustained.
165
Analysis
Preliminary analyses examined linearity between communicative behaviours and language scores. Likelihood ratio tests compared regression models where behaviours were modelled as categorical (quartiles) and continuous, to models including continuous terms only. All p-values were >.05, providing no evidence against linearity, so analyses used continuous terms. Language scores were analysed as continuous variables, rescaled to z-scores (M=0, SD=1); maternal behaviours and FC was analysed as continuous variables.
The associations between communicative behaviours and expressive and receptive language scores at 24-, 36- and 48-months-old were examined using unadjusted linear regression (Aim 1). Potential confounders (described above) were added to determine whether the associations remained after adjustment. The final adjusted model examining the associations with 36- and 48-months-old language scores additionally included 24-month-old language scores (receptive/expressive as appropriate).
To investigate whether the association between directives and language scores changed alongside a co-occurring responsive behaviour (Aim 2), Pearson’s correlation coefficients between all communicative behaviours were examined. A responsive and directive behaviour that were weakly correlated to avoid collinearity were selected.
Effect modification, whereby the effect of maternal behaviours on language scores differed by interaction quality, was examined in the cross-sectional models where there was evidence of an association in the unadjusted model (Aim 3). FC × maternal
166
behaviour was added as an interaction term and likelihood ratio tests compared models with and without interaction. Since it can be difficult to observe subtle interaction effects (Kirkwood and Sterne, 2003), further analysis was undertaken using simple slopes (UCLA, Statistical Consulting Group). Predicted 24-month-old language scores were fitted for the maternal behaviour, holding the interaction term constant at each FC level (1-7) to illustrate how the association changed over the full range of interaction quality.
Results
52.3% of the participants were male, 37.1% were first-born, and 9.6% lived in a household where a language other than English was spoken. Despite efforts to recruit across high, medium and low socio-economic areas, the sample was more socially advantaged (M=1026.5, SD=53.3) than the Australian population on average (M=1000,
SD=100) indicated by the SEIFA score (see Table 3). This study comprised 197 of the original 251 families video-recorded. Differences were not evident between the 197 participants who were in this study compared with the remainder of the 251 who were video-recorded, except for maternal age (mothers in the study-sample were 1.3 years older, 95% CI 0.46, 2.14).
167
Table 3 Sample Characteristics
Total n n (%) or M (SD) Gender, male 197 103 (52.3) First-born child 197 73 (37.1) Birthweight (g) 190 3380.3 (620.2) SEIFA index score of disadvantage 197 1026.5 (53.3) Maternal age at child age 12-months 197 34.3 (4.5) NESB 197 19 (9.6) In intervention arm of trial 197 100 (50.8) Parental education: Not completed high school 38 (19.4) Completed high school 65 (33.2) Completed diploma/tertiary qual/postgrad 93 (47.5)
a b Child language standard scores P n M (SD) Low Language n (%) P P Receptive language: 24-months-old 195 90.6 (14.0) 63 (32.3) 36-months-old 190 98.4 (15.2) 24 (12.7) 48-months-old 193 94.7 (14.0) 33 (17.1) Expressive language: 24-months-old 195 91.8 (11.9) 27 (13.9) 36-months-old 189 101.1 (14.5) 16 (8.5) 48-months-old 192 97.7 (14.3) 26 (13.5)
a b Note: SEIFA Socio-Economic Index for Areas; NESB Non-English-speaking background. P P PLS-4 at 24 & 36 months, CELF-P2 at 48 months; P P Low language defined as scoring ≥1.25 SD below the mean on expressive and/or receptive language score
168
Maternal and interactive behaviours are summarised in Table 4. Praise occurred most
frequently (approximately once every minute), missed-opportunities and directives occurred
every 2-3 minutes. The mean FC rating of 3.82 indicated that, on average, interactions lacked smoothness, were dominated by one partner (usually, although not always, the mother), and
were not strongly cohesive.
Table 4 Summary Statistics of Maternal and Interactive behaviours a Communicative behaviours (rate Reliability P n M (SD), min, max per min, rpm) Inter-rater Intra-rater Successful directives 196 0.41 (0.35), 0, 2.1 .987 .994 Unsuccessful directives 196 0.33 (0.42), 0, 2.4 .997 .995 Missed-opportunities 196 0.37 (0.37), 0, 1.9 .990 .986 Praise 196 0.82 (0.71), 0, 3.8 .997 .999 Fluency and Connectedness .644 .857 195 3.82 (1.55), 1, 7 rating (scale 1-7) Previously coded responsive behaviours (rpm) N M (SD), min, max Expansions 197 0.6 (0.61), 0, 3.2 Imitations 197 0.47 (0.48), 0, 3.3 Interpretations 197 0.57 (0.44), 0, 2.3 Labels 197 1.17 (0.68), 0, 4.2 Supportive directives 197 0.6 (0.45), 0, 2.6 Responsive questions 197 0.71 (0.53), 0, 2.7
a Note. P P Intra-class correlation random effects for maternal behaviours; Cohen’s kappa’s (Cohen, 1960) for Fluency and Connectedness
.
169
Aim 1: Maternal behaviours, Interaction Quality and Language
Outcomes
There was no evidence that missed-opportunities or praise were associated with language scores at any age if examined across the continuum of language scores in the unadjusted models or the models adjusted for confounders (see Table 5 – coefficients reported for models adjusted for confounders, and for models adjusted for confounders plus 24-month language score. Unadjusted models are not reported to reduce the size of the tables).
The data were then dichotomised into low (scoring ≥1.25 SD below the normative mean) versus typical expressive language groups (as per Reilly et al. 2010) using the 24-month-old expressive language scores (n=27 low expressive language, n=168 typical expressive language). The mean rate of missed-opportunities was significantly higher for children with typical expressive language scores than for those with low expressive language scores (M = 0.40/min versus M = 0.22/min, p = 0.023).
There were no evident differences for praise rates between participants with low expressive language scores at 24-months and those with typical expressive language scores. No further analyses were undertaken for missed-opportunities or praise.
Successful directives were associated with lower 24-month-old receptive and expressive language scores in the unadjusted models but not after adjusting for confounders, and with lower 36- and 48-month-old receptive and expressive language scores in the unadjusted and adjusted models. For example, one additional successful directive was associated with 0.45 SD lower receptive language score at 36-months-old
[95% CI -0.84, -0.06]. After adding 24-month-old language scores to the models, the 170
association with 36-month-old language scores attenuated, but the 48-month-old associations remained.
Unsuccessful directives were associated with 24-month-old receptive language after adjusting for confounders. Each additional unsuccessful directive was associated with an estimated 0.41 SD lower receptive language score [95% CI -0.74, -0.08], but an association was not evident for 24-month-old expressive language. Unsuccessful directives were associated with poorer 36- and 48-month-old receptive and expressive language scores after adjusting for confounders. For example, each additional unsuccessful directive was associated with an estimated 0.37 SD lower receptive language score at 36-months-old [95% CI -0.69, -0.04] and 0.66 SD lower score at 48- months-old [95% CI -0.99, -0.33]. These associations remained after adding participants’ 24-months language scores to the adjusted models (Table 5). The magnitude of the associations with 36- and 48-month-old language scores was similar for successful and unsuccessful directives.
FC was positively associated with expressive and receptive language scores at
24- and 36-months-old after adjusting for confounders (e.g., 24-months-old receptive language, ES= 0.21, 95% CI [0.12, 0.29]), but with only receptive language scores at
48-months-old (ES=0.18, 95% CI [0.08, 0.27]). The association with 36- and 48-month- old language was no longer evident when 24-month-old language scores were added to the model
171
Table 5 Association Between Maternal Behaviours and Interaction Quality and Child Language Scores
a a Behaviour Adjusted P Adjusted P P + 24m language score
b 2 b 2 Missed-opportunities ES P 95% CI p RP ES P 95% CI p RP 24m Receptive -0.11 -0.48, 0.25 0.547 0.14 Expressive 0.14 -0.23, 0.50 0.453 0.14
36m Receptive -0.22 -0.59, 0.15 0.248 0.18 -0.18 -0.46, 0.11 0.229 0.50 Expressive -0.11 -0.48, 0.26 0.556 0.17 -0.12 -0.43, 0.18 0.427 0.44 48m Receptive 0.06 -0.33, 0.45 0.761 0.11 0.20 -0.14, 0.54 0.237 0.39 Expressive -0.09 -0.46, 0.29 0.645 0.16 0.01 -0.34, 0.36 0.939 0.35 Praise 24m Receptive 0.17 -0.02, 0.36 0.075 0.16
Expressive 0.03 -0.16, 0.22 0.743 0.14
36m Receptive 0.10 -0.10, 0.30 0.344 0.18 -0.08 -0.24, 0.08 0.316 0.49 Expressive 0.06 -0.14, 0.26 0.532 0.17 -0.06 -0.23, 0.11 0.483 0.44 48m Receptive -0.00 -0.20, 0.20 0.997 0.11 -0.11 -0.29, 0.07 0.220 0.39 Expressive 0.06 -0.13, 0.25 0.527 0.17 -0.01 -0.19, 0.17 0.939 0.35 Successful directives 24m Receptive -0.37 -0.75, 0.02 0.063 0.16
Expressive -0.33 -0.72, 0.06 0.095 0.15
36m Receptive -0.45 -0.84, -0.06 0.023 0.20 -0.22 -0.53, 0.10 0.177 0.50 Expressive -0.50 -0.90, -0.09 0.016 0.20 -0.27 -0.61, 0.08 0.130 0.45 48m Receptive -0.70 -1.09, -0.30 <0.001 0.16 -0.49 -0.85, -0.13 0.008 0.42 Expressive -0.74 -1.12, -0.36 <0.001 0.23 -0.58 -0.95, -0.22 0.002 0.38
172
a a Behaviour Adjusted P Adjusted P P + 24m language score b 2 b 2 Unsuccessful directives ES P 95% CI p RP ES P 95% CI p RP 24m Receptive -0.41 -0.74, -0.08 0.015 0.17
Expressive -0.16 -0.49, 0.18 0.350 0.14
36m Receptive -0.37 -0.69, -0.04 0.027 0.20 -0.19 -0.46, 0.08 0.159 0.50 Expressive -0.53 -0.85, -0.22 0.001 0.22 -0.41 -0.69, -0.14 0.004 0.47 48m Receptive -0.66 -0.99, -0.33 <0.001 0.18 -0.54 -0.85, -0.24 <0.001 0.43 Expressive -0.55 -0.87, -0.22 0.001 0.21 -0.46 -0.77, -0.14 0.005 0.38 Fluency & Connectedness 24m Receptive 0.21 0.12, 0.29 <0.001 0.24 Expressive 0.10 0.01, 0.19 0.034 0.16 36m Receptive 0.19 0.10, 0.28 <0.001 0.25 0.06 -0.02, 0.13 0.151 0.50 Expressive 0.15 0.06, 0.24 0.002 0.22 0.03 -0.05, 0.11 0.476 0.44 48m Receptive 0.18 0.08, 0.27 <0.001 0.17 0.08 -0.01, 0.17 0.066 0.40 Expressive 0.06 -0.03, 0.15 0.208 0.17 -0.02 -0.11, 0.07 0.673 0.34
Note. CI = confidence interval. Language measured using PLS-4 at 24 & 36 months-old, CELF-P2 at 48 months-old, z- a scores; P P Adjusted for child gender, birthweight, birth-order, treatment group, neighbourhood disadvantage score b (SEIFA), parental education, maternal age;P P ES = effect size: interpret as the average standard deviation difference in language score for 1 rpm higher communicative behaviour, or one-point higher rating on Fluency and Connectedness
173
Aim 2: Maternal Directives and Co-occurring Maternal Responsive
Behaviour
Unsuccessful directives were selected as the intrusive maternal behaviour for this analysis as they were associated with poorer language scores at each age. They were weakly correlated with both imitations and expansions (r = -0.18, Table 6). Since expansions were previously found to have the stronger association with concurrent and later language scores than imitations (Levickis et al., 2014), they were selected as the responsive behaviour.
174
Table 6 Pearson’s Correlation Matrix Between Maternal Behaviours and Interaction Quality Variable, r-value 1: 2: 3: 4: 5: 6: 7: 8: 9: 10: 11: p-value 1. Praise 1
2. Missed -0.28 1 opportunities <0.001 3. Successful 0.13 -0.1 1 directives 0.078 0.167 4. Unsuccessful 0.04 -0.17 0.46 1 directives 0.538 0.015 <0.001 5. Expansions 0.02 -0.1 -0.19 -0.18 1 0.784 0.183 0.008 0.012 6. Imitations 0.01 -0.07 -0.16 -0.18 0.58 1 0.914 0.331 0.028 0.014 <0.001 7. Interpretations 0.1 -0.03 -0.16 -0.18 0.28 0.19 1 0.183 0.708 0.023 0.013 <0.001 0.01 8. Labels 0.05 -0.25 0.001 -0.01 -0.18 -0.07 0.07 1 0.484 <0.001 0.993 0.9 0.01 0.35 0.327 9. Supportive 0.33 -0.2 0.25 0.16 -0.07 -0.02 0.02 0.1 1 directives <0.001 0.006 0.001 0.024 0.304 0.774 0.8 0.162 10. Responsive 0.14 -0.11 -0.14 -0.09 0.25 0.28 0.03 -0.01 -0.01 1 questions 0.057 0.126 0.052 0.2 <0.001 <0.001 0.633 0.849 0.879 11. Fluency & 0.11 -0.12 -0.06 -0.24 0.45 0.35 0.38 0.08 0.07 0.16 1 Connectedness 0.128 0.086 0.41 0.001 <0.001 <0.001 <0.001 0.29 0.354 0.027 Note: Items 1-10 are rate per minute, Item 11 (Fluency and Connectedness) is on a rating scale from 1 to 7
175
After controlling for expansions, receptive language coefficients at each age for one-unit higher unsuccessful directives attenuated: 24-month-old coefficients attenuated from -0.41 to -0.34, 36-month-old coefficients from -0.37 to -0.29, 48-month-old coefficients attenuated from -0.66 to -0.59. A similar pattern was observed for 36- and
48-months-old expressive language scores after controlling for expansions. In contrast, the receptive and expressive language coefficients associated with one additional expansion after controlling for unsuccessful directives remained almost unchanged (e.g.
24-month-old receptive language without adjustment = 0.48, after adjustment = 0.47)
(see Table 7).
176
Table 7: Associations between Unsuccessful Maternal Directives and Expansions Together and Child Language Scores at 24-, 36- and 48-months-old
a b b Language P Adjusted P Adjusted P P + 24m language score
c 2 c 2 ES P 95% CI p RP ES P 95% CI p RP
24m Receptive: Directive -0.34 [-0.66, -0.03] .034 0.24 Expansions 0.47 [0.25, 0.69] <.001 24m Expressive: Directive -0.08 [-0.40, 0.23] .608 0.25 Expansions 0.56 [0.33, 0.78] <.001 36m Receptive: Directive -0.29 [-0.60, 0.02] .066 0.28 -0.11 [-0.37, 0.15] .421 0.52 Expansions 0.50 [0.28, 0.72] <.001 0.24 [0.04, 0.43] .016 36m Expressive: Directive -0.45 [-0.75, -0.15] .003 0.33 -0.44 [-0.71, -0.16] .002 0.45 Expansions 0.56 [0.35, 0.78] <.001 0.36 [0.15, 0.57] <.001 48m Receptive: Directive -0.59 [-0.91, -0.27] <.001 0.24 -0.42 [-0.71, -0.13] .005 0.42 Expansions 0.44 [0.21, 0.67] <.001 0.23 [0.02, 0.44] .032 48m Expressive: Directive -0.48 [-0.79, -0.17] .003 0.28 -0.47 [-0.77, -0.17] .002 0.35 Expansions 0.45 [0.23, 0.68] <.001 0.34 [0.11, 0.56] .004 Coefficients for expansions without adjusting for directives, for comparison purposes 24m Receptive 0.48 [0.25, 0.70] <0.001 0.22 24m Expressive 0.55 [0.34, 0.77] <0.001 0.25 36m Receptive 0.51 [0.28, 0.73] <0.001 0.26 0.16 [-0.04, 0.35] 0.115 0.50 36m Expressive 0.59 [0.37, 0.80] <0.001 0.29 0.26 [0.06, 0.46] 0.013 0.46 48m Receptive 0.47 [0.23, 0.71] <0.001 0.17 0.22 [-0.00, 0.45] 0.054 0.41 48m Expressive 0.48 [0.25, 0.70] <0.001 0.24 0.27 [0.04, 0.50] 0.023 0.37
a Note. CI = confidence interval. P P Language measured using PLS-4 at 24- & 36-months-old, CELF- b P2 at 48-months-old (z-scores); P P Adjusted for child gender, birthweight, birth-order, treatment c group, neighbourhood disadvantage score (SEIFA), parental education, maternal age; P P ES = effect size: interpret as the average standard deviation difference in language score for 1rpm higher communicative behaviour
177
Aim 3: Maternal Behaviours in the Context of Mother-Child Fluency and Connectedness
Examination of FC as an interaction term in the maternal behaviour and 24- month-old language models found that a significant interaction effect was only evident for the association between successful directives and receptive language (Table 8). For each point higher FC rating, the predicted slope became shallower by 0.28 SD (95% CI
[0.03, 0.54], p=0.025). Thus, in dyads with lower levels of FC, the estimated mean difference in receptive language score associated with each additional successful directive was more substantive than in dyads with higher FC. That is, the negative association between successful directives and language scores was stronger in poorly- connected dyads.
Further investigation of effect modification using simple slopes suggested that the negative association between successful directives and receptive language primarily existed for dyads with no or low FC (rated 1-3) (see Figure 1 and accompanying Table).
Simple slopes were examined for the associations between 24-month-old language scores and expansions, and between 24-month-old language scores and unsuccessful directives (with and without controlling for co-occurring expansions). As the initial associations were not statistically significant these results are included in the supplementary Figure and Table to avoid over-interpretation
178
Table 8: Interaction Effects of Fluency and Connectedness on the Association between Language Score and Maternal Behaviours at 24-months-old a Unadjusted Adjusted P
b c b c Behaviour Coef.P 95% CI p P Coef.P 95% CI p P d Receptive language P Unsuccessful directives -0.18 [-0.42, 0.05] .123 -0.15 [-0.63, 0.95] .199 Successful directives 0.29 [0.05, 0.54] .018 0.28 [0.03, 0.54] .025 Expansions -0.08 [-0.21, 0.05] .237 -0.06 [-0.19, 0.07] .366 Expansions and -0.11 [-0.24, 0.03] .250 -0.08 [-0.22, 0.05] .430 e unsuccessful directives P -0.11 [-0.36, 0.14] -0.08 [-0.33, 0.18] d Expressive language P Unsuccessful directives -0.22 [-0.47, 0.03] .076 -0.19 [-0.44, 0.07] .138 Successful directives 0.03 [-0.23, 0.30] .794 0.07 [-0.21, 0.34] .608 Expansions -0.12 [-0.26, 0.01] .069 -0.12 [-0.26, 0.01] .058 Expansions and -0.14 [-0.27, 0.00] .133 -0.14 [-0.27, 0.00] .130 e unsuccessful directives P -0.09 [-0.34, 0.17] -0.06 [-0.32, 0.20]
a Note: CI = confidence interval. P P Adjusted for child gender, birthweight, birth-order, treatment group, neighbourhood disadvantage score (SEIFA), parental b c education, and maternal age; P P interaction coefficient; P P p-value for likelihood ratio test comparing regression model containing the interaction term with d e regression model without the interaction term. P P Language measured using PLS-4 at 24- and 36-months-old, CELF-P2 at 48-months-old (z-scores). P P Model includes both behaviours adjusted for one another: expansions and language adjusted for unsuccessful directives, and unsuccessful directives and language adjusted for expansions.
179
Figure 1: Simple slopes modelling effect modification by Fluency and Connectedness on the association between maternal successful directives and child receptive language scores at 24-months-old *denotes p<0.05
Simple slopes graph Estimated association at each level of Fluency and Connectedness Fluency & Receptive language score Connectedness ES 95% CI p 1 -1.18 -1.92, -0.44 .002 2 -0.89 -1.43, -0.35 .001 3 -0.59 -0.98, -0.20 .003 4 -0.30 -0.67, 0.07 .111 5 -0.01 -0.50, 0.49 .980 6 0.29 -0.40, 0.98 .410 7 0.58 -0.33, 1.49 .208
180
These findings showed consistency with those of the association between successful directives and receptive language scores, insofar as the significant associations varied by FC level. For example, in contrast to successful directives, the association between unsuccessful directives and receptive language scores was statistically significant and stronger at higher FC levels (rated 3-7). The associations between expansions and receptive and expressive language scores were significant at levels 1-6, but weakened at higher FC levels.
Discussion
This study highlights the importance of considering maternal behaviours and the interactional context for ongoing language learning amongst children with low expressive language at 18-months. We explored the association between maternal behaviours and interaction quality at age 24-months and language outcomes at 24-, 36- and 48-months. Nuanced associations between language scores and directives, which a child did or did not follow, were identified. Directives which the child did not follow were associated with poorer 24-month receptive language skills, yet both types were associated with poorer 36- and 48-month expressive and receptive language skills.
These associations were weaker when maternal expansions were considered, revealing how combinations of maternal input may be differentially associated with language- learning. Mothers of toddlers with expressive language difficulties at 24-months missed fewer response opportunities than mothers of children with typical language, suggesting mothers adapted to their children’s poorer expressive skills. Higher Fluency and
Connectedness were associated with better child language skills, and modified the effects of maternal successful directives. The negative association between successful
181
directives and 24-month-old receptive language scores was only evident for parents and children with low Fluency and Connectedness.
Aim 1: Maternal Behaviours, Interaction Quality and Language
Outcomes
In this study, mothers were more likely to respond (verbally or non-verbally) to communicative acts of toddlers with poorer expressive language abilities than those with typical language. These mothers may have adapted their behaviour to their children’s delayed skills and became more vigilant of their communicative attempts.
This explanation would be in line with reports suggesting that parents of children with developmental difficulties are highly responsive to child initiations and opportunities to interact, adjusting their social communication over time to support their children’s level of development (e.g. Guralnick et al. 2008). We found no evidence that more maternal praise was associated with higher language scores. This might have been due to the unstructured task lacking contextual “draw” (Blacher et al. 2013) to elicit the levels and types of praise meaningful to the child, and the broad definition employed. Further research is therefore required. Whether intervention-specific praise has therapeutic value also requires investigation.
The findings that unsuccessful intrusive directives were associated with lower concurrent receptive language scores independent of confounders supports previous investigations that have found directives which lead children’s attention are associated with poorer language skills (Akhtar et al. 1991; Masur et al. 2005; Tomasello and Todd,
1983). Successful directives were also associated with poorer receptive language, but only before adjusting for confounders. This finding is at odds with our hypothesis that 182
successful directives might provide a subsequent language-learning opportunity, based on Shimpi and Huttenlocher’s (2007) research into maternal labels. Perhaps directives are more likely to be unsuccessful than successful in children with lower receptive language skills because they may fail to understand the directives. Mothers of children with low receptive language may have repeated directives until their children responded, resulting in a stronger association between receptive language scores and unsuccessful directives than between receptive language scores and successful directives. Whilst this could indicate a mismatch between the level at which the mother was targeting her language and the child’s competency level, it is also feasible that it was appropriate and adaptive for slower language-learners. By usually referring to the immediate environment and so having high referential transparency, directives conceivably could facilitate word-learning for children with poor language-learning skills. The negative association could reflect the fact that mothers of children with low language adapt by using more directives than mothers of those with typically developing language.
Both successful and unsuccessful directives used at 24-months-old were associated with poorer receptive and expressive language scores at 36- and 48-months- old. In the case of 48-month language scores, this association was evident even after adjusting for 24-month-old language scores. This might be explained by slow language learners continuing to learn language slowly and falling behind their peers at 48- months. Another possibility is that mothers’ continued use of directives (assuming directive use at 24-months is indicative of ongoing directive use) moves from being appropriately adaptive to being maladaptive, having a detrimental effect on language abilities at 36- and 48-months. Cumulatively, directives may limit children’s
183
opportunities for conversation, disrupt the children’s attention (Tomasello and Todd,
1983), fail to stimulate lexical hierarchies and representations in their minds (Diaz et al.
1991), and result in less effective storage and retrieval from their lexicon, resulting in poorer language development. However, we need repeated observations of directives over time to determine whether maternal directive use at 24-months is indicative of directive use at later time points.
The final behaviour, Fluency and Connectedness (FC), assessed the quality of the mother-child dyadic interaction. We found that more fluent and connected interactions were associated with higher concurrent and later expressive and receptive language scores, expanding upon previous findings regarding expressive language skills only in a younger sample (Hirsh-Pasek et al. 2015). Interactions where the mother and child are poorly connected may provide a sub-optimal language-learning foundation, whilst well-connected interactions may provide a strong scaffold within which the child can learn, practice and build upon new words. In the absence of such a scaffold, parents’ words may pass by “like background noise”, having no impact on child learning (Hirsh-
Pasek et al. 2015, p.1081).
It is unlikely that the association between child language skills and FC is unidirectional. Better parental language skills likely facilitate their capacity to engage with their children, and better child language abilities likely facilitate children’s capacity to engage and connect with their parents (Hudry et al. 2013). Poorer child language skills in contrast are associated with lower rates of initiation and responsiveness (Conti-Ramsden et al. 1995), which impact children’s ability to engage with, and be engaged by, their parents. Even with parents’ best intentions, poorer child language skills may restrict the degree of flow possible during an interaction. This is 184
illustrated by the association between FC and later language scores attenuating once concurrent language scores were included in the models; children’s 24-month-old language skills likely contributed to how well they interacted with their mother, as well as their later language skills.
Aim 2: Directives and Co-occurring Expansions
The negative association between directives and language scores weakened in the presence of expansions, whilst the positive association between expansions and language scores was maintained. This provides empirical support for the idea that directiveness does not necessarily occur at the expense of the facilitative aspects of responsivity (Marfo, 1990). Unlike directives, expansions provide enriched linguistic input in the form of greater semantic and phonological information about words (Hoff,
2003), resulting in more robust lexical hierarchies. The language-learning opportunity afforded to the child by these expansions may have been greater than any deficits associated with unsuccessful directives.
Aim 3: Maternal Behaviours in the Context of Mother-Child Interaction
Quality
The final part of this study considered whether the dyadic environment in which communicative behaviours were used modified their association with language scores.
As Ratner (2013) proposed, individual differences in adult-child dyadic interactions could “create differences in how children best exploit their language environments to learn” (p214). Our findings partially support this. The negative association between successful directives and 24-month-old receptive language scores was strongest in
185
minimally-connected dyads (rated 1-3), whilst there was no evidence of this association in well-connected dyads (rated 4-7).
Children in minimally-connected dyads may be particularly susceptible to the negative effects of directives on language learning, whilst those in well-connected dyads may be buffered from them. Alternatively, mothers may use directives to engage children who have difficulties initiating and/or sustaining a connection, as suggested by
Hudry et al.’s (2013) findings of parent-child interaction styles among young children with ASD. The fact that these directives were successfully followed by the child points to the success of the mothers' strategy. This success may reinforce the mothers' use of directives. Maternal on-going use of directives may continue to be adaptive, as long as the child is learning language within their zone of proximal development (Vygotsky,
1978). However, since directives typically do not foster turn-taking or conversation, their continued use may restrict a dyad from developing along the FC scale once the child’s language learning is ready to be extended. This may mean mother and child perpetuate a pattern of interaction that is sub-optimal for language-learning, and continue to function in a lower quality interactive system.
Whilst there was no evidence for a significant interaction effect among the other associations examined, exploratory simple slope analysis revealed how the connectedness between mother and child might modify the associations between other maternal communicative behaviours and child language learning. This provides a more nuanced understanding of the developmental precursors of children’s language (Hirsh-
Pasek et al. 2015). Repeated measures of interaction quality are needed to explore these issues in more detail, along with repeated measures of maternal behaviours to monitor change or stability in use over time. The resulting rich understanding may inform which 186
parenting behaviours might be most appropriately targeted in dyads of different qualities and at different stages of development.
Strengths and Limitations
The study sample included children with low expressive vocabulary at 18- months-old and thus limits the generalisability of the findings. This is important given that subsequent studies revealed that early expressive vocabulary delays alone do not accurately predict later language difficulties (e.g. Reilly et al. 2010). This sample contained a similar percentage of children with language difficulties at 36- and 48- months-old as found in a community population sample (Reilly et al. 2010).
Nevertheless, since studies have identified differences in the interactions of carers of typically developing children compared to those of children with DLD (e.g. Conti-
Ramsden et al. 1995), there may be differences in the maternal behaviours used in this sample compared to either a general population or a late-talker sample (usually defined by 24-months-old language skills).
Our ability to evaluate the direction of effects was restricted by having no direct measures of the child’s contributions during the interaction (e.g. interest levels, locomotion, initiations), and only having a single observation of the interaction. A recent examination of treatment mechanisms for an RCT for children with ASD incorporated three parent-child observations over 12-months (Pickles et al. 2015). These repeated observations enabled the researchers to use mediation techniques to uncover causal relationships between changes in parental and child behaviours. Repeated observations in studies such as the current one might allow similar techniques to be used to evaluate the direction of effect between parental and interactive behaviours and
187
child language scores. Of course, the observations would need to take place within a sufficient timeframe for developmental changes to occur. As noted earlier, this approach could also reveal how parental and interactive behaviours change over time.
Further limitations include the relatively socially-advantaged sample. This means dyads with higher rates of missed-opportunities and lower rates of praise, due to experiencing greater stressors associated with economic and social adversity, might be under-represented. Changing from the PLS-4 to the CELF-P2 at 48-months, which is not unprecedented or avoidable in longitudinal studies, was not ideal as it introduced the potential for non-equivalence. The potential for measurement error also existed due to observed maternal behaviours potentially not being representative of typical, non- observed behaviour (Uziel, 2010). Further, a measure of maternal language ability would have enabled us to examine how this might be associated with maternal behaviours used and the FC measure. Finally, the inter-rater reliability for the FC scale was moderate at 0.64. Although all ratings were within one point of each other, as considered acceptable by the developers of the scale (Adamson et al. 2012; also see
Bakeman and Quera, 2011), refinement would be necessary for use in future studies.
The strengths of this study include: a large sample size relative to other studies in this area; recruitment from the community rather than concerned families who present to clinics; the development of a reliable coding scheme that can be implemented with minimal cost and thorough training, and; the combined consideration of maternal behaviours and the interactional context on language development.
188
Practical Implications
The findings have practical implications for early language promotion and intervention strategies. Firstly, addressing how caregivers interpret their child’s response to directives may be important for changing their communicative behaviours.
For example, parents who frequently use directives might benefit from learning to use other strategies which might break maladaptive interaction patterns. Secondly, promoting parental behaviours associated with better language development (e.g. expansions), rather than advising against certain behaviours (e.g. directives) might be an effective strategy for clinicians and one that is more acceptable to parents.
A third clinical implication is that it may be important to target interaction quality in interventions alongside teaching specific communicative behaviours, as has been recommended previously (Hirsh-Pasek et al. 2015; Roberts and Kaiser, 2011). The
FC measure may be a practical tool for clinicians. However, research is required to determine the components of this single item, with the aim of isolating the skills needed by parent and child to participate at each FC level; to determine how clinicians can best coach parents to use these skills; and how parents can teach their children to use these skills. This includes investigating the appropriate dosage, that is, what intensity and frequency is required to influence children’s language outcomes (Dunst et al. 2014).
Finally, our finding that interaction quality modified the association between successful directives and poorer receptive language outcomes, whereby the association was strongest in poorly connected dyads whilst not evident in well-connected dyads, has implications for the design of intervention strategies. It supports the idea that interaction quality may form the foundation of future language learning (Hirsh-Pasek et al. 2015)
189
and differences in dyadic interaction may create differences in how children use their environmental inputs (Ratner, 2013). Coaching caregivers about fostering high quality interactions might be best integrated throughout all intervention sessions. Parents of children with poorer language could benefit from support to recognise poorly-formed communication bids from their child and provide rich and individually-tailored language stimulation. In cases where the dyads are already well-connected the potential impact of these strategies may be lower, as there will be fewer opportunities for change.
The FC ratings might inform clinical decision-making and be straightforward to use in a sensitive manner, both recommended prerequisites for parent-child interaction assessment to become routine practice (Mahoney et al. 1996). An assessment of a dyad’s connectedness could assist clinicians to work with parents to foster an interactive environment that is most conducive to an individual child’s learning needs.
Future Research
Our findings verify the need to incorporate the bidirectional model into language development research and practice. Research should continue to investigate other communicative behaviours and aspects of the interaction that may be associated with language outcomes at different ages and stages. It would benefit from incorporating child behavioural measures and using repeated observations to disentangle the complex parent-child transactional relationships. Research is required to design and test practical strategies to teach caregivers to engage in more fluent and connected interactions with their children, and finally whether assessing interaction quality is appropriate and helpful during a language assessment.
190
Conclusion
Using a sample of slow-to-talk children, the study adds to the existing literature about children’s home language-learning environment in several ways. Firstly, by identifying that maternal intrusive directives, whilst being associated with poorer language scores at 24-, 36- and 48-months-old, also appear to have a weakened association with language outcomes in the presence of maternal responsive expansions.
Secondly, by confirming that investigating the mother-child interactive dyad as its own entity is essential to child language research. Thirdly, by illustrating how the combined effect of different types of maternal communicative behaviours and features of the interaction itself might interact to predict child language outcomes. Most importantly, our findings contribute to the ongoing effort to generate evidence-based recommendations for clinical practice to help children with, or at risk of, developmental language disorder.
191
Supplementary Figure 1: Simple slopes for associations between maternal behaviours and 24 month old language z-scores by Fluency and Connectedness level
Receptive language Expressive language
192
Supplementary Table 1 Simple slopes statistics for the association between maternal behaviours and 24-month language scores by Fluency and Connectedness level (1-7)
a d d FC P P rating Receptive language P Expressive language P b b ES P 95% CI c p ES P 95% CI p Successful directives 1 -1.18 -1.92, -0.44 .002 -0.46 -1.25, 0.33 .253 2 -0.89 -1.43, -0.35 .001 -0.43 -1.0, 0.15 .146 3 -0.59 -0.98, -0.21 .003 -0.39 -0.81, 0.03 .066 4 -0.30 -0.67, 0.07 .111 -0.36 -0.75, 0.04 .077 5 -0.01 -0.50, 0.49 .980 -0.32 -0.85, 0.21 .231 6 0.29 -0.40, 0.98 .410 -0.29 -1.02, 0.45 .442 7 0.58 -0.33, 1.49 .208 -0.25 -1.22, 0.72 .608 Unsuccessful directives 1 0.02 -0.55, 0.60 .932 0.30 -0.31, 0.91 .334 2 -0.16 -0.56, 0.24 .435 0.08 -0.35, 0.50 .723 3 -0.34 -0.66, -0.03 .034 -0.15 -0.48, 0.19 .388 4 -0.53 -0.92, -0.14 .008 -0.37 -0.78, 0.04 .078 5 -0.71 -1.27, -0.15 .014 -0.59 -1.19, 0.00 .050 6 -0.89 -1.66, -0.12 .023 -0.82 -1.63, -0.00 .049 7 -1.08 -2.07, -0.09 .033 -1.04 -2.09, 0.01 .051 Expansions 1 0.69 0.11, 1.27 .020 1.07 0.48, 1.66 <.001 2 0.61 0.15, 1.07 .010 0.95 0.48, 1.41 <.001 3 0.53 0.18, 0.89 .003 0.82 0.47, 1.18 <.001 4 0.45 0.19, 0.72 .001 0.70 0.43, 0.97 <.001 5 0.38 0.15, 0.60 .001 0.58 0.35, 0.81 <.001 6 0.30 0.04, 0.56 .026 0.46 0.19, 0.722 .001 7 0.22 -0.13, 0.56 .211 0.34 -0.01, 0.68 .058
193
Supplementary Table 1 continued.
a d d FC P P rating Receptive language P Expressive language P b b ES P 95% CI c p ES P 95% CI p Unsuccessful directives controlling for expansions 1 -0.11 -0.70, 0.47 .702 0.04 -0.56, 0.64 .883 2 -0.23 -0.62, 0.17 .265 -0.04 -0.45, 0.36 .839 3 -0.34 -0.65, -0.02 .035 -0.13 -0.45, 0.19 .430 4 -0.45 -0.85, -0.04 .030 -0.22 -0.63, 0.20 .307 5 -0.56 -1.16, 0.04 .066 -0.30 -0.91, 0.31 .331 6 -0.67 -1.49, 0.15 .108 -0.39 -1.23, 0.45 .363 7 -0.78 -1.84, 0.27 .146 -0.48 -1.56, 0.61 .387 a b P P FC= Fluency and Connectedness Level; P P ES=Estimated effect size for the association between maternal behaviour and concurrent language at each level of Fluency and c d Connectedness; P P CI=Confidence Interval; P P Language assessed using PLS-4
194
References Adamson, L.B., Bakeman, R., Deckner, D.F., & Nelson, P.B. (2012). Rating Parent-Child Interactions: Joint Engagement, Communication Dynamics, and Shared Topics in Autism, Down Syndrome, and Typical Development. Journal of Autism and Developmental Disorders, 42(12), 2622-2635. doi:10.1007/s10803-012-1520-1 Akhtar, N., Dunham, F., & Dunham, P.J. (1991). Directive interactions and early vocabular development: The role of joint attentional focus. Journal of Child Language, 18(1), 41-49. Alston, E., & James‐Roberts, I.S. (2005). Home environments of 10‐month‐old infants selected by the WILSTAAR screen for pre‐language difficulties. International Journal of Language & Communication Disorders, 40(2), 123-136. Australian Bureau of Statistics (2001). Socio-economic Indexes for Areas. Canberra, Australia: Australian Bureau of Statistics. Bagner, D.M., Garcia, D., & Hill, R. (2016). Direct and indirect effects of behavioral parent training on infant language production. Behavior Therapy, 47(2), 184-197. Bakeman, R., & Quera, V. (2011). Sequential analysis and observational methods for the behavioral sciences. Cambridge University Press. Bedford, R., Gliga, T., Frame, K., Hudry, K., Chandler, S., Johnson, M.H., Charman, T. & The Basis Team. (2013). Failure to learn from feedback underlies word learning difficulties in toddlers at risk for autism. Journal of Child Language, 40(1), 29-46. doi:10.1017/s0305000912000086 Beitchman, J.H., Brownlie, E.B., & Bao, L. (2014). Age 31 Mental Health Outcomes of Childhood Language and Speech Disorders. Journal of the American Academy of Child and Adolescent Psychiatry, 53(10), 1102-1110. doi:10.1016/j.jaac.2014.07.006 Bishop, D.V.M., Snowling, M.J., Thompson, P.A., Greenhalgh, T., & CATALISE Consortium (2016). CATALISE: A Multinational and Multidisciplinary Delphi Consensus Study. Identifying Language Impairments in Children. Plos One, 11(7), 26. doi:10.1371/journal.pone.0158753 Blacher, J., Baker, B.L., & Kaladjian, A. (2013). Syndrome specificity and mother-child interactions: examining positive and negative parenting across contexts and time. Journal of Autism and Developmental Disorders, 43(4), 761-774. doi:10.1007/s10803-012-1605-x Bruner, J.S. (1983). Play, thought, and language. Peabody Journal of Education, 60(3), 60- 69. doi:10.1080/01619568309538407 Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychosocial Measurement, 20, 37-46. Conti-Ramsden, G., Hutcheson, G.D., & Grove, J. (1995). Contingency and breakdown: children with SLI and their conversations with mothers and fathers. Journal of Speech and Hearing Research, 38(6), 1290-1302. Crawley, S.B., & Spiker, D. (1983). Mother-child interactions involving two-year-olds with Down syndrome: A look at individual differences. Child Development, 54, 1312- 1323. Diaz, R.M., Neal, C.J., & Vachio, A. (1991). Maternal teaching in the zone of proximal development: A comparison of low-and high-risk dyads. Merrill-Palmer Quarterly, 37(1), 83-107. Dunst, C.J., Trivette, C.M., & Raab, M. (2014). Everyday Child Language Learning Early Intervention Practices. Infants & Young Children, 27(3), 207-219. doi:10.1097/iyc.0000000000000015 195
Funamoto, A., & Rinaldi, C.M. (2015). Measuring parent-child mutuality: A review of current observational coding systems. Infant Mental Health Journal, 36(1), 3-11. doi:10.1002/imhj.21481 Gaertner, B.M., Spinrad, T.L., & Eisenberg, N. (2008). Focused attention in toddlers: Measurement, stability, and relations to negative emotion and parenting. Infant and Child Development, 17(4), 339-363. doi:10.1002/icd.580 Guralnick, M.J., Neville, B., Hammond, M.A., & Connor, R.T. (2008). Mothers' social communicative adjustments to young children with mild developmental delays. American Journal on Mental Retardation, 113(1), 1-18. doi:10.1352/0895- 8017(2008)113[1:mscaty]2.0.co;2 Hirsh-Pasek, K., Adamson, L. B., Bakeman, R., Owen, M.T., Golinkoff, R.M., Pace, A., Yust, P., & Suma, K. (2015). The Contribution of Early Communication Quality to Low-Income Children's Language Success. Psychological Science, 26(7), 1071- 1083. doi:10.1177/0956797615581493 Hoff, E. (2003). The Specificity of Environmental Influence: Socioeconomic Status Affects Early Vocabulary Development Via Maternal Speech. Child Development, 74(5), 1368-1378. doi:http://dx.doi.org/10.1111/1467-8624.00612 Hudry, K., Aldred, C., Wigham, S., Green, J., Leadbitter, K., Temple, K., Barlow, K., McConachie, H., & The PACT Consortium. (2013). Predictors of parent-child interaction style in dyads with autism. Research in Developmental Disabilities, 34(10), 3400-3410. doi:10.1016/j.ridd.2013.07.015 Huttenlocher, J., Waterfall, H., Vasilyeva, M., Vevea, J., & Hedges, L.V. (2010). Sources of variability in children's language growth. Cognitive Psychology, 61(4), 343-365. doi:10.1016/j.cogpsych.2010.08.002 Kirkwood, B.R., & Sterne, J.A.C. (2003). Essential Medical Statistics (2nd ed.). Malden, Mass: Blackwell Science. Landry, S.H., Smith, K.E., Miller-Loncar, C.L., & Swank, P.R. (1997). Predicting cognitive-language and social growth curves from early maternal behaviors in children at varying degrees of biological risk. Developmental Psychology, 33(6), 1040-1053. Levickis, P., Reilly, S., Girolametto, L., Ukoumunne, O.C., & Wake, M. (2014). Maternal Behaviors Promoting Language Acquisition in Slow-to-Talk Toddlers: Prospective Community-based Study. Journal of Developmental and Behavioral Pediatrics, 35(4), 274-281. doi:10.1097/dbp.0000000000000056 Mahoney, G., Spiker, D., & Boyce, G. (1996). Clinical Assessments of Parent-Child Interaction Are Professionals Ready To Implement This Practice? Topics in Early Childhood Special Education, 16(1), 26-50. Marfo, K. (1990). Maternal directiveness in interactions with mentally handicapped children: An analytical commentary. Journal of Child Psychology and Psychiatry, 31(4), 531-549. Masur, E.F., Flynn, V., & Eichorst, D.L. (2005). Maternal responsive and directive behaviours and utterances as predictors of children's lexical development. Journal of Child Language, 32(1), 63-91. Morgan, P.L., Farkas, G., Hillemeier, M.M., Hammer, C.S., & Maczuga, S. (2015). 24- Month-Old Children With Larger Oral Vocabularies Display Greater Academic and Behavioral Functioning at Kindergarten Entry. Child Development, 86(5), 1351- 1370. doi:10.1111/cdev.12398 Noldus Information Technology. (2008). The Observer XT (Version 8.0). Norbury, C.F., Gooch, D., Baird, G., Charman, T., Simonoff, E., & Pickles, A. (2016). Younger children experience lower levels of language competence and academic 196
progress in the first year of school: evidence from a population study. Journal of Child Psychology and Psychiatry, 57(1), 65-73. doi:10.1111/jcpp.12431 Paul, R., & Shiffer, M.E. (1991). Communicative initiations in normal and late-talking toddlers. Applied Psycholinguistics, 12(4), 419-431. Pickles, A., Harris, V., Green, J., Aldred, C., McConachie, H., Slonims, V., Le Couteur, A., Hudry, H., Charman, T. & The PACT Consortium. (2015). Treatment mechanism in the MRC preschool autism communication trial: implications for study design and parent- focussed therapy for children. Journal of Child Psychology and Psychiatry, 56(2), 162-170. doi:10.1111/jcpp.12291 Pungello, E.P., Iruka, I.U., Dotterer, A.M., Mills-Koonce, R., & Reznick, J.S. (2009). The effects of socioeconomic status, race, and parenting on language development in early childhood. Developmental Psychology, 45(2), 544-557. doi:http://dx.doi.org/10.1037/a0013917 Ratner, N.B. (2013). Why Talk with Children Matters: Clinical Implications of Infant- and Child-Directed Speech Research. Seminars in Speech and Language, 34(4), 203- 214. doi:10.1055/s-0033-1353449 Reilly, S., Wake, M., Ukoumunne, O.C., Bavin, E., Prior, M., Cini, E., Conway, L., Eadie, P & Bretherton, L. (2010). Predicting language outcomes at 4 years of age: findings from Early Language in Victoria Study. Pediatrics, 126(6), e1530-e1537. doi:10.1542/peds.2010-0254 Roberts, M.Y., & Kaiser, A.P. (2011). The effectiveness of parent-implemented language interventions: A meta-analysis. American Journal of Speech-Language Pathology, 20, 180-199 Roy, P., Kersley, H., & Law, J. (2005). The Sure Start Language Measure Standardisation Study. Retrieved from http://tna.europarchive.org/20070101101348/http://www.dfes.gov.uk/research/ programmeofresearch/projectinformation.cfm?projectid=14628&resultspage=1 website: Rowe, M.L. (2012). A longitudinal investigation of the role of quantity and quality of child- directed speech in vocabulary development. Child Development, 83(5), 1762-1774. doi:10.1111/j.1467-8624.2012.01805.x Sameroff, A.J. (2009). The transactional model. In A. Sameroff (Ed.), The transactional model of development: How children and contexts shape each other. (pp. 3-21). Washington, DC: American Psychological Association Semel, E., Wiig, E.H., & Secord, W.A. (2006). The Clinical Evaluation of Language Fundamentals-Preschool, Second Edition (CELF-P2): Australian Standardised Edition (2 ed.). NSW, Australia: Harcourt Assessment Inc. Shimpi, P.M., & Huttenlocher, J. (2007). Redirective labels and early vocabulary development. Journal of Child Language, 34(4), 845-859. doi:http://dx.doi.org/10.1017/S0305000907008112 Tomasello, M., & Todd, J. (1983). Joint attention and lexical acquisition style. First Language, 4, 197-212. UCLA: Statistical Consulting Group. How can I explain a continuous by continuous interaction? Stata FAQ. Accessed August 2016 at http://www.ats.ucla.edu/stat/stata/faq/conconb12.htm Uziel, L. (2010). Look at Me, I'm Happy and Creative: The Effect of Impression Management on Behavior in Social Presence. Personality and Social Psychology Bulletin, 36(12), 1591-1602. doi:10.1177/0146167210386239 Vygotsky, L.S. (1978). Mind in society: The development of higher mental process. Cambridge, MA: Harvard University Press. 197
Wake, M., Tobin, S., Girolametto, L., Ukoumunne, O. C., Gold, L., Levickis, P., Sheenan, J., Goldfeld, S., Reilly, S. (2011). Outcomes of population based language promotion for slow to talk toddlers at ages 2 and 3 years: Let's Learn Language cluster randomised controlled trial. BMJ, 343, d4741-d4741. doi:10.1136/bmj.d4741 Wake, M., Levickis, P., Tobin, S., Zens, N., Law, J., Gold, L., Ukoumunne, O. C., Goldfeld, S., Le, H.N., Skeat, J., Reilly, S. (2012). Improving outcomes of preschool language delay in the community: protocol for the Language for Learning randomised controlled trial. BMC Pediatrics,12, 96. http://www.biomedcentral.com/1471-2431/12/96 Winsler, A., Diaz, R.M., McCarthy, E.M., Atencio, D.J., & Chabay, L.A. (1999). Mother– child interaction, private speech, and task performance in preschool children with behavior problems. Journal of Child Psychology and Psychiatry, 40(6), 891-904. doi:10.1111/1469-7610.00507 Zimmerman, I.L., Steiner, V.G., & Pond, R.E. (2002). The Preschool Language Scale, 4th Edition: San Antonio, TX: The Psychological Corporation.
198
6.3 Supplementary Results
6.3.1 Overview
To recap, the preceding paper considered aspects from all four parts of Aim
Two. That is, firstly how maternal behaviours observed at age 24 months (i.e. praise, missed opportunities, successful and unsuccessful directives) and interactive behaviours
(i.e. Fluency and Connectedness) were associated with receptive and expressive language at 24, 36 and 48 months. Secondly, whether a maternal responsive behaviour
(i.e. expansions) modified the effect of a maternal intrusive behaviour (i.e. successful directives). Thirdly, whether the associations between maternal behaviours (i.e. directives, expansions, imitations, responsive questions) and child language scores were modified by an interactive behaviour (i.e. Fluency and Connectedness).
The following section completes the consideration of maternal behaviours that might be associated with language outcomes by examining our findings for the association between Emotion and Mental State Talk (EMST) and language outcomes at
24, 36 and 48 months. A review of the relevant literature was presented in Chapter
Three (section 3.7), and the methods and analysis plan were presented in Chapter Four.
6.3.2 Results
There was evidence that EMST was positively associated with receptive and expressive language at 24 and 36 months in the unadjusted linear regression models
(Table 11). For example, for each additional EMST per minute, 24 month receptive language scores were 0.16 SD higher (95% CI [0.02, 0.31], p = 0.028). The regression
199
analyses revealed no evidence for the association at 48 months (receptive language: ES
= 0.10 [-0.04, 0.25], p = 0.17); expressive language ES = 0.14 [-0.01, 0.29], p = 0.07).
After adjusting for potential confounders (child gender, birthweight, birth order,P
treatmentP group, SEIFA index score of disadvantage, parental education, and maternal age), the associations at 24 and 36 months were no longer statistically significant (e.g.
24 month receptive language: ES RadjR = 0.13 [-0.02, 0.28], p = 0.085)
Table 11 Association between Emotion and Mental State Talk and Language Scores a Unadjusted Adjusted P Language b c 2 b 2 d DomainP β P 95% CI p RP β P 95% CI p RP 24m Receptive 0.16 [0.02, 0.31] .028 2.5 0.13 [-0.02, 0.28] .085 15.6 Expressive 0.19 [0.05, 0.33] .009 3.5 0.12 [-0.03, 0.27] .107 15.3 36m Receptive 0.22 [0.07, 0.37] .005 4.1 0.15 [-0.00, 0.30] .056 19.0 Expressive 0.22 [0.07, 0.37] .005 4.2 0.15 [-0.00, 0.30] .056 18.7 48m Receptive 0.10 [-0.04, 0.25] .166 1.0 0.05 [-0.10, 0.21] .486 10.8 Expressive 0.14 [-0.01, 0.29] .065 1.8 0.04 [-0.11, 0.19] .613 16.4
2 a Note. CI = confidence interval, RP P in %; P AdjustedP for child gender, birthweight, birth-order,P b treatmentP group, SEIFA index score of disadvantage, parental education, and maternal age; P P
Language measured usingP PLSP -4 at 24 and 36 months, CELF-P2 at 48 months (converted to z- c scores); P βP = effect size: interpret as the average SD difference in language score for one additional d 2 EMST utterance per minute; P RP P P is for the full model, including the variation that is explained by the confounders
6.3.3 Discussion
Maternal use of EMST was associated with better language scores at 24 and 36 months. However, the effect sizes were small according to the criteria specified by J
Cohen (1960), ranging from 0.16 to 0.22, and the associations were no longer
200
statistically significant after controlling for biological and environmental confounders.
There was no evidence that EMST was associated with language outcomes at 48 months in the unadjusted or adjusted models. It is worth noting that the measure changed from the PLS-4 to the CELF-P2 at this data point (see section 4.8.1) although since both assessments measure the same constructs it is unlikely that this explains the finding. It is possible that by 48 months other factors begin to influence language development, over and above early maternal behaviours. Reilly et al. (2010) previously reported that the variation in language abilities explained by environmental factors increased between 24 and 48 months.
Post hoc tests revealed that EMST was positively associated with SEIFA index scores of relative social disadvantage. The association between EMST and child language at 24 and 36 months in the unadjusted model may therefore have been confounded by the SEIFA score. The post hoc test indicated that a one standard deviation higher SEIFA score was associated with a 0.34 standard deviation higher
EMST score (95% CI [0.09, 0.59], p = 0.007). This suggests that mothers living in higher SES areas use higher rates of EMST than mothers who live in lower SES areas.
Future research should explore whether the association between EMST and praise, and child language differs between high, medium and low SES households.
Potentially, low SES households contending with social adversity and deprivation may use emotion socialisation behaviours such as discussing emotion (Eisenberg et al.,
1998) less frequently than higher SES households who may have more resources to do this in terms of time, emotional energy and cultural capital. As Swenson et al. (2016, p.
55) notes, some parents raising children in under-resourced circumstances may feel the need to “‘toughen’ their children to the realities of life”. They may also be parenting in 201
the best way they know how from their own childhood, which could have contained little emotion socialisation or praise. There is an opportunity for health professionals to educate these parents about the importance of providing children with a supportive environment in which to develop the skills required to face the difficult realities of life
(Swenson et al., 2016).
A further consideration is whether EMST, praise and missed opportunities might be associated with different aspects of language proficiency, such as interpreting language use, for example social communication or pragmatic language. Currently researchers are unclear how to define and reliably measure pragmatic language skills, especially at younger ages (Matthews, 2014). Examining how maternal behaviours might be associated with pragmatic language skills would therefore need to be integrated into a wider plan of research.
There are two limitations relating to the data collection and coding that may account for the findings. The first is that the free play task may not have been appropriate to measures different types of emotion and mental state talk. Other studies have used shared book reading tasks, where the content of the book is about emotions
(e.g. Reese et al. 2016). In the current study, the EMST was incidental, and not a focus of the play, so for many parents the free play may not have elicited talk about emotions and needs. See section 9.3.2 for further discussion of this point.
The second limitation relates to how EMST was defined, and is similar to the limitation relating to praise addressed in the paper. The EMST definition may have been too broad, as it combined mental and emotional state words in one code for ease of coding. Previous researchers have used more refined definitions that distinguished
202
between the types of talk. For example, Drummond et al. (2014) distinguished between simple affect talk, desire talk, emotion explanations or elaborations, other internal state terms, mental state terms and empathy statements. In fact, a recent US population-based study of mother-toddler interaction (n = 1,114) concluded that maternal talk about mental states was a parenting component distinct from maternal talk about desires and emotions (Jessee, McElwain, & Booth-LaForce, 2016). These authors draw attention to the specificity principle (Bornstein, 2015) which states that specific parenting behaviours are associated with specific, unique effects on child development. Therefore, combining EMST in one code may have obscured differences that may be important, and may explain why no evidence on an association with language was found in this sample.
Future community-based research should examine EMST using a more detailed coding system, and using a task that requires explicit discussion of emotion.
6.4 Summary
This chapter addressed aspects of Aim Two by examining whether maternal praise, missed opportunities, emotion and mental state talk, successful and unsuccessful directives, and Fluency and Connectedness were associated with language outcomes at
24, 36 and 48 months. The next Chapter completes the presentation of results addressing Aim Two by presenting a paper examining successful and unsuccessful directives, and Supported and Coordinated Joint Engagement.
203
7 Mother-child Interaction and Language
Outcomes: PART TWO
7.1 Overview
The following Chapter presents the final results addressing Aim Two of the thesis. I have been invited to re-submit this paper to the Journal of Child Language for a third-round of peer review. The paper examined Supported and Coordinated Joint
Engagement, coded as part of the second coding scheme. It also examined whether the level of joint engagement modified the associations found between maternal responsive behaviours and child language outcomes found in the earlier Maternal Responsive
Behaviours Study (MRBS), on which this thesis builds (Levickis et al., 2014).
204
Title: The role of joint engagement in the development of language in a community- derived sample of slow-to-talk children
Running head: Joint engagement and language development
205
7.2 Abstract
We explored whether Supported (S-JE) or Coordinated Joint Engagement (C-JE) between 24 month old children who were slow-to-talk at 18 months old, and their mothers were associated with language scores at 24, 36 and 48 months of age (n=195).
Further, whether S-JE or C-JE modified the concurrent associations between maternal responsive behaviours and language scores. Previous research showed that S-JE, maternal expansions, imitations and responsive questions were associated with better language scores. Our main finding was that S-JE but not C-JE was positively associated with 24 and 36 month expressive and receptive language scores, but not with 48 month scores. S-JE modified how expansions and imitations, but not responsive questions, were associated with language scores; the associations were evident in all but the highest levels of S-JE. Further research is necessary to test these findings in other samples before clinical recommendations can be made.
Key words: joint engagement, mother-child interaction, attention, responsivity
7.3 Background
Maternal responsive behaviours used during joint engagement (JE) with their young children predict better language skills (Levickis et al., 2014). JE can vary in duration, quality, and in how the child coordinates their attention (Adamson et al., 2004,
2009). To what extent these JE variations are associated with language outcomes, and whether they modify the effect of maternal responsive behaviours on language development is unclear. Better understanding could inform early language interventions which teach parents to use these responsive behaviours. Since unresolved language
206
difficulties are associated with poorer educational, interpersonal and psychosocial outcomes into adulthood (e.g. Beitchman et al., 2001), improving the efficacy of these interventions is important. The current paper explores the contribution of JE to expressive and receptive language learning both directly and via its effect on maternal responsive behaviours.
Research into maternal behaviours, JE and language development has tended to examine children with typically developing language and with language delay separately. This truncates the distribution of language abilities considered in each study.
The present study focuses on children identified with low expressive vocabulary at an earlier age (18-months compared to 24-months) and below a more liberal vocabulary cut-point (20th compared to 10th percentile) than usual. A large proportion show language scores within the normal range at 24-, 36- and 48-months, which is useful for investigating the associations across a wide distribution of language abilities.
JE is when parent and child are actively focused on the same object or event at the same time and are aware of each other doing so (Adamson & Bakeman, 1991;
Dunham and Dunham, 1995; Moore and Dunham, 1995). JE skills, for example pointing, showing and using eye contact, are positively associated with language skills
(Farrant, Devine, Maybery, & Fletcher, 2012; Kasari, Gulsrud, Freeman, Paparella, &
Hellemann, 2012; Tomasello & Todd, 1983). Subsequently, JE difficulties might contribute to problems with language learning (Mundy, Kasari, Sigman, & Ruskin,
1995). Indeed, children with autism spectrum disorders and late talkers have been found to have difficulties with JE compared with typically developing children (Patterson,
Elder, Gulsrud, & Kasari, 2014; Paul & Shiffer, 1991; Vuksanovic & Bjekic, 2013;
Wetherby, Yonclas, & Bryan, 1989).
207
JE is thought to underpin language acquisition by creating a shared referential framework which helps children make the correct connection between their parent’s spoken word and its referent (Akhtar, 2005; Bruner, 1975; Carpenter et al., 1998;
Tomasello, 2001; Yu & Ballard, 2007). Within the JE context, children can look at the referent for enough time to learn and practise its word, approximately 1-2 seconds before and after it is spoken (Dixon & Salley, 2006; Kannass & Oakes, 2008; Trueswell et al., 2016). Hence, variations in how children coordinate their visual attention during
JE may explain differences in language learning.
During ‘supported’ JE (S-JE) children visually attend to the shared object only, whilst their parents scaffold the interaction, follow their child’s interests, and maintain turns (Adamson et al., 2009). In contrast, during ‘coordinated’ JE (C-JE) children help maintain the interaction by attending to the object and parent, usually by alternating their gaze (Adamson et al., 2009). C-JE is therefore likely to tax children’s cognitive and affective resources more so than S-JE (Adamson et al., 2009), and to reduce the time children can look at a spoken word’s referent. S-JE may therefore provide a better contextual framework for word learning than C-JE. Indeed, one small scale study of typically developing children (n=56) found time in S-JE at 18-months-old was associated with better expressive and receptive vocabulary scores at 30-months, but time spent in C-JE was not (Adamson et al., 2004, 2009). This has yet to be tested in a larger sample or with measures of language skills beyond vocabulary knowledge.
The shared referential framework created in JE gives children the opportunity to benefit from the language-facilitating aspects of maternal responsive behaviours (e.g.
Diaz et al., 1991). Seminal work by Tomasello and Farrar (1986) demonstrated that maternal verbal input used inside JE at 15-months was correlated with language abilities 208
at 21-months, whilst input used outside of JE was not. Previous findings from the sample in the present study showed that maternal responsive behaviours (expansions, imitations and responsive questions) used during JE at 24-months were positively associated with concurrent child receptive and expressive language scores (Levickis et al., 2014). It is feasible that these associations between responsive behaviours and concurrent language skills might be modified by the JE state in which mother and child are engaged. Specifically, children’s sustained attention during S-JE may enable them to benefit more from the maternal responsive behaviours than when they are alternating their attention during C-JE. However, to date whether JE state modifies the strength of the association between maternal behaviours and child language skills has not been tested.
7.3.1 Purpose of the Current Investigation
The current investigation aimed to answer two questions in a community- derived sample of children who were slow-to-talk at 18 months.
1. Are Supported and Coordinated Joint Engagement at 24 months old associated
with child receptive and expressive language scores at 24, 36 and 48 months-old?
2. Does the level of S-JE or C-JE modify the association between maternal
expansions, imitations and responsive questions and 24 month old language
outcomes?
We anticipated that S-JE would be associated with language outcomes at 24, 36 and
48 months, but that there would be no such evidence for C-JE. We examined
receptive and expressive language separately to explore whether JE is associated
with both domains of language.
209
7.4 Method
7.4.1 Participants
The study was nested within a cluster randomized-controlled trial based in a population-level survey, Let’s Learn Language (LLL) (NHMRC #384491) (Wake et al.,
2011) and its follow-up, Language for Learning (L4L) (NHMRC #60740) (Wake et al.,
2012). Recruitment is described elsewhere (Wake et al., 2012; Wake et al., 2011).
Briefly, participants were recruited at their 12 month check-up, offered universally to families in Victoria, Australia. Exclusion criteria were if the child had developmental delay, suspected autism spectrum disorder, or a major medical condition, or if parents had insufficient English to complete written questionnaires at grade six reading level
(typically age 11 to 12).
Parents completed a baseline questionnaire collecting demographic information.
Six months later, at age 18 months, they completed a screening survey for expressive vocabulary skills, the Sure Start Language Measure (Roy et al., 2005). Children scoring at or below the 20th percentile were invited into an RCT for a low-intensity parent- toddler language promotion programme (n = 301). There were no differences evident in language outcomes between the RCT intervention and comparison groups (Levickis et al., 2014), so the participants are analysed together as a single group in the current study, with all adjusted analyses controlling for intervention group status.
At 24 months, 251 mothers agreed to be video-recorded with their child. There were no differences in demographic characteristics between participants who were and were not video-recorded (Levickis et al., 2014). Of the 251 participants, data were
210
available for 197 participants in the current study; 49 participants did not give consent for their data to be used in other studies, a further four were excluded after receiving an autism spectrum disorder diagnosis at age three or four years, and another was excluded because a different caregiver took part. There was no evidence that the participants differed from the 251 for whom we had video-recordings except for maternal age (1.3 years older in this sample).
As can be seen in Table 12, about half of the participants took part in the intervention, half were boys, and nearly half of the parents had completed further education. The sample was more socially advantaged than the Australian population on average, evident by a Socio-Economic Indices for Area (SEIFA) score of disadvantage higher than the Australian mean (M = 1000, SD = 100) (Statistics, 2001). As shown in
Table 12, although these children scored at or below the 20th percentile for expressive vocabulary at 18-months, their language skills, as a group, largely normalised as indicated by means near the normative mean of 100 on face-to-face language assessments at 24, 36 and 48 months for receptive and expressive language scores (see
Measures section below for further details). Moreover, Table 12 shows that only a minority of children scored 1.25 standard deviations below the mean, a typical cut-off for identifying language impairments in research (Reilly et al., 2010; Tomblin et al.,
1997), at 24, 36 and 48 months for either receptive or expressive language.
211
Table 12 Sample Characteristics and Language Scores Total Sample characteristics n (%) or M (SD) n
197 103 (52.28) Gender, male First-born child 197 73 (37.06) Birthweight (g) 190 3380.3 (620.2) a SEIFA index score of disadvantage P 197 1026.5 (53.3) Maternal age at child age 12-months (years) 197 34.31 (4.5) In intervention arm of trial 197 100 (50.8) Parent education: Not complete high school 38 (19.4) Completed high school 65 (33.2) Completed diploma/tertiary qual/postgrad 93 (47.5) Low language b c Child language scoresP n M (SD) n (%)P P Receptive language score: 24 months 195 90.58 (14.0) 63 (32.3) 36 months 190 98.44 (15.2) 24 (12.7) 48 months 193 94.66 (14.0) 33 (17.1) Expressive language score: 24 months 195 91.76 (11.9) 27 (13.9) 36 months 189 101.10 (14.5) 16 (8.5) 48 months 192 97.66 (14.3) 26 (13.5)
a b Note: P P SEIFA = Socio-Economic Indices for Area; P P Preschool Language Scale - 4th Edition at 24 and 36 months, and Clinical Evaluation of Language Fundamentals - Preschool 2nd Edition at 48 c months; P P Low language defined as scoring ≥ 1.25 SD below the standardised mean on expressive and/or receptive language tests
7.4.2 Procedure
Parents completed postal questionnaires when their children were 24, 36 and 48 months old. At 24 months old, research assistants who were blind to the children’s treatment group status, conducted the video-recording of mothers and their children in their homes. Mothers were asked to play with their child as they normally would for 15-
212
minutes using two sets of toys (farmyard and accessories, and doll and accessories).
This was recorded using Hitachi DZ-GX5060SW DVD camcorders and recorded onto
Sony DVD DS DVD-RW discs. The language assessment was administered at this 24 month visit, and again at home visits when the children were 36 and 48 months old.
In an earlier study, the video recordings were uploaded into Observer® XT coding software (Noldus, 2008), and the second author coded maternal responsive behaviours using a continuous coding methodology (Levickis et al., 2014). There were no differences between the control and intervention groups on maternal responsive behaviours, which suggests that the intervention did not have an effect on maternal use of responsive behaviours in the intervention arm of the trial. For the current study, the middle five minutes were observed in Windows Media Player and rated for S-JE and C-
JE. These ratings were entered into an excel spreadsheet by the first author.
7.4.3 Measures
7.4.3.1 Supported and Coordinated Joint Engagement
S-JE and C-JE were rated on a seven-point Likert Scale using two items adapted from Adamson et al. (2012), as shown in Table 13. The coder (first author) noted the start and end time of each episode when the mother and child were actively involved with the same object or activity for at least three seconds. The coder noted whether the
JE was supported or coordinated, the latter evidenced by the child looking at or speaking to the mother. The coder also made note of the quality of the interaction by considering the matched affect, energy and intensity between mother and child, as detailed in the technical manual developed by Adamson et al. (2012). Finally, the coder calculated the total time in S-JE and C-JE to use to select a point on the rating scale.
213
Each point on the rating scale corresponded to an approximate JE duration (in seconds), calculated based on the five minute observation. A score of one indicated there were no
JE episodes, four indicated that mother and child were in moderate-quality JE for about a third of the time or briefly in high-quality JE, and seven indicated the dyad being frequently in rich and varied JE episodes (Adamson et al., 2012). The rating was increased by one scale point if the quality was appraised to be high.
To test agreement, 10% of the samples were independently rated by a second coder (author #3), blind to the first coder’s rating. Unweighted kappas (J Cohen, 1960), were 0.47 and 0.66 for S-JE and C-JE respectively. However, no disagreements exceeded one scale point on the 1-7 rating scale. Consequently, weighted kappas
(Cohen, 1968) that counted one-point disagreements as agreements were 1.0 for S-JE and C-JE. 10% were also re-rated by the primary coder. Intra-rater unweighted kappas were 0.77 and 0.74 for S-JE and C-JE respectively, and weighted kappas were 1.0 for both.
214
Table 13 Joint Engagement States Coding Scheme Joint engagement (JE) Mother and child are actively attending to the same object or event with sustained interest for minimum 3 seconds. (Adamson et al., 2012; Active attending is usually evidenced by manipulation of the object, or being actively focused on the shared Hirsh-Pasek et al., 2015) activity while one partner manipulates the object. Defined as either supported or coordinated. Supported Child is focusing almost exclusively on the shared object or event (but beyond just listening). Caregiver’s participation influences the child’s activity/experience with object/event but the child does not acknowledge this involvement. The caregiver is often actively manipulating object or making statements/commands that alter child’s actions. Coordinated Child is coordinating attention from object or event to the caregiver. Child is acknowledging the caregiver, often with repeated glances to their face. The caregiver’s level of activity directly on the object may be minimal. Rating scale Describes both the quantity and quality of the child’s time in joint engagement
Level 1= 2= 3= 4= 5= 6= 7=
Approx. duration (secs) 0 15 30 90 120 150 >200secs
Rating definition No episodes of 1/3 of time spent Frequently in rich JE in moderate and varied JE quality, or episodes. briefly in highly striking manner
215
7.4.3.2 Language Measures
The Preschool Language Scale 4th edition (PLS-4) (Zimmerman et al., 2002) was used at 24 and 36 months-old and the Clinical Evaluation of Language
Fundamentals Preschool Second edition (CELF-P2) (Semel et al., 2006) at 48 months.
Subscales from both yield norm-referenced expressive and receptive language standard scores with a mean (M) of 100 and SD of 15. To more readily compare the different measures in analysis, the scores were rescaled to z-scores (M = 0, SD = 1).
7.4.3.3 Maternal Responsive Behaviours
The maternal responsive behaviours that were the focus of this study, described in Table 14, were expansions (where the mother repeated and added to the child’s vocalisation), imitations (where the mother repeated the child’s vocalisation or verbalisation) and responsive questions (asking wh- questions that were immediate and dependent on child’s preceding act). These were chosen as all were shown to be associated with better language scores at 24 and 36 months in Levickis et al. (2014).
The analysis used rate per minute to control for variation in the recording durations.
216
Table 14 Definitions of Maternal Responsive Communicative Behaviours Behaviour Definition Example Expansion Mother repeats one or all the child’s preceding C: “Ball” words and adds to the child’s preceding M: “It’s a red ball” verbalisation Imitation Mother repeats the child’s preceding vocalisation C: “Ball”, or verbalisation exactly or with a reduction in the M: “Ball” words. Responsive Mother asks a “wh” question (e.g. “what”, C: Playing with horse Question “when”, “who”), which is immediate and M: “What’s that?” dependent on the child’s preceding act.
Note: Definitions from Maternal Responsive Behaviours Study (Levickis et al., 2014)
7.4.3.4 Demographics
Child, maternal, and environmental information identified as potential confounders a priori were collected in the baseline questionnaire. These were maternal age, parental education, birth order (first or later born), birthweight (as a proxy for biological risk), and gender. Neighbourhood disadvantage was measured using the
Socio-Economic Indexes for Areas (SEIFA) score of disadvantage, which is calculated using census data about household education, employment and occupation, income, and composition. (Statistics, 2001). Higher scores indicate less social disadvantage.
7.5 Results
As shown in Table 15, on average, the participants were engaged in S-JE for at least 90 to 100 secs out of the total 300 secs (mean rating = 4.1), compared to C-JE in which they were engaged for less than 30-secs (mean rating = 2.7). This reveals that as a group, the children were more likely to be visually attending to only the shared object whilst playing with their mother, than they were to be alternating their gaze between the 217
object and their mother. There was no evidence of any differences in time spent in S-JE or C-JE between participants who were in the intervention arm of the RCT or those in the control group.
Table 15 Joint Engagement and Maternal Responsive Behaviours Summary Approximate Interactive or Maternal Behaviour n M (SD) duration (secs) Supported Joint Engagement (S-JE) 195 4.1 (1.5) 90-100 Coordinated Joint Engagement (C-JE) 195 2.7 (1.3) 20-30 Previously coded responsive behaviours
(Levickis et al., 2014) Expansion 197 0.6 (0.6) Imitation 197 0.5 (0.5) Responsive question 197 0.7 (0.5)
The correlation matrix in Table 16 shows that higher quality or more time spent in S-JE was moderately correlated with better receptive and expressive language scores at 24, 36 and 48 months. As anticipated, there was no consistent evidence for C-JE being correlated with language scores.
218
Table 16 Correlation Matrix of Responsive Behaviours, Joint Engagement, and Language Scores
Variable 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 1. Expansion
2. Imitation .58**
3. Responsive Question .25** .28**
4. S-JE .29** .22* .26**
5. C-JE .25** .23* .06 .11
6. 24m Receptive .35** .24** .15* .39** .14
7. 24m Expressive .40** .30** .18* .29** .03 .61**
8. 36m Receptive .36** .17* .20* .35** .16* .67** .55**
9. 36m Expressive .41** .22* .19* .34** .16* .67** .55** .80**
10. 48m Receptive .31** .14 .11 .26** .12 .59** .41** .71** .68**
11. 48m Expressive .34** .13 .14 .20* .06 .50** .40** .67** .74** .74**
Note: Items 1 to 3 are rate per minute, items 4 and 5 are ratings on 7-point Likert scale; Language skills assessed using Preschool Language Scale 4th Edition at ages 24 and 36 months, and Clinical Evaluation of Language Fundamentals Preschool 2nd Edition at 48 months; * p<0.05, ** p<0.001
.
219
7.5.1 Question 1: S-JE and C-JE and Language Scores
The first research question asked whether S-JE and C-JE were associated with receptive and expressive language scores at 24, 36 and 48 months-old. After checking that the assumption of linearity was met by comparing models fitting S-JE and C-JE as continuous to models fitting them as categorical (divided into quartiles), linear regression modelling was used. Multivariate models were run adjusting for potential confounders (gender, birthweight, birth order, SEIFA score of disadvantage, parental education, maternal age, RCT treatment group). Finally, to see whether the variability in
36 and 48 months language scores explained by S-JE or C-JE was independent of the child’s earlier language skills, 24 month language scores (expressive or receptive as appropriate) were added to the models (as Adamson et al., 2009).
As shown in Table 17, more time spent in S-JE was associated with better receptive and expressive language scores at 24 and 36 months after controlling for confounders. Language scores were approximately 0.2 SD higher for every one-point higher S-JE rating. Moreover, Table 17 shows that S-JE was associated with change in language scores from 24 to 36 months, although the magnitude of this effect was very small. There was no evidence that the independent association between S-JE and child language continued to age 48 months. As can be seen in Table 17, there was also no consistent evidence that C-JE was associated with language scores at 24, 36 or 48 months. Although C-JE was positively associated with language scores at 36 months in the unadjusted models, the effect sizes were < 0.2, and the associations did not hold once confounders were added.
220
Table 17 Associations between Joint Engagement State and Language Scores a b b Age Domains P Unadjusted Adjusted P Adjusted P P + 24 month language c 2 c 2 c 2 S-JE β P 95% CI p RP β P 95% CI p RP β P 95% CI p RP 24 Receptive 0.25 [0.17, 0.34] <.001 0.15 0.22 [ 0.13, 0.30] <.001 0.25 Expressive 0.19 [0.10, 0.28] <.001 0.08 0.17 [ 0.08, 0.26] <.001 0.20
36 Receptive 0.23 [0.14, 0.32] <.001 0.12 0.20 [ 0.11, 0.29] <.001 0.26 0.08 [-0.00, 0.16] .05 0.52 Expressive 0.22 [0.13, 0.31] <.001 0.11 0.19 [ 0.10, 0.28] <.001 0.25 0.12 [ 0.04, 0.21] .01 0.4
48 Receptive 0.17 [0.08, 0.26] <.001 0.07 0.15 [ 0.05, 0.24] .003 0.15 0.04 [-0.05, 0.12] .41 0.38 Expressive 0.13 [0.04, 0.22] .006 0.04 0.07 [-0.02, 0.17] .124 0.18 0.03 [-0.06, 0.12] .542 0.28
C-JE 24 Receptive 0.10 [-0.00, 0.20] .061 0.02 0.02 [-0.08, 0.13] .694 0.14 Expressive 0.02 [-0.08, 0.12] .695 0.00 -0.07 [-0.17, 0.04] .229 0.15
36 Receptive 0.12 [ 0.01, 0.22] .031 0.02 0.05 [-0.06, 0.16] .389 0.18 0.03 [-0.05, 0.11] .49 0.51 Expressive 0.11 [ 0.01, 0.22] .034 0.02 0.04 [-0.06, 0.15] .429 0.18 0.07 [-0.02, 0.17] .129 0.38
48 Receptive 0.09 [-0.02, 0.19] .108 0.01 0.05 [-0.06, 0.16] .355 0.11 0.05 [-0.05, 0.14] .347 0.38 Expressive 0.04 [-0.06, 0.15] .434 0.00 -0.02 [-0.13, 0.09] .741 0.17 0.01 [-0.10, 0.11] .90 0.28
a Note: n = 195; P P Preschool Language Scale-4th Edition at 24 and 36 months and Clinical Evaluation of Language Fundamentals-Preschool 2nd Edition at b 48-months, scores rescaled to M=0, SD=1; P P Adjusted for child gender, birthweight, birth-order, treatment group, SEIFA disadvantage score, parent c education, maternal age; P P β = effect size: interpret as the average SD difference in language score for one-point higher rating on S-JE or C-JE scale
221
7.5.2 Question 2: JE as an Effect Modifier
The second question asked whether S-JE or C-JE modified the association between three maternal responsive behaviours (expansions, imitations, and responsive questions) and concurrent child language outcomes (Levickis et al., 2014). Since part one revealed no evidence that C-JE and language scores were associated, only S-JE continued to be explored.
First, an ‘S-JE by Maternal Responsive Behaviour’ interaction term was incorporated into the 24 months language and maternal responsive behaviour regression models and compared to models excluding the interaction terms, using likelihood ratio tests. As shown in Table 18, only one model for the association between expansions and receptive language supported the interaction term, showing the association was weaker when more time was spent in S-JE. Although a similar pattern can be observed in Table
18 between expansions and expressive language, and between maternal imitations and receptive and expressive language, the interaction terms were not supported. However, because standard tests of interaction can miss subtle effects (Kirkwood & Sterne, 2013), simple slopes were fitted (UCLA, 2015) to explore the question further
222
Table 18 S-JE Interaction Effects on Association between Responsive Behaviours and Language a Maternal behaviour Unadjusted Adjusted P b c b c β.P 95% CI p P β.P 95% CI p P d Receptive language P Expansion -0.12 [-0.24, -0.00] .047 -0.10 [-0.23, 0.02] .097 Imitation -0.06 [-0.22, 0.10] .476 -0.05 [-0.21, 0.11] .501 Responsive Question -0.02 [-0.17, 0.13] .811 -0.03 [-0.20, 0.14] .720 d Expressive language P Expansion -0.10 [-0.22, 0.02] .109 -0.07 [-0.19, 0.06] .306 Imitation -0.13 [-0.30, 0.03] .101 -0.14 [-0.30, 0.02] .079 Responsive Question 0.01 [-0.14, 0.17] .886 0.04 [-0.13, 0.21] .638
a Note: CI = confidence interval. P P Adjusted for child gender, birthweight, birth-order, treatment b group, SEIFA score of disadvantage, parental education, and maternal age;P P interaction c coefficient; P pP -value for likelihood ratio tests comparing regression models with and without d the interaction term. P P Language measured using PLS-4, scores rescaled to M = 0, SD = 1
Figure 4 plots the estimated 24 month expressive and receptive language scores by maternal expansions, imitations and responsive questions at each S-JE rating (one to seven). The positive associations with both maternal expansions and imitations were observed in all but the highest S-JE ratings. The strength of the associations was weaker for higher S-JE ratings, which can also be seen in the coefficients in Table 19. Unlike the other two maternal responsive behaviours, the simple slopes for responsive questions did not support the hypothesis that their association with language scores would differ by S-JE rating. There were no S-JE levels at which a statistically significant association could be observed (Table 19).
223
Figure 4 Associations between Maternal Behaviours and Language Scores at 24 months by S-JE
224
Table 19 Simple Slope Statistics for Association between Maternal Behaviours and Language Scores by S-JE a a S-JE rating Expressive language P Receptive language P b b β P 95% CI p β P 95% CI p Expansions 1 0.92 [ 0.41, 1.43] <.001 0.90 [ 0.40, 1.40] .001 2 0.82 [ 0.42, 1.22] <.001 0.77 [ 0.38, 1.17] <.001 3 0.72 [ 0.42, 1.02] <.001 0.65 [ 0.36, 0.95] <.001 4 0.62 [ 0.39, 0.85] <.001 0.53 [ 0.30, 0.76] <.001 5 0.52 [ 0.30, 0.73] <.001 0.41 [ 0.20, 0.62] <.001 6 0.42 [ 0.15, 0.68] .002 0.29 [ 0.02, 0.55] .033 7 0.32 [-0.03, 0.67] .076 0.16 [-0.19, 0.51] .358 Imitations 1 0.97 [ 0.33, 1.61] .003 0.56 [-0.07, 1.20] .079 2 0.84 [ 0.34, 1.33] .001 0.51 [ 0.02, 1.0] .043 3 0.70 [ 0.33, 1.07] <.001 0.45 [ 0.08, 0.82] .017 4 0.57 [ 0.28, 0.86] <.001 0.39 [ 0.11, 0.68] .007 5 0.44 [ 0.16, 0.72] .002 0.34 [ 0.06, 0.61] .018 6 0.30 [-0.05, 0.66] .097 0.28 [-0.08, 0.63] .122 7 0.17 [-0.31, 0.65] .489 0.22 [-0.25, 0.69] .359 Responsive questions 1 0.16 [-0.44, 0.75] .604 0.15 [-0.43, 0.73] .616 2 0.17 [-0.29, 0.63] .474 0.13 [-0.32, 0.58] .570 3 0.18 [-0.17, 0.52] .306 0.11 [-0.22, 0.45] .513 4 0.19 [-0.08, 0.46] .162 0.09 [-0.17, 0.35] .480 5 0.20 [-0.07, 0.47] .141 0.08 [-0.19, 0.34] .570 6 0.21 [-0.13, 0.56] .229 0.06 [-0.28, 0.40] .738 7 0.22 [-0.24, 0.69] .344 0.04 [-0.41, 0.49] .865
a b Note: S-JE Supported Joint Engagement; P P Language measured at 24 months using PLS-4; P P β = Estimated effect size for the association between the maternal behaviour and 24 month language score at S-JE level
.
225
Supplementary Figure 5 Expressive and Receptive Language Score Distributions Receptive Language Scores Expressive Language Scores
24 months 24 months
36 months 36 months
48 months 48 months
Note: PLS-4 at 24 and 36 months, CELF-P2 at 48-months; Standard scores, M = 100, SD = 15
226
7.6 Discussion
The purpose of this study was to explore: (1) whether S-JE and C-JE in mother- child interaction at 24 months were associated with child language scores at 24, 36 and
48 months old, and; (2) whether S-JE and C-JE modified the positive cross-sectional associations previously found between maternal expansions, imitations and responsive questions and child language scores in the Maternal Responsive Behaviours Study
(Levickis et al., 2014). S-JE was positively associated with 24 and 36 month language scores, but not with 48 month scores. There was no evidence that C-JE was consistently associated with language scores at 24, 36 or 48 months. S-JE modified how expansions and imitations were associated with cross-sectional language scores with the associations being evident in all but the highest levels of S-JE. In contrast, there was no evidence that S-JE modified the association between responsive question and language abilities. These findings will be considered in turn, followed by a broader discussion of their contribution to the literature.
7.6.1 Question 1: The Association Between JE and Language Scores
These findings expand upon previous research reporting that S-JE, and not C-JE, at 18 months old was associated with later expressive vocabulary (Adamson et al.,
2004; Adamson et al., 2009). S-JE continues to be important to language development at 24 months, and is associated with concurrent language skills as well as skills one year later. These findings also suggest that time spent in S-JE may be important to broader domains of language development than just word learning, including comprehension and expressive language more generally. This might be because during S-JE, children can use their attentional resources to monitor more than verbal labels, including other
227
aspects of the social interaction. Compared to C-JE, whilst in S-JE children may have more time be able to process, consolidate, and learn from their experiences (Hirsh-Pasek et al., 2015). As such, time in S-JE may provide the optimal language-learning environment, at least up to age 36 months.
The finding that S-JE was no longer associated with expressive or receptive language scores at 48 months old may indicate that other child, family or environmental factors become more important to later language development. For example, social disadvantage reportedly becomes more influential in language development between the ages of two and four, perhaps due to an accumulated effect of exposure to poorer language input (Reilly et al., 2010). Perhaps the null finding at 48 months could be accounted for by a developmental or environmental change that is common to pre- schoolers. For example, most Australian children start formal kindergarten when they are four years-old, increasing their exposure to interactions with non-family members, e.g. teachers, peers. These interactions may be important to on-going language learning and may become more influential than earlier mother-child S-JE, therefore accounting for greater variability in children’s language skills at 48 months.
As hypothesised, C-JE was not associated with language outcomes. However, it does leave unanswered the question of what role C-JE may have in language development (Adamson et al., 2009). After all, young children who use frequent eye contact and talk directly to their caregivers, both indicators of C-JE, are engaging and rewarding to interact with. An engaged and motivated adult is more likely to continue interacting with that child, providing verbal input and feedback important for language development. Perhaps C-JE is associated with other aspects of communication or social
228
development that were not examined in this study, such as pragmatic language or social cognition. This is yet to be investigated to our knowledge.
7.6.2 Question 2: JE as an Effect Modifier
The second aim of this paper was to examine whether S-JE or C-JE modified the associations previously found in this sample between maternal responsive behaviours and language outcomes (Levickis et al., 2014). Whilst C-JE was not found to be associated with language outcomes, S-JE was observed to modify the association between maternal expansions and imitations and concurrent language outcomes.
However, there was no evidence that S-JE modified the association between 24 month language scores and responsive questions.
Maternal expansions and imitations were positively associated with receptive and expressive language scores at almost all but the highest two levels of S-JE. At these highest ratings, mother-child dyads are frequently in rich and varied S-JE where maternal behaviours may not have the same potential to influence an already optimal language-learning environment. The simple slopes illustrate that even though the association between the expansions and imitations and child language scores was observed at most levels of S-JE, the strength of the association was greatest at the lower ratings, when the child spends less time in S-JE. Where S-JE is less frequent or of lower quality, expanding upon or imitating a child’s utterance, may be especially important.
There was no evidence that S-JE modified the association between language scores at 24 months and responsive questions. This might be an artefact of the coding protocol for S-JE and C-JE. A responsive question necessitates a response from the child. The child’s response may lead to a transition in engagement state. For example, if 229
the child answers the question whilst continuing to play with the object and glancing at or responding to their mother, their engagement state will become ‘co-ordinated’, and the S-JE episode will end. If the child does not respond the mother might interpret this as the child lacking interest or attention, and terminate the S-JE episode. Therefore, the reason why an effect modification was not detected could be that regardless of the child’s level of S-JE, maternal responsive questions usually result in the S-JE episode coming to an end.
An alternative explanation could be that the association between some maternal behaviours and child language is more stable than others, and less likely to be modified by the interactional context in which they are used. For example, the degree to which the maternal behaviour is dependent on the child’s preceding behaviour, and hence stage of language development, might be important. Expansions and imitations can only be used in response to a child’s preceding verbalisation or vocalisation. In comparison, responsive questions can be asked about anything the child is attending to, regardless of the child’s verbalisations or vocalisations. Therefore, use of responsive questions depends less on the child’s behaviour, and more on the mother’s ability to create opportunities to initiate or maintain the conversation. It might be maternal responsive behaviours that proactively create these conversational opportunities which have the more stable association with language development regardless of the interactional context. Further exploration of the characteristics of maternal behaviours is required to explore these ideas further.
230
7.6.3 Strengths and Limitations
Study strengths include the large sample size, the prospective, community-based design, the observational measure of S-JE and C-JE, and repeated face-to-face language assessments. Observational and direct measures can provide a more objective measure than parent report (Hayden, Durbin, Klein, & Olino, 2010). Recording the mother-child playing in the home may also have increased our chances of capturing a more naturalistic interaction. Finally, the concurrent measures of language skills and joint engagement enabled analytical adjustment for the effects of language ability at the time of the mother-child interaction.
The sample of children in this study were those with expressive vocabulary
th scores below the 20P P percentile at age 18 months in a larger community sample of children. It is possible that the parent-child interactions may have been qualitatively different from interactions between parents and children who did not have low expressive vocabulary at 18 months. Caution is therefore required in generalising the findings to other populations. Socially disadvantaged households were under- represented in our sample, despite efforts to recruit from high, medium and low socio- economic areas. This is often the case in longitudinal studies (Kiernan & Mensah,
2009), and may mean that the full range of parent-child interactions was not captured, particularly those disproportionately affected by family disadvantage. The change in language measure at 48 months, from the PLS-4 to the CELF-P2, means there is the potential for introduced non-equivalence. The change in association between S-JE and language scores at 48 months may be due to the change in language measure.
231
Finally, the quantity of child directed speech heard by the children during the interaction and during each JE episode was not calculated. It was therefore not possible to quantify the strength of the associations between JE, maternal behaviours and child language scores over and above maternal input quantity and diversity (McGillion, Pine,
Herbert & Matthews, 2017).
7.6.4 Future Directions
Further investigation is necessary before drawing clinical recommendations from this study. However, possibilities include trialling a version of the S-JE scale to use within clinical assessment to identify a child’s abilities to participate in interactions that are optimal for language-learning. Investigating how S-JE and C-JE are associated with existing measures of child attentional skills to determine what additional skills are being captured by the JE ratings over and above attention skills. Understanding which maternal behaviours are modified by S-JE and which are not, might be useful for clinicians tailoring individual intervention strategies for young children. Finally, these findings suggest that the supported, or scaffolded, characteristics of joint engagement may be especially important to language learning at 24 months, rather than just the joint engagement alone. By having little responsibility for orchestrating turns or monitoring their mother’s interests, the child can take full advantage of the shared attention for language learning. Greater understanding of this scaffolding might inform language promotion approaches to coaching caregivers on how best to support their toddlers during interactions.
232
7.7 Conclusions
This investigation has brought together two separate but related strands of research into child language development; research into the parent-child interactional context, and research into specific types of maternal input. It found that time spent in S-
JE at 24 months made an important contribution to the language-learning environment up to age 36 months, and to the beneficial outcomes associated with some maternal responsive behaviours used within this environment. Future research should progress this approach with the intention of improving the efficacy of early language intervention strategies.
7.8 Summary
This chapter completed the investigation of Aim Two of the thesis. Along with
Chapter Six, the findings revealed that Fluency and Connectedness, Supported Joint
Engagement, and successful and unsuccessful directives were associated with child language outcomes during the period 24 to 48 months. The next Chapter addresses the third and final aim of the thesis. That is, whether the maternal and interactive behaviours examined in relation to children’s language development were also associated with children’s SEB scores at 24, 36 and 48-months.
233
8 Mother-child Interaction and Social,
Emotional and Behavioural Outcomes
8.1 Overview
This Chapter addressed Aim Three of the thesis, namely whether behaviours used in mother-child interactions at age 24 months are associated with concurrent or later child internalising and externalising behaviour scores. Up to this point, the thesis has examined how language and SEB abilities are associated in this sample, and how maternal and interactive behaviours are associated with language outcomes. This
Chapter draws together the thesis by examining whether these same behaviours are associated with SEB outcomes. Unlike the previous three chapters, which are manuscripts, this Chapter presents only results and the related discussion. The relevant literature can be found in the Literature Review Chapters (sections 2.10, 3.5 to 3.10), and the methods and analysis plan in Chapters Four and Five.
234
8.2 Introduction
The maternal and interactive behaviours examined were maternal praise, missed opportunities, emotion and mental state talk (EMST), successful and unsuccessful directives, Supported Joint Engagement (S-JE), Coordinated Joint Engagement (C-JE), and Fluency and Connectedness (FC). In addition, the six maternal responsive behaviours coded in the Maternal Responsive Behaviours Study (MRBS) were examined, that is, expansions, imitations, labels, responsive questions, supportive directives, and interpretations (Levickis et al., 2014).
Internalising and externalising behaviours were measured by parent-report using the Child Behaviour Checklist (Achenbach & Rescorla, 2000) at 24 and 36 months and the Strengths and Difficulties Questionnaire (Goodman, 2001) at 48 months (see section
4.8.2). Internalising difficulties at these young ages include anxiety and separation anxiety (e.g. expressing fear or worry over little things, being clingy to a parent), depression (e.g. being unhappy, sad, fearful), being withdrawn (e.g. refusing to play, avoiding eye contact, moving away), and somatic complaints (e.g. having stomach aches, headaches, vomiting) (Bayer et al., 2006). Externalising difficulties at this age include aggression (e.g. screaming, being mean to others, destroying things) and inattention (e.g. can’t sit still or concentrate, failing to finish, fidgeting).
8.3 Results
Table 20 presents the results of the regression analyses investigating associations between the maternal and interactive behaviours and SEB scores at all three time-points. There was no evidence that praise or EMST were associated with
235
internalising or externalising behaviour scores. For example, the estimated mean difference in 24 month internalising behaviour score for each additional use of praise per minute was -0.07, but the 95% confidence interval (CI) within which the true value existed was wide and spanned zero (CI -0.28 to 0.14, p = .502).
Similarly, there was no evidence that successful directives were associated with internalising or externalising difficulties at the three ages (e.g. 48 month internalising behaviour score, β = 0.07, CI -0.34 to 0.49, p = .723). In contrast, there was evidence of a positive association between unsuccessful directives and internalising behaviours at
24 and 36 months. At 24 months, one additional unsuccessful directive per minute was associated with half a standard deviation higher internalising problem score (β = 0.54,
CI 0.20 to 0.89, p = .002). At 36 months, one additional unsuccessful directive per minute was associated with 0.42 SD higher internalising problem score (β = 0.42, CI
0.07 to 0.76, p = .019). By 48 months, there was no evidence of this association (β =
0.13, CI -0.22 to 0.48, p = .458). There was no evidence that externalising behaviour scores were associated with unsuccessful directives at any age.
Regarding missed opportunities, at 48 months they were associated with lower externalising problem scores (β = -0.40, CI -0.79 to -0.00, p = .050). For each additional missed opportunity per minute, externalising problem behaviours were 0.40 SD lower.
However, at 24 and 36 months there was no evidence for this association (e.g. 24 month externalising behaviour, β = -0.06, CI -0.47 to 0.35, p = .762; 36 month externalising behaviour, β = -0.21, CI -0.59 to 0.18, p = .285).
There were no consistent findings that any of the maternal responsive communicative behaviours previously coded (Levickis et al., 2014) were associated
236
with either internalising or externalising behaviour scores. There were three statistically significant findings among the 36 comparisons made: expansions were associated with externalising behaviour problems at 24 months old (β = 0.27, CI 0.01 to 0.53, p = .039); labels with internalising behaviour problems at 36 months (β = 0.22, CI 0.00 to 0.44, p
= .048), and supportive directives with externalising behaviours at 36 months (β = -
0.37, CI -0.69 to -0.05, p = .023). However, in light of the cautious approach being taken for interpreting the findings, these were not considered convincing.
Turning to the association between the interactive communicative behaviours and SEB scores, there was no evidence that S-JE, C-JE or FC were associated with either internalising or externalising behaviour scores at 24, 36 or 48 months.
237
Table 20 Association between Maternal and Interactive Behaviours and SEB Scores
a a Behaviour Age Internalising behaviours P Externalising behaviours P (months) β [95% CI], p β [95% CI], p
b Maternal behaviours coded for MIBS (this thesis) (rate per minute, rpm) P
Praise 24 -0.07 [-0.28, 0.14], .502 -0.02 [-0.23, 0.19], .861 36 -0.12 [-0.33, 0.10], .285 -0.04 [-0.25, 0.17], .695 48 -0.07 [-0.28, 0.13], .480 -0.03 [-0.24, 0.19], .812
Missed opportunity 24 -0.08 [-0.48, 0.33], .702 -0.06 [-0.47, 0.35], .762 36 0.1 [-0.30, 0.50], .615 -0.21 [-0.59, 0.18], .285 48 -0.32 [-0.71, 0.08], .115 -0.40 [-0.79, -0.00], .050
Emotion and mental 24 0.01 [-0.15, 0.17], .879 0.01 [-0.15, 0.17], .916 state talk 36 -0.05 [-0.21, 0.11], .503 0.04 [-0.11, 0.20], .573 48 0.01 [-0.15, 0.16], .944 0.01 [-0.15, 0.16], .944
Successful directive 24 0.21 [-0.21, 0.64], .325 0.1 [-0.33, 0.54], .632 36 -0.04 [-0.46, 0.37], .834 0.04 [-0.37, 0.44], .860 48 0.07 [-0.34, 0.49], .723 -0.07 [-0.49, 0.35], .733
Unsuccessful 24 0.54 [0.20, 0.89], .002 0.27 [-0.09, 0.63], .135 directive 36 0.42 [0.07, 0.76], .019 0.27 [-0.07, 0.60], .125 48 0.13 [-0.22, 0.48], .458 0.22 [-0.13, 0.57], .218
b Maternal responsive behaviours coded in MRBS (rpm) P
Expansion 24 -0.01 [-0.27, 0.24], .923 0.27 [0.01, 0.53], .039 36 -0.06 [-0.32, 0.19], .622 0.09 [-0.16, 0.34], .473 48 -0.06 [-0.31, 0.19], .624 0.07 [-0.18, 0.32], .590
Label 24 -0.06 [-0.28, 0.17], .609 -0.11 [-0.34, 0.11], .325 36 0.22 [0.00, 0.44], .048 0.09 [-0.12, 0.31], .395 48 0.21 [-0.01, 0.43], .063 -0.04 [-0.26, 0.19], .746
Responsive 24 0.06 [-0.23, 0.36], .665 -0.03 [-0.32, 0.27], .849 Question 36 0.04 [-0.25, 0.33], .787 0.05 [-0.22, 0.33], .701 48 0.09 [-0.20, 0.38], .536 -0.08 [-0.37, 0.21], .592
238
a a Behaviour Age Internalising behaviours P Externalising behaviours P (months) β [95% CI], p β [95% CI], p
Supportive 24 -0.20 [-0.53, 0.12], .222 -0.29 [-0.62, 0.04], .083 Directive 36 -0.37 [-0.69, -0.05], .023 -0.14 [-0.45, 0.17], .381 48 -0.15 [-0.47, 0.17], .362 -0.13 [-0.46, 0.20], .437
Interpretation 24 -0.01 [-0.35, 0.34], .966 0.19 [-0.15, 0.54], .270 36 0.16 [-0.18, 0.50], .366 0.05 [-0.28, 0.39], .753 48 -0.12 [-0.46, 0.22], .482 -0.02 [-0.33, 0.36], .947
Imitation 24 0.06 [-0.24, 0.37], .677 0.05 [-0.26, 0.35], .768 36 -0.06 [-0.36, 0.24], .713 0.02 [-0.27, 0.32], .867 48 -0.15 [-0.45, 0.15], .317 -0.02 [-0.32, 0.27], .869
c Interactive behaviours coded in MIBS (this thesis) P
Supported Joint 24 -0.03 [-0.13, 0.07], .528 0.01 [-0.09, 0.11], .836 Engagement 36 -0.04 [-0.14, 0.06], .392 -0.01 [-0.11, 0.09], .861 48 -0.03 [-0.13, 0.07], .585 0.00 [-0.10, 0.11], .925
Coordinated Joint 24 0.02 [-0.09, 0.14], .707 0.03 [-0.08, 0.15], .562 Engagement 36 -0.03 [-0.15, 0.08], .605 -0.05 [-0.16, 0.06], .341 48 -0.03 [-0.14, 0.08], .590 0.03 [-0.09, 0.14], .633
Fluency and 24 -0.02 [-0.12, 0.08], .724 0.04 [-0.07, 0.14], .480 Connectedness 36 -0.04 [-0.15, 0.06], .391 0.01 [-0.09, 0.11], .798 48 -0.06 [-0.16, 0.04], .207 -0.01 [-0.11, 0.09], .897
Note: MRBS Maternal Responsive Behaviour Study; MIBS Maternal and Interactive Behaviours a Study; CI = Confidence Interval; Statistically significant findings are in bold. P P Child Behaviour b Checklist at 24 and 36 months, and Strengths and Difficulties Questionnaire at 48 months; P P Interpret c β coefficients as the SD difference in SEB score for one additional maternal behaviour per minute; P P Interpret β coefficients as the SD difference in SEB score associated with one higher interactive behaviour rating
.
239
8.4 Discussion
The findings show that whilst unsuccessful directives were associated with higher internalising behaviours at 24 and 36 months, there was no other convincing evidence for associations between the other maternal and interactive behaviours and
SEB outcomes.
The finding that unsuccessful directives were associated with internalising behaviours resonates with other research that shows that controlling parenting is associated with internalising problems at age 24 months (Bayer et al., 2006) (see section
2.7.2). This association is likely to be bidirectional, with controlling practices, such as using intrusive directives, reinforcing children’s existing internalising symptoms. The finding that the association was no longer evident at 48 months also resonates with research that shows the strongest predictors of internalising behaviours by age five include harsh discipline, maternal stress, and having no older siblings (Bayer et al.,
2012). These factors may begin to exert a greater influence on children’s SEB development, mitigating the effect of controlling behaviours used at 24 months.
However, since discipline and maternal stress were not measured in this study this would need to be examined in future studies. Furthermore, it would be useful to also determine whether the early association between unsuccessful directives and internalising symptoms holds using SEB measures beyond parent-report, although this might not be feasible at these young ages.
There was no evidence that the other behaviours were associated with SEB scores, despite responsiveness being positively associated with other aspects of child development (e.g. Landry et al., 2006). The communicative responsive behaviours
240
appear to be very specific to language development. The null findings might also be due to how interventions for SEB difficulties may work. Brassart et al. (2017) found that participating in an intervention for SEB difficulties was associated with increases in parental self-efficacy, that is, parents’ confidence in themselves as parents (Bandura,
1977; Jones & Prinz, 2005). This was also a theme that emerged from qualitative interviews of parents who participated in the Solihull Approach Parenting Group for children with moderate behaviour difficulties, with the majority reporting improved confidence in being effective parents (Vella, Butterworth, Johnson, & Law, 2015).
Greater self-efficacy has been linked to greater parental responsiveness (Dekovic et al.,
2010). The information parents receive during interventions could help them recognise they are more competent than they realised, leading to improved self-efficacy scores.
This could lead to greater responsiveness, resulting in an actual or perceived decrease in child behaviour problems (Brassart et al., 2017; Dekovic et al., 2010; Roskam et al.,
2015). Therefore, the null findings could be because it is participation in an intervention promoting parental responsiveness that leads to changes in SEB outcomes through increasing parental self-efficacy. This requires further examination in a carefully designed intervention study for children with SEB difficulties.
There are several study limitations that effect the interpretation of these findings.
Firstly, it is possible that the definitions of the maternal and interactive behaviours did not capture components of responsiveness most important for SEB development, e.g. warmth or affection. Warmth is associated with increased social-emotional competence
(e.g. Eisenberg et al., 2003; Treyvaud et al., 2009) and is defined as “the physical or verbal expression of affection, positive emotions, positive evaluations, and positive regard” (Roggman, Cook, Innocenti, Norman, & Christiansen, 2013, p. 291). Whilst
241
Fluency and Connectedness captured matched affect, it did not distinguish between positive versus negative affect. The praise definition captured some elements of warmth. However, all types of praise were included in a single code, which could have concealed distinctions important to SEB development, such as process versus person praise (Dweck, 2006; Gunderson et al., 2013) (see section 3.5). Similarly, all types of
EMST were included in one code, possibly obscuring distinctions important to SEB development (e.g. Drummond et al., 2014).
Secondly, it is feasible that using the metric of frequencies (i.e. rate per minute) rather than proportion of maternal input might have resulted in the associations with
SEB difficulties being less evident in the adjusted models. This is because amount of input might be more closely associated with the maternal and environmental characteristics adjusted for in the analyses than proportion of input (e.g. Hoff, 2003), and this could be why the associations were no longer evident in the adjusted models.
Future research analysing these linguistic behaviours as a proportion of input in addition to frequency of input is warranted.
Thirdly, many analytical comparisons were conducted to address the research question of Aim Three, increasing the risk of finding spurious associations. Therefore, the results are considered exploratory and should be interpreted cautiously. It was not appropriate to reduce the comparisons by pooling the data as for Aim One (Chapter
Five) because maternal and interactive behaviours were only measured at 24 months, whilst SEB behaviours were measured at 24, 36 and 48 months. For Aim One, both language and SEB difficulties were measured at all three time-points, meaning each comparison was cross-sectional, clustered within participants so there were three rows per participant (24m language x 24m SEB; 36m language x 36m SEB; 48m language x 242
48m SEB). This cautious approach means the individual statistically significant findings for missed opportunities, expansions, labels, and supportive directives were interpreted as spurious. The confidence intervals around the coefficients were wide, and the coefficients for the respective associations at the other ages were statistically insignificant. The finding that unsuccessful directives were associated with internalising behaviours at 24 and 36 months was more convincing because it was observed at two time-points, and the confidence intervals were tighter.
Finally, the analytical approach could not examine more complex associations between the behaviours and SEB outcomes. It is possible that EMST was associated with SEB development in a more complicated way than linear regression modelling could measure. For example, child language skills might mediate its association with
SEB scores. That research question was beyond the scope of this PhD project and will be addressed in future research.
The strengths of the study included the rigorous development of the coding schemes, the repeated measures of SEB outcomes using high quality measures, and the large sample size. These are discussed in more detail in the following Chapter.
8.5 Summary
This Chapter addressed Aim Three of this PhD, to determine whether maternal and interactive behaviours were associated with SEB difficulties. Only one maternal behaviour, unsuccessful directives, was found to be associated with SEB outcomes.
Reasons for the lack of findings regarding the other behaviours were discussed and
243
include how the behaviours were defined, and how SEB interventions might actually work. The next Chapter will discuss the overall findings of the thesis.
244
9 Discussion and Conclusions
9.1 Overview
This chapter will discuss how the findings from this PhD sit within the broader literature, and then present the study’s strengths and limitations, clinical implications, and recommendations for future research.
245
9.2 Summary and Interpretation of Key Findings
The overarching aim of this thesis was to examine whether language and SEB difficulties were associated between 24 and 48 months, and whether specific maternal and interactive behaviours might be associated with both difficulties. The key findings will be discussed below.
9.2.1 Association between language and SEB difficulties
Higher rates of SEB difficulties were found in children with poor receptive language abilities. Although it is not possible to determine causality from this study design, it is highly likely the association is bidirectional. Being unable to understand what people are saying could result in children developing internalising problems, such as clinginess, anxiety, and chronic refusal to join in with peers, which may further restrict their language learning opportunities. As stated in Chapter Five, the findings also suggest that there may be a ‘comprehension threshold’, below which social interactions are so difficult and uncomfortable that children externalise their behaviour, for example fidgeting, screaming, and destroying things. The children’s receptive language difficulties may prevent them from understanding their parents’ attempts to help them self-regulate (Vallotton & Ayoub, 2011), and their continuing behavioural difficulties may mask their comprehension difficulties. This hypothesis needs further investigation using prospective population-based data to compare parental concern about children’s language and behaviour between parents of children identified with and without co-occurring receptive language and behaviour difficulties.
246
The different patterns of association found between receptive language skills and internalising and externalising behaviours (i.e. across the spectrum of language abilities versus only at the extremes, respectively), and the lack of evidence of an association with expressive language skills, suggest that there may be multiple pathways to the associations between language and SEB difficulties. For example, receptive language difficulties may be integral to the emerging association in the pre-school years, whilst expressive language difficulties may only begin to play a role in SEB difficulties in the school years, mediated by academic difficulties (e.g. Beitchman, Wilson, et al., 1996).
The study design precludes establishing the underlying mechanisms of the association between language and SEB difficulties. However, finding that the co- occurrence is evident from 24 months, and that receptive language scores explained less than 5% of the variability in SEB outcomes may support models that propose that neurobiological and genetic mechanisms have a strong role in its onset (Zubrick et al.,
2007). Language difficulties may be one risk factor among many others for SEB difficulties, including family history of difficulties (e.g. Clegg et al., 2015), and similarly SEB difficulties may be one risk factor for language difficulties. It is therefore important for clinicians to ask about both behaviour and language difficulties in addition to other risk factors when taking case histories. This may help identify young children with language problems that are hidden by their behaviour difficulties, resulting in more timely and targeted intervention for both difficulties. Clinical practice will continue to be informed by research examining the nuances of the association between language and SEB difficulties, including the role of child resilience (Conti-
Ramsden & Durkin, 2015) and the complex interactions between language difficulties and risk and protective factors for SEB difficulties (Yew & O'Kearney, 2015a).
247
9.2.2 Parenting behaviours
Interventions for language and SEB difficulties are usually delivered separately, in small groups, over several months, and are parent-mediated (see section 2.8).
Whether children with co-occurring difficulties should be enrolled in language intervention and SEB intervention simultaneously or consecutively, or whether only one difficulty is addressed depends on several factors including the judgement of health professionals, service availability and importantly, parental capacity. For many parents, maintaining fidelity to two interventions would be challenging. This section discusses the findings regarding whether the same parenting behaviours might address both child difficulties.
Many of the parenting behaviours targeted in interventions for language and
SEB difficulties are similar (Tempel et al., 2009). As such, it seemed reasonable to hypothesise that the same parenting behaviours would be associated with both difficulties. If so, these parenting behaviours could be trialled in a single intervention for young children with these co-occurring problems. If successful, this might lead to an intervention pathway that is less resource-intensive for services, children, and parents.
There is also a possibility that the intervention could be more effective through an additive or synergistic effect of targeting two difficulties together.
However, only one parenting behaviour out of 13 (unsuccessful directives) was found to be associated with both language and SEB difficulties (see Table 21). Higher use of unsuccessful directives was associated with poorer language and poorer internalising behaviour scores at 24 and 36 months. This aligns with existing evidence that intrusive parenting can be detrimental to child development (e.g. Bayer et al., 2006;
248
Parker, 1983), and that intrusive directives disrupt language-learning (Akhtar et al.,
1991; Tomasello & Todd, 1983). Although seven other behaviours were associated with language outcomes (successful directives, Supported Joint Engagement, Fluency and
Connectedness, expansions, imitations, responsive questions and labels), the hypothesis that the same behaviours would be associated with both difficulties was not supported.
Table 21 Summary of Findings Maternal or Interactive Receptive and/or Internalising and/or Behaviour Expressive Language Externalising Behaviour 24m 36m 48m 24m 36m 48m Praise Emotion and Mental State Talk Unsuccessful directive Successful directive Supported Joint Engagement Coordinated Joint Engagement Fluency and Connectedness Responsive behaviours (Levickis et al., 2014) Expansion Imitation Responsive Question Label Supportive directive Interpretation
Note: Summary generated from findings adjusted for potential confounders
These findings might be explained by the underlying mechanisms by which maternal behaviours may bring about change in child language and SEB abilities. For child language development, these mechanisms are thought to be direct. For example, 249
by helping word-learning by labelling and recasting children’s utterances, modelling appropriate turn-taking by responding to child utterances, and scaffolding conversations by sharing joint attention. In this way, the parenting behaviours may afford children direct language-learning opportunities.
For child SEB skills, the mechanisms by which parenting behaviours may bring about change may be more complex than for language development. For example, as discussed in section 8.4, the effect may be indirect through increasing parental self- efficacy (Roskam et al., 2015). Studies that found that language interventions also address SEB difficulties, and those that have tested interventions that successfully address both difficulties (Brassart & Schelstraete, 2015; Delaney & Kaiser, 2001) could be tapping into this phenomenon. The reported improvements in language skills could be due to direct language-learning, whilst those observed in SEB skills could be indirect, due to increasing parental self-efficacy leading to improvements in responsiveness for example, rather than due to the specific learned parenting behaviours per se. Therefore, when parental behaviours are observed without the intervention context, as in this study, no association with SEB difficulties is evident.
Alternatively, the social learning theory and the family coercion model
(Bandura, 1977; G. R. Patterson, 1982; Sitnick et al., 2015) suggest that child behaviour problems decline during intervention due to a reduction in maladaptive coercive exchanges between parent and child. Interventions for SEB difficulties teach parents to increase their use of positive behaviours and effective limit-setting, thereby reducing their use of negative behaviours (e.g. harsh discipline) and increasing positive parent- child interactions (Forgatch & Patterson, 2010). Hence it might be the balance of positive and negative parenting behaviours that is important to SEB development. 250
Measuring precise behaviours in isolation, as occurred in the current study, does not capture this balance. Generating a composite score or ratio of the parenting behaviours
(e.g. positive to coercive) might be a more effective method to examine their associations with SEB outcomes, and requires further investigation.
Successful directives, Supported Joint Engagement, and Fluency and
Connectedness were the maternal and interactive behaviours found to be associated with language outcomes. These findings suggest there are specific linguistic features of parental input that are key to child language development. These include the quantity of talk (Hart & Risley, 1995), vocabulary and syntactic diversity (Rowe, 2012), imitations, expansions and responsive questions (Levickis et al., 2014), and now, successful and unsuccessful intrusive directives. The finding that directives were negatively associated with language scores supports and extends previous research (Akhtar et al., 1991) by showing that the negative association exists regardless of whether the child complies with the directive. These findings also demonstrate how the degree of connectedness and joint engagement is important for language development, thereby contributing to knowledge about how situational or stylistic factors of an interaction, such as its quality and responsiveness, are associated with language development (e.g. Adamson et al.,
2012; Hirsh-Pasek et al., 2015; Landry et al., 2006).
Finally, this study highlighted the dynamic nature of the contribution of maternal behaviours to child language development. For example, it revealed that although directives were associated with lower language scores, this association was weaker when mothers also used expansions. It is feasible that maternal use of responsive behaviours protects against the effects of intrusive behaviours. This resonates with the strength-based social learning approach that informs interventions for 251
SEB difficulties (Dishion et al., 2008; Sitnick et al., 2015). Such approaches may also be applicable to language interventions. That is, clinicians may be able to promote parenting behaviours associated with better language outcomes, rather than focusing on reducing behaviours associated with poorer language outcomes.
Further, this study demonstrated how maternal behaviours are associated with child language outcomes within the interactional context. The negative association between directives and language skills was weaker in highly connected dyads, and the positive association between expansions and imitations and language scores was stronger among dyads with lower levels of Supported Joint Engagement. These findings again reinforce the dynamic contribution made by maternal behaviours. They bear similarities to Tomasello and Farrar (1986) who found that only maternal input given inside versus outside joint attention was associated with language outcomes, and Yu and
Smith (2012) who showed labelling objects resulted in word learning only when the object was in infants’ direct line of sight. Together, this body of work suggests that when maternal linguistic behaviours and the interactional or situational context are aligned in a manner advantageous to language acquisition, children are more likely to learn language (Rowe, 2017).
The following sections will present the strengths and limitations of the study, the clinical implications and finally, recommendations for future research.
252
9.3 Strengths and Limitations
9.3.1 Strengths
Many strengths of this study lie in the characteristics of the sample and the study design. The sample was drawn from a community-based survey rather than a clinically- referred population. This means the findings describe associations across all language and SEB abilities, rather than only clinical levels of difficulties. The sample was also large relative to existing studies of parent-child interaction which often contain fewer than 100 participants (e.g. Masur et al., 2005). It was therefore adequate to detect significant effects and to result in more precise, meaningful confidence intervals for the estimated effects.
The young age of participants at recruitment, and the prospective design were also strengths of the study. This design means that data were collected at the emerging language phase of development and at the earliest age at which problem behaviours can become apparent, and at regular intervals over time. This minimises reporting bias, where recall of past child language and behaviour can be affected by current experiences, particularly the child’s current language skills and behaviours. Further, language and SEB difficulties could be measured concurrently at three time-points, using high-quality measures (see section 4.8), which enabled cross-sectional and longitudinal analysis. This was crucial for examining whether associations that were present at 24 months also existed at 36 and 48 months.
Language abilities were measured using gold-standard face-to-face assessments, rather than shorter parent-report measures. It was therefore possible to obtain objective
253
measures of expressive and receptive language skills. SEB difficulties were measured using well-recognised parent-report measures, which have strong psychometric properties (see section 4.8). By assessing both expressive and receptive language skills, and measuring internalising and externalising behaviours it was possible to examine the language and SEB association across different profiles for the first time at this age in a community-derived sample. It was also possible to examine whether the maternal and interactive behaviours were associated with receptive versus expressive language skills, and with internalising versus externalising behaviours.
The likelihood of observing naturalistic mother-child interactions was increased by conducting the video-recordings at the family home rather than in an unfamiliar location. This helped to obtain an objective measure of mother-child interaction that was as realistic as is possible in observational research. The development of the coding scheme for the maternal behaviours, and the refinement of the existing coding scheme for the interactive behaviours (Adamson et al., 2012) was rigorous (see sections 4.5 and
4.6). Behaviours that could not be coded reliably were excluded (e.g., child initiations and responses, direct and indirect directives, fillers). The process of piloting, refining, and re-piloting, and the reliability-checking that occurred at intervals throughout the coding process strengthened the confidence in the findings.
9.3.2 Limitations
There are several limitations which have important implications for the interpretation of the findings. Participants were recruited to the Let’s Learn Language trial based on their expressive vocabulary scores at 18 months (Wake et al., 2011).
Although by 36 months the sample’s mean language score was close to the population
254
norm (see section 4.3), caution is required when generalising the study findings to other populations. There is a possibility that the slower rate of vocabulary acquisition up to age 18 months may have resulted in different mother-child interaction patterns developing compared to dyads where the children’s expressive vocabulary was acquired without “delay”. Also, families from low SES backgrounds were under-represented so it is feasible that the findings are not generalisable to this group. Low SES is a risk factor for language and SEB difficulties (see section 2.6.3), and can present a unique set of challenges to parents, such as financial stress and insecurity that may impact on parent- child interactions. Replication of the study using general population samples, and samples of families experiencing high levels of adversity, is recommended.
There were some limitations regarding the data collected. At 48 months, the
SEB and language measures changed in order to align and harmonise the study measures with other similar child development studies to permit cross-study comparison. The Strengths and Difficulties Questionnaire and Child Behaviour
Checklist have moderate concurrent validity (Stone et al., 2010), and the CELF-P2 and
PLS-4 are both commonly used measures. However, the tools may differ in some unknown way. This is important to keep in mind, especially when interpreting associations that were present at 24 and 36 months, but were no longer observed at 48 months. Further, although non-verbal ability was measured at 48 months, it was not measured at 24 and 36 months due to the lack of reliable measures at these ages.
Therefore, non-verbal IQ could not be included as a potential confounder in the analyses.
There are limitations relating to the protocol used for the observations of mother-child interactions. It is possible that examining mother-child interaction in only 255
one scenario, that of unstructured free-play, might have missed context-specific effects of the maternal and interactive behaviours. Parenting behaviours differ by environmental settings and across time such that more positive behaviours are used during unstructured play whilst more negative behaviours are used during structured tasks (Blacher, Baker, & Kaladjian, 2013; Gardner, 2000). A research protocol using a range of scenarios, such as free-play, a challenging co-operation task, and a clean-up task might have captured more variability in the effects of the behaviours, particularly in relation to children’s SEB development. For example, a challenging task might capture the use of praise in response to the child’s efforts, and a clean-up task that children find undesirable might capture different effects of intrusive directives. Whilst there are many benefits to positioning a study within an existing high-quality research project, there are compromises involved including being restricted by data collection protocols.
A further consideration about the observation protocol is whether dyads may have altered their behaviour due to being observed and recorded by the research assistant (Uziel, 2010). If so, the recordings might not have been representative of their normal interaction patterns. Bennetts, Mensah, Westrupp, et al. (2017) conducted qualitative interviews with ten parents who had recently taken part in a free-play research task with their toddlers. Their findings suggest that mother-child interaction patterns may have changed due to being observed. For example, some mothers reported modifying their behaviour, from over-acting to freezing, whilst others seemed to have less awareness about their behaviour and did not report altering it. This is a limitation of all observational research, and must be balanced alongside the limitation of alternative measurement options, such as parent-report. It is also important to note that due to the observational nature of this study, it is not possible to infer causality. Suggestions about
256
other designs and analytical techniques that could improve our understanding about direction of effects are presented in section 9.5 below.
The final set of limitations relates to the coding of the mother-child interaction.
Initial plans to code child initiations and responses were suspended during piloting when it became apparent that reliably attending to and coding the mother and child’s behaviours simultaneously was not feasible. To code both partners’ behaviours would have required an extra investment of time that was outside the scope of this PhD study.
Furthermore, the single camera angle restricted observations of children’s non-verbal communication when their backs were turned or they were off-screen. Subsequently, it was not possible to code child behaviours.
Maternal language ability was not measured so it was not possible to determine how this was associated with the maternal or interactive behaviours. Furthermore, maternal utterances during the interactions were not transcribed due to the large sample size and restricted time available in a single PhD study. Therefore, it was not possible to adjust for the quantity of child-directed speech heard by the children during the interaction. Such a measure would have provided an estimate of how much speech children hear in contexts they are likely to learn from (McGillion, Pine, Herbert &
Matthews, 2017). This would have enabled examination of the strength of the associations between the maternal behaviours and child outcomes over and above maternal input quantity.
257
9.4 Clinical Implications
Before discussing the clinical implications of the study findings, it is important to recognise that the findings are reported at the group rather than the individual level. It is therefore possible that for individual mother-child dyads some maternal or interactive behaviours will be associated with language and/or SEB development even though at a group level evidence of an association was not found. A corollary of this caveat is that the sample contained few participants with severe SEB or language difficulties. Mother- child interactions could differ between mothers and children with and without severe difficulties. These findings therefore need replication in samples of children with clinical levels of difficulties.
Receptive language difficulties are difficult to detect, so knowing that they co- occur with SEB difficulties between 24 and 48 months may be critical to detecting them, because behaviour difficulties may be easier to recognise. Clinicians who identify
SEB difficulties in young children should be alert to looking for co-occurring receptive language difficulties. Without checking for these difficulties, verbal approaches used by the clinicians to remediate the SEB difficulties may not be effective because the children may not understand them. This information warrants inclusion in pre-service education, and in professional development programmes for health and education professionals.
Greater awareness among clinicians may result in more children being identified earlier and referred for treatment. Clinicians treating children with co-occurring language and SEB difficulties may need to continue addressing the difficulties separately, at least until further research is carried out (see section 9.5). This is because
258
the direct associations found between the maternal and interactive behaviours and language outcomes did not generalise to the SEB outcomes, except for directives. It will be critical for clinicians to collaborate with each other and researchers to determine which difficulties to address first, to share information for adapting their interventions, and to re-assess the children’s needs regularly.
Currently it is challenging to distinguish between transient and persistent child language difficulties in the pre-school years. Clinicians could find that including a mother-child interaction activity in the assessment process helps them to identify or prioritise children in need of intervention, including children at higher risk because of poorer maternal communication style. For example, a profile of low language scores, low connectedness, and high directive use could indicate higher need for intervention than a similar language profile but high connectedness and high responsive behaviour use. A feasibility study would be necessary to determine a suitable protocol for this, including what mother-child interaction task to use, whether to record the interaction or observe it in real time, which rating scales to use (see section 9.5 below), and its acceptability to clinicians and parents. Following this study, a RCT with long term follow-up would be necessary to determine whether inclusion of the interaction activity improved the identification of children in need of intervention.
Clinicians may also find that observing and monitoring mother-child interactions during interventions helps them to tailor treatment by choosing parenting behaviours most likely to be effective for individual mother-child dyads. By using a strengths-based approach, clinicians may find parents respond well to being encouraged to use ‘positive’ behaviours. Parent may perceive being discouraged from using ‘negative’ behaviours as undermining, which might cause parental hostility or disengagement from the 259
intervention. A strengths-based approach may include clinicians weaving interaction skills training throughout treatment to regularly reinforce the use of high quality interactions. A programme of research is necessary to determine whether altering clinical practice in these ways results in changes to targeted parenting behaviours and greater improvements in child language skills compared to current intervention protocols.
9.5 Future Research
The association between language and SEB difficulties offers a rich area for on- going investigation in this sample, and in others that represent low SES families, the general population, and families with children who have severe language and/or SEB difficulties. Exploring the trajectories of the association into the school years may reveal when expressive language difficulties begin to be associated with SEB difficulties, the role played by non-verbal IQ, and whether academic skills begin to mediate the association (Beitchman, Brownlie, et al., 1996). Investigating how the association between language and SEB difficulties affects participation in school and extra- curricular activities will inform the new approach of diagnosing Developmental
Language Disorder according to its impact on social and educational functioning rather than on assessment cut-points (see section 2.2.2) (Bishop et al., 2017).
Finally, investigating the role that pragmatic language skills play in the association between language and behaviour is another promising avenue of investigation. Law et al. (2014) found that pragmatic language skills substantially mediated the association between language difficulties and hyperactivity and peer problems, but not conduct problems or emotional symptoms, in children living in a low
260
SES neighbourhood (n = 138). They suggested that, if replicated in a large-scale population-based study, this finding would have implications for the development and delivery of interventions for children with these co-occurring difficulties. For example, incorporating pragmatic skills training into the intervention. Advancing this work requires detailed examination of the overlap between definitions of pragmatic language skills and SEB skills.
Several recommendations relate to the protocols for data collection and analysis for future parent-child interaction research. Ideally samples representative of the general population, including low SES families, should be used, and child interactions with their other key caregivers be investigated e.g. fathers, grandparents. This will add to the growing body of work that examines how caregivers contribute to development in different ways (e.g. Magill-Evans & Harrison, 2001; Möller, Nikolić, Majdandžić, &
Bögels, 2016). It is recommended that systematic observation occurs at regular intervals over time to allow researchers to examine more complex patterns of associations than was possible in the present study. For example, pathway analysis could be adopted to examine direction of effects, and mediation analysis to explore the role of parenting behaviours in the association between language and SEB difficulties. Two cameras should be used so both child and parent behaviours can be coded reliably.
Consideration should be given to coding the sub-types of praise and emotion and mental state talk, and incorporating a measure of parenting warmth. If possible, a sub- sample of recordings could be transcribed in order to derive a quantitative measure of child directed speech. This would allow examination of the extent to which the maternal behaviour codes predict language outcomes over and above the amount of language children hear. It would also allow examination of whether the amount and diversity of 261
words heard during supported and coordinated joint engagement is important beyond time spent in each of those engagement states. This information would make an important contribution to parent-child interaction research by helping to determine whether the significant time and resources required to transcribe maternal input is justified by greater predictive power.
Ideally, the parent and child should be observed during several tasks to elicit a broad range of behaviours, for example a challenging task, clean-up task, and a co- operation task. This would also allow examination of context-specific effects of these behaviours (Blacher et al., 2013; Gardner, 2000). It would be helpful to ask parents to rate how typical they believe their child’s behaviour was during each task to indicate how representative the observation might be of the child’s usual behaviour (Bennetts,
Mensah, Green, et al., 2017). This question would require careful development and validation.
If further research into parenting behaviours identified behaviours associated with both language and SEB difficulties, the next step would be to target these behaviours in a trial of a parent-mediated intervention for young children with co- occurring difficulties. The aim of the trial would be to examine whether teaching parents to use these specific behaviours results in improvements in children’s language and SEB difficulties. It would be important to measure parental self-efficacy, parent and interactive behaviours, and child language and SEB abilities pre- and post-intervention.
In this way, it would be possible to examine whether any changes to parenting or interactive behaviours were associated with changes in child outcomes, and to examine the role of parental self-efficacy.
262
The final recommendation for future research relates to the finding that
Fluency and Connectedness was positively associated with language scores between
24 and 48 months, and modified the associations between some maternal behaviours and language outcomes. The mean rating of Fluency and Connectedness in our sample
(3.82, SD 1.55) was similar to that found by Hirsh-Pasek et al (2015) in their smaller, low-income sample of 24 month old children (3.85, SD 1.13). The similarity in ratings in two separate studies of children of the same age is promising. A programme of research is warranted to investigate the feasibility of using the Fluency and
Connectedness rating scale to identify young children at risk of DLD, and/or to use in
DLD interventions. This programme might include determining whether it is possible to deconstruct Fluency and Connectedness into its key components, for example, turn- taking, matched affect, and extending the topic of conversation. It would be necessary to examine the distribution of abilities for each component in a pilot sample as a first step. These data could be used to determine cut-offs below which intervention would be recommended. For example, mother-child dyads with ratings in the bottom 10% for the child’s age group, equating to limited turn-taking or no child initiations for example, may indicate a need for intervention. Skills taught in the intervention would be aligned to the Fluency and Connectedness components to help clinicians monitor participants’ progress, and allow the intervention to be evaluated. Hence, it would be necessary to operationalise the components of Fluency and Connectedness into teachable constructs for clinicians and parents.
9.6 Conclusions
This thesis contributes knowledge that helps to understand the relationship between language and SEB difficulties in early childhood, and whether these difficulties 263
are associated with the same maternal and interactive behaviours. Revealing that receptive language difficulties can co-occur with SEB problems from ages two to four represents a promising means of identifying these children, and for referring them to early intervention. Whilst it is likely that there is an underlying neurobiological or genetic vulnerability that predisposes children to these difficulties, it is possible to conceptualise how once either or both difficulties emerge, they may lead to or exacerbate the other difficulty.
This thesis also contributes to our understanding about how maternal linguistic behaviours and the interactional context together are associated with children’s language development. Although interventions for both language and SEB difficulties target parenting behaviours, only one maternal behaviour was found that was directly associated with language and SEB scores. Further research is recommended to continue investigating how to best support children with these co-occurring difficulties and provide advice to their families.
264
References
Achenbach, T. M., & Rescorla, L. A. (2000). Manual for the ASEBA Pre-school Forms
and Profiles. Burlington, VT: University of Vermont, Research Center for
Children, Youth, and Families.
Adamson, L. B., & Bakeman, R. (1991). The development of shared attention during
infancy. In R. Vasta (Ed.), Annals of Child Development (Vol. 8, pp. 1-41).
London, UK: Jessica Kingsley Publishers.
Adamson, L. B., Bakeman, R., & Deckner, D. F. (2004). The development of symbol-
infused joint engagement. Child Development, 75(4), 1171-1187.
doi:10.1111/j.1467-8624.2004.00732.x
Adamson, L. B., Bakeman, R., Deckner, D. F., & Nelson, P. B. (2012). Rating parent-
child interactions: Joint engagement, communication dynamics, and shared
topics in autism, Down's Syndrome, and typical development. Journal of Autism
and Developmental Disorders, 42(12), 2622-2635. doi:10.1007/s10803-012-
1520-1
Adamson, L. B., Bakeman, R., Deckner, D. F., & Romski, M. (2009). Joint engagement
and the emergence of language in children with autism and Down's Syndrome.
Journal of Autism and Developmental Disorders, 39(1), 84-96.
doi:10.1007/s10803-008-0601-7
Akhtar, N. (2005). Is joint attention necessary for early language learning? In The
Development of Social Cognition and Communication. (pp. 165-179): Lawrence
Erlbaum Associates Publishers, Mahwah, NJ.
265
Akhtar, N., Dunham, F., & Dunham, P. J. (1991). Directive interactions and early
vocabulary development: The role of joint attentional focus. Journal of Child
Language, 18(1), 41-49.
Aksan, N., Kochanska, G., & Ortmann, M. R. (2006). Mutually responsive orientation
between parents and their young children: Toward methodological advances in
the science of relationships. Developmental Psychology, 42(5), 833-848.
doi:10.1037/0012-1649.42.5.833
Andersen, C. E., & Marinac, J. V. (2007). Using an observational framework to
investigate adult language input to young children in a naturalistic environment.
Child Language Teaching and Therapy, 23(3), 307-324.
Anselmi, L., Piccinini, C. A., Barros, F. C., & Lopes, R. S. (2004). Psychosocial
determinants of behaviour problems in Brazilian pre-school children. Journal of
Child Psychology and Psychiatry, 45(4), 779-788. doi:10.1111/j.1469-
7610.2004.00271.x
Astington, J. W., & Baird, J. A. (2005). Why language matters for theory of mind:
Oxford University Press.
Australian Bureau of Statistics, (2001). Socio-economic Indexes for Areas Canberra,
Australia: Australian Bureau of Statistics.
Bagner, D. M., Garcia, D., & Hill, R. (2016). Direct and indirect effects of behavioral
parent training on infant language production. Behavior Therapy, 47(2), 184-
197.
Bakeman, R., & Adamson, L. B. (1984). Coordinating attention to people and objects in
mother-infant and peer-infant interaction. Child Development, 55(4), 1278-1289.
Bakeman, R., & Quera, V. (2011). Sequential analysis and observational methods for
the behavioral sciences: Cambridge University Press.
266
Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioral change.
Psychological Review, 84(2), 191.
Barlow, J., & Stewart-Brown, S. (2000). Behavior problems and group-based parent
education programs. Journal of Developmental and Behavioral Pediatrics,
21(5), 356-370. doi:10.1097/00004703-200010000-00006
Barnes, S., Gutfreund, M., Satterly, D., & Wells, G. (1983). Characteristics of adult
speech which predict children's language development. Journal of Child
Language, 10(1), 65-84.
Barrett, P. M., Rapee, R. M., Dadds, M. M., & Ryan, S. M. (1996). Family
enhancement of cognitive style in anxious and aggressive children. Journal of
Abnormal Child Psychology, 24(2), 187-203. doi:10.1007/bf01441484
Bavin, E. L., & Bretherton, L. (2013). The Early Language in Victoria Study: Late-
talkers, predictors, and outcomes In L. Rescorla & P. Dale (Eds.), Late talkers:
Language development, interventions, and outcomes. Baltimore, USA: Brooks
Publishing Co.
Bayer, J. K., Hiscock, H., Ukoumunne, O. C., Price, A., & Wake, M. (2008). Early
childhood aetiology of mental health problems: A longitudinal population-based
study. Journal of Child Psychology and Psychiatry, 49(11), 1166-1174.
doi:10.1111/j.1469-7610.2008.01943.x
Bayer, J. K., Sanson, A. V., & Hemphill, S. A. (2006). Parent influences on early
childhood internalizing difficulties. Journal of Applied Developmental
Psychology, 27(6), 542-559. doi:10.1016/j.appdev.2006.08.002
Bayer, J. K., Sanson, A. V., & Hemphill, S. A. (2009). Early childhood aetiology of
internalising difficulties: A longitudinal community study. International Journal
of Mental Health Promotion, 11(1), 4-14.
267
Bayer, J. K., Ukoumunne, O. C., Mathers, M., Wake, M., Abdi, N., & Hiscock, H.
(2012). Development of children's internalising and externalising problems from
infancy to five years of age. Australian and New Zealand Journal of Psychiatry,
46(7), 659-668. doi:10.1177/0004867412450076
Beitchman, J. H., Brownlie, E. B., & Bao, L. (2014). Age 31 Mental Health Outcomes
of Childhood Language and Speech Disorders. Journal of the American
Academy of Child and Adolescent Psychiatry, 53(10), 1102-1110.
doi:10.1016/j.jaac.2014.07.006
Beitchman, J. H., Brownlie, E. B., Inglis, A., Wild, J., Ferguson, B., Schachter, D., . . .
Mathews, R. (1996). Seven-year follow-up of speech/language impaired and
control children: Psychiatric outcome. Journal of Child Psychology and
Psychiatry, and Allied Disciplines, 37(8), 961-970.
Beitchman, J. H., Nair, R., Clegg, M., & Patel, P. G. (1986). Prevalence of speech and
language disorders in 5-year-old kindergarten children in the Ottawa-Carleton
Region. Journal of Speech and Hearing Disorders, 51(2), 98-110.
Beitchman, J. H., Wilson, B., Brownlie, E. B., Walters, H., Inglis, A., & Lancee, W.
(1996). Long-term consistency in speech/language profiles .2. Behavioral,
emotional, and social outcomes. Journal of the American Academy of Child and
Adolescent Psychiatry, 35(6), 815-825. doi:10.1097/00004583-199606000-
00022
Beitchman, J. H., Wilson, B., Douglas, L., Young, A., & Adlaf, E. (2001). Substance
use disorders in young adults with and without LD: Predictive and concurrent
relationships. Journal of Learning Disabilities, 34(4), 317-332.
doi:10.1177/002221940103400407
268
Bell, R. Q., & Harper, L. V. (1977). The Effect of Children on Parents. Hillsdale, NJ:
ErJbaum.
Benner, G. J., Nelson, J. R., & Epstein, M. H. (2002). Language skills of children with
EBD: A literature review. Journal of Emotional and Behavioral Disorders,
10(1), 43-59. doi:10.1177/106342660201000105
Bennetts, S. K., Mensah, F. K., Green, J., Hackworth, N. J., Westrupp, E. M., & Reilly,
S. (2017). Mothers' experiences of parent-reported and video-recorded
observational assessments. Journal of Child and Family Studies.
doi:10.1007/s10826-017-0826-1
Bennetts, S. K., Mensah, F. K., Westrupp, E. M., Hackworth, N. J., Nicholson, J. M., &
Reilly, S. (2017). Establishing agreement between parent-reported and directly-
measured behaviours. Australasian Journal of Early Childhood, 42(1), 105-115.
Bercow, J. (2008). The Bercow Report: A review of services for children and young
people (0-19) with speech, language and communication needs. Department for
Children, Schools and Families, UK. Downloaded from
http://dera.ioe.ac.uk/id/eprint/8405
Beuker, K. T., Rommelse, N. N. J., Donders, R., & Buitelaar, J. K. (2013).
Development of early communication skills in the first two years of life. Infant
Behavior and Development, 36(1), 71-83.
doi:http://dx.doi.org/10.1016/j.infbeh.2012.11.001
Bigelow, A. E., MacLean, K., Proctor, J., Myatt, T., Gillis, R., & Power, M. (2010).
Maternal sensitivity throughout infancy: Continuity and relation to attachment
security. Infant Behavior and Development, 33(1), 50-60.
doi:10.1016/j.infbeh.2009.10.009
269
Bishop, D. V. M., Snowling, M. J., Thompson, P. A., Greenhalgh, T., & CATALISE 2
Consortium. (2017). Phase 2 of CATALISE: A multinational and
multidisciplinary Delphi consensus study of problems with language
development: Terminology. Journal of Child Psychology and Psychiatry,
58(10), 1068-1080. doi:10.1111/jcpp.12721
Bishop, D. V. M., Snowling, M. J., Thompson, P. A., Greenhalgh, T., & CATALISE
Consortium. (2016). CATALISE: A multinational and multidisciplinary Delphi
consensus study: Identifying language impairments in children. Plos One, 11(7),
26. doi:10.1371/journal.pone.0158753
Blacher, J., Baker, B. L., & Kaladjian, A. (2013). Syndrome specificity and mother-
child interactions: Examining positive and negative parenting across contexts
and time. Journal of Autism and Developmental Disorders, 43(4), 761-774.
doi:10.1007/s10803-012-1605-x
Blackwell, L. S., Trzesniewski, K. H., & Dweck, C. S. (2007). Implicit theories of
intelligence predict achievement across an adolescent transition: A longitudinal
study and an intervention. Child Development, 78(1), 246-263.
doi:10.1111/j.1467-8624.2007.00995.x
Bohannon, J., & Bonvillian, J. (1997). Theoretical approaches to language acquisition.
In J. Berko-Gleason (Ed.), The Development of Language (4th edition ed., pp.
259-316). Boston: Allyn & Bacon.
Bornstein, M. H. (2015). Children’s parents. In R. M. Lerner (Ed.), Handbook of Child
Psychology and Developmental Science (7 ed., pp. 55-132). Hoboken, NJ:
Wiley.
Bornstein, M. H., Hendricks, C., Haynes, O. M., & Painter, K. M. (2007). Maternal
sensitivity and child responsiveness: Associations with social context, maternal
270
characteristics, and child characteristics in a multivariate analysis. Infancy,
12(2), 189-223.
Bornstein, M. H., & Tamis-LeMonda, C. S. (1989). Maternal responsiveness and
cognitive development in children. New Directions for Child Development, 43,
49-61.
Bornstein, M. H., Tamis-LeMonda, C. S., Hahn, C. S., & Haynes, O. M. (2008).
Maternal responsiveness to young children at three ages: Longitudinal analysis
of a multidimensional, modular, and specific parenting construct. Developmental
Psychology, 44(3), 867-874. doi:10.1037/0012-1649.44.3.867
Boyle, M. H., Jenkins, J. M., Georgiades, K., Cairney, J., Duku, E., & Racine, Y.
(2004). Differential-maternal parenting behavior: Estimating within- and
between-family effects on children. Child Development, 75(5), 1457-1476.
doi:10.1111/j.1467-8624.2004.00751.x
Brassart, E., & Schelstraete, M. A. (2015). Enhancing the communication abilities of
pre-schoolers at risk for behavior problems: Effectiveness of a parent-
implemented language intervention. Infants and Young Children, 28(4), 337-
354. doi:10.1097/iyc.0000000000000049
Brassart, E., Schelstraete, M. A., & Roskam, I. (2017). What are the effects of a parent-
implemented verbal responsive intervention on pre-schoolers with externalizing
behavior problems? Child and Family Behavior Therapy, 39(2), 108-130.
doi:10.1080/07317107.2017.1307679
Breitenstein, S. M., Gross, D., Fogg, L., Ridge, A., Garvey, C., Julion, W., & Tucker, S.
(2012). The Chicago Parent Program: Comparing one-year outcomes for African
American and Latino parents of young children. Research in Nursing & Health,
35(5), 475-489. doi:10.1002/nur.21489
271
Brennan, L. M., Shaw, D. S., Dishion, T. J., & Wilson, M. (2012). Longitudinal
predictors of school-age academic achievement: Unique contributions of
toddler-age aggression, oppositionality, inattention, and hyperactivity. Journal
of Abnormal Child Psychology, 40(8), 1289-1300. doi:10.1007/s10802-012-
9639-2
Briggs-Gowan, M. J., Carter, A. S., Skuban, E. M., & Horwitz, S. M. (2001).
Prevalence of social-emotional and behavioral problems in a community sample
of 1 and 2-year-old children. Journal of the American Academy of Child and
Adolescent Psychiatry, 40(7), 811-819. doi:10.1097/00004583-200107000-
00016
Brookman-Frazee, L., Stahmer, A., Baker-Ericzen, M. J., & Tsai, K. (2006). Parenting
interventions for children with autism spectrum and disruptive behavior
disorders: Opportunities for cross-fertilization. Clinical Child and Family
Psychology Review, 9(3-4), 181-200. doi:10.1007/s10567-006-0010-4
Brophy, M., & Dunn, J. (2002). What did mummy say? Dyadic interactions between
young "hard-to-manage" children and their mothers. Journal of Abnormal Child
Psychology, 30(2), 103-112. doi:10.1023/a:1014705314406
Brownell, C. A., Svetlova, M., Anderson, R., Nichols, S. R., & Drummond, J. (2013).
Socialization of early prosocial behavior: Parents' talk about emotions is
associated with sharing and helping in toddlers. Infancy, 18(1), 91-119.
doi:10.1111/j.1532-7078.2012.00125.x
Bruner, J. S. (1975). From communication to language: Psychological perspective.
Cognition, 3(3), 255-287. doi:10.1016/0010-0277(74)90012-2
Bruner, J. S. (1978). Acquiring the uses of language. Canadian Journal of Psychology-
Revue Canadienne De Psychologie, 32(4), 204-218. doi:10.1037/h0081691
272
Bruner, J. S. (1986). Play, thought and language. Prospects, 16(1), 77-83.
doi:10.1007/bf02197974
Buschmann, A., Jooss, B., Rupp, A., Feldhusen, F., Pietz, J., & Philippi, H. (2009).
Parent-based language intervention for two-year-old children with specific
expressive language delay: A randomised controlled trial. Archives of Disease in
Childhood, 94(2), 110-116. doi:10.1136/adc.2008.141572
Caldwell, B. M., & Bradley, R. H. (2003). HOME Inventory Administration Manual.
Arkansas, US: University of Arkansas for Medical Sciences.
Can, D. D., Ginsburg-Block, M., Golinkoff, R. M., & Hirsh-Pasek, K. (2013). A long-
term predictive validity study: Can the CDI Short Form be used to predict
language and early literacy skills four years later? Journal of Child Language,
40(4), 821-835. doi:10.1017/S030500091200030X
Carpendale, J., & Lewis, C. (2004). Constructing an understanding of mind: The
development of children's social understanding within social interaction.
Behavioral and Brain Sciences, 27(1), 79-+.
Carpendale, J., & Lewis, C. (2006). How Children Develop Social Understanding.
Malden: Blackwell Publishing.
Carpenter, J., & Drabick, D. A. (2011). Co-occurrence of linguistic and behavioural
difficulties in early childhood: A developmental psychopathology perspective.
Early Child Development and Care, 181(8), 1021-1045.
doi:http://dx.doi.org/10.1080/03004430.2010.509795
Carpenter, M., Nagell, K., & Tomasello, M. (1998). Social cognition, joint attention,
and communicative competence from 9 to 15 months of age. Monographs of the
Society for Research in Child Development, 63(4), 176.
doi:http://dx.doi.org/10.2307/1166214
273
Cartmill, E. A., Armstrong, B. F., 3rd, Gleitman, L. R., Goldin-Meadow, S., Medina, T.
N., & Trueswell, J. C. (2013). Quality of early parent input predicts child
vocabulary three years later. Proc Natl Acad Sci U S A, 110(28), 11278-11283.
doi:10.1073/pnas.1309518110
Caulfield, M. B., Fischel, J. E., DeBaryshe, B. D., & Whitehurst, G. J. (1989).
Behavioral correlates of developmental expressive language disorder. Journal of
Abnormal Child Psychology, 17(2), 187-201.
Christensen, D., Zubrick, S. R., Lawrence, D., Mitrou, F., & Taylor, C. L. (2014). Risk
factors for low receptive vocabulary abilities in the pre-school and early school
years in the Longitudinal Study of Australian Children. Plos One, 9(7), 19.
doi:10.1371/journal.pone.0101476
Cicchetti, D., & Toth, S. L. (1998). The development of depression in children and
adolescents. American Psychologist, 53(2), 221-241. doi:10.1037/0003-
066x.53.2.221
Cimpian, A., Arce, H. M. C., Markman, E. M., & Dweck, C. S. (2007). Subtle linguistic
cues affect children's motivation. Psychological Science, 18(4), 314-316.
doi:10.1111/j.1467-9280.2007.01896.x
Clegg, J., Hollis, C., Mawhood, L., & Rutter, M. (2005). Developmental language
disorders: A follow-up in later adult life. Cognitive, language and psychosocial
outcomes. Journal of Child Psychology and Psychiatry, 46(2), 128-149.
doi:10.1111/j.1469-7610.2004.00342.x
Clegg, J., Law, J., Rush, R., Peters, T. J., & Roulstone, S. (2015). The contribution of
early language development to children's emotional and behavioural functioning
at six years: An analysis of data from the Children in Focus sample from the
274
ALSPAC birth cohort. Journal of Child Psychology and Psychiatry, 56(1), 67-
75. doi:10.1111/jcpp.12281
Clegg, J., Stackhouse, J., Finch, K., Murphy, C., & Nicholls, S. (2009). Language
abilities of secondary age pupils at risk of school exclusion: A preliminary
report. Child Language Teaching and Therapy, 25(1), 123-139.
doi:10.1177/0265659008098664
Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and
Psychosocial Measurement, 20, 37-46.
Cohen, J. (1968). Weighted kappa: Nominal scale agreement provision for scaled
disagreement or partial credit. Psychological Bulletin, 70(4), 213.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155-159.
doi:10.1037/0033-2909.112.1.155
Cohen, N. J., Menna, R., Vallance, D. D., Barwick, M. A., Im, N., & Horodezky, N. B.
(1998). Language, social cognitive processing, and behavioral characteristics of
psychiatrically disturbed children with previously identified and unsuspected
language impairments. Journal of Child Psychology and Psychiatry and Allied
Disciplines, 39(6), 853-864. doi:10.1017/s0021963098002789
Cole, P. M., Armstrong, L. M., & Pemberton, C. K. (2010). The role of language in the
development of emotion regulation. In Child Development at the Intersection of
Emotion and Cognition (pp. 59-77). Washington, DC: American Psychological
Association; US.
Conti-Ramsden, G., & Botting, N. (2008). Emotional health in adolescents with and
without a history of specific language impairment. Journal of Child Psychology
and Psychiatry, 49(5), 516-525. doi:10.1111/j.1469-7610.2007.01858.x
275
Conti-Ramsden, G., & Durkin, K. (2015). What factors influence language impairment?
Considering resilience as well as risk. Folia Phoniatrica Et Logopaedica, 67(6),
293-299.
Conti-Ramsden, G., & Durkin, K. (2016). Language impairment and adolescent
outcomes. Chichester: John Wiley & Sons Ltd.
Conti-Ramsden, G., & Friel-Patti, S. (1986). Mother–child dialogues: Considerations of
cognitive complexity for young language learning children. British Journal of
Disorders of Communication, 21(2), 245-255.
doi:http://dx.doi.org/10.3109/13682828609012281
Corkum, V., & Dunham, P. (1996). The Communicative Development Inventory
WORDS Short Form as an index of language production. Journal of Child
Language, 23(3), 515-528.
Corpus, J. H., & Lepper, M. R. (2007). The effects of person versus performance praise
on children's motivation: Gender and age as moderating factors. Educational
Psychology, 27(4), 487-508. doi:10.1080/01443410601159852
Crawley, S. B., & Spiker, D. (1983). Mother-child interactions involving two-year-olds
with Down syndrome: A look at individual differences. Child Development,
1312-1323.
Cross, M. (2011). Children with social, emotional and behavioural difficulties and
communication problems: There is always a reason: Jessica Kingsley
Publishers.
Culp, A. M., Hubbs-Tait, L., Culp, R. E., & Starost, H. J. (2000). Maternal parenting
characteristics and school involvement: Predictors of kindergarten cognitive
competence among Head Start children. Journal of Research in Childhood
Education, 15(1), 5-17.
276
Cytryn, L., & McKnew, D. (1996). Growing up sad. In: New York: Norton.
Dale, P. S., Dionne, G., Eley, T. C., & Plomin, R. (2000). Lexical and grammatical
development: A behavioural genetic perspective. Journal of Child Language,
27(3), 619-642. doi:10.1017/s0305000900004281
Dale, P. S., & Hayiou-Thomas, M. E. (2013). Outcomes for Late Talkers: A Twin Study
In L. Rescorla & P. Dale (Eds.), Late talkers: Language development,
interventions, and outcomes. Baltimore, USA: Brooks Publishing Co. de Rosnay, M., & Hughes, C. (2006). Conversation and theory of mind: Do children
talk their way to socio-cognitive understanding? British Journal of
Developmental Psychology, 24, 7-37. doi:10.1348/02615005x82901
Deater-Deckard, K., & Petrill, S. A. (2004). Parent-child dyadic mutuality and child
behavior problems: An investigation of gene-environment processes. Journal of
Child Psychology and Psychiatry, 45(6), 1171-1179. doi:10.1111/j.1469-
7610.2004.00309.x
Dekovic, M., Asscher, J. J., Hermanns, J., Reitz, E., Prinzie, P., & van den Akker, A. L.
(2010). Tracing changes in families who participated in the Home-Start
Parenting Program: Parental sense of competence as mechanism of change.
Prevention Science, 11(3), 263-274. doi:10.1007/s11121-009-0166-5
Delaney, E. M., & Kaiser, A. P. (2001). The effects of teaching parents blended
communication and behavior support strategies. Behavioral Disorders, 26(2),
93-116.
Denham, S. A., Zoller, D., & Couchoud, E. A. (1994). Socialization of preschoolers
emotion understanding. Developmental Psychology, 30(6), 928-936.
doi:10.1037/0012-1649.30.6.928
277
Desmarais, C., Sylvestre, A., Meyer, F., Bairati, I., & Rouleau, N. (2008). Systematic
review of the literature on characteristics of late-talking toddlers. International
Journal of Language and Communication Disorders, 43(4), 361-389.
doi:10.1080/13682820701546854
Deutscher, B., Fewell, R. R., & Gross, M. (2006). Enhancing the interactions of teenage
mothers and their at-risk children: Effectiveness of a maternal-focused
intervention. Topics in Early Childhood Special Education, 26(4), 194-205.
doi:10.1177/02711214060260040101
Diaz, R. M., Neal, C. J., & Vachio, A. (1991). Maternal teaching in the zone of
proximal development: A comparison of low and high-risk dyads. Merrill-
Palmer Quarterly (1982-), 83-107.
Dickinson, D. K., & Porche, M. V. (2011). Relation between language experiences in
pre-school classrooms and children's kindergarten and fourth-grade language
and reading abilities. Child Development, 82(3), 870-886. doi:10.1111/j.1467-
8624.2011.01576.x
Dieterich, S. E., Landry, S. H., Smith, K. E., Swank, P. R., & Hebert, H. M. (2006).
Impact of community mentors on maternal behaviors and child outcomes.
Journal of Early Intervention, 28(2), 111-124.
doi:10.1177/105381510602800203
Dionne, G., Tremblay, R., Boivin, M., Laplante, D., & Perusse, D. (2003). Physical
aggression and expressive vocabulary in 19-month-old twins. Developmental
Psychology, 39(2), 261-273. doi:10.1037/0012-1649.39.2.261
Dishion, T. J., Patterson, G. R., & Kavanagh, K. A. (1992). An experimental test of the
coercion model: linking theory, measurement, and intervention. New York, US:
Guilford Press.
278
Dishion, T. J., Shaw, D., Connell, A., Gardner, F., Weaver, C., & Wilson, M. (2008).
The Family Check-Up with high-risk indigent families: Preventing problem
behavior by increasing parents' positive behavior support in early childhood.
Child Development, 79(5), 1395-1414. doi:10.1111/j.1467-8624.2008.01195.x
Dixon, W. E., & Salley, B. J. (2006). "Shhh! We're tryin' to concentrate": Attention and
environmental distracters in novel word learning. Journal of Genetic
Psychology, 167(4), 393-414.
Dockrell, J., Lindsay, G., Roulstone, S., & Law, J. (2014). Supporting children with
speech, language and communication needs: An overview of the results of the
Better Communication Research Programme. International Journal of Language
and Communication Disorders, 49(5), 543-557. doi:10.1111/1460-6984.12089
Dodge, K. A. (1990). Nature versus nurture in childhood conduct disorder - It is time to
ask a different question. Developmental Psychology, 26(5), 698-701.
doi:10.1037//0012-1649.26.5.698
Dougherty, L. R., Leppert, K. A., Merwin, S. M., Smith, V. C., Bufferd, S. J., &
Kushner, M. R. (2015). Advances and directions in pre-school mental health
research. Child Development Perspectives, 9(1), 14-19. doi:10.1111/cdep.12099
Douglas, H. (2006). The Solihull Approach Under-fives Practitioner's Manual (4th ed.).
Solihull: CAMHS, Solihull Care Trust.
Drummond, J., Paul, E. F., Waugh, W. E., Hammond, S. I., & Brownell, C. A. (2014).
Here, there and everywhere: Emotion and mental state talk in different social
contexts predicts empathic helping in toddlers. Frontiers in Psychology, 5, 11.
doi:10.3389/fpsyg.2014.00361
Dubois-Comtois, K., Moss, E., Cyr, C., & Pascuzzo, K. (2013). Behavior problems in
middle childhood: The predictive role of maternal distress, child attachment, and
279
mother-child interactions. Journal of Abnormal Child Psychology, 41(8), 1311-
1324. doi:10.1007/s10802-013-9764-6
Dunham, P. J., & Dunham, F. (1995). Optimal social structures and adaptive infant
development. In C. Moore & P. J. Dunham (Eds.), Joint attention: Its origins
and role in development. (pp. 159-188). Hillsdale, New Jersey:: Lawrence
Erlbaum Associates, Publishers.
Dunn, J., Bretherton, I., & Munn, P. (1987). Conversations about feeling states between
mothers and their young children. Developmental Psychology, 23(1), 132-139.
doi:10.1037//0012-1649.23.1.132
Dunn, J., Brown, J., & Beardsall, L. (1991). Family talk about feeling states and
children's later understanding of others' emotions. Developmental Psychology,
27(3), 448-455. doi:10.1037/0012-1649.27.3.448
Dunn, J., Brown, J., Slomkowski, C., Tesla, C., & Youngblade, L. (1991). Young
children's understanding of other people's feeleings and beliefs - Individual
differences and their antecedents. Child Development, 62(6), 1352-1366.
doi:10.1111/j.1467-8624.1991.tb01610
Dunst, C. J., Trivette, C. M., & Cross, A. H. (1986). Mediating influences of social
support: Personal, family and child outcomes. American Journal of Mental
Deficiency, 90(4), 403-417.
Dweck, C. S. (2006). Mindset: The New Psychology of Success. New York, NY:
Ballantine Books.
Eadie, P., Nguyen, C., Carlin, J., Bavin, E. L., Bretherton, L., & Reilly, S. (2014).
Stability of language performance at four and five years: Measurement and
participant variability. International Journal of Language and Communication
Disorders, 49(2), 215-227.
280
Eisenberg, N., Cumberland, A., & Spinrad, T. L. (1998). Parental socialization of
emotion. Psychological Inquiry, 9(4), 241-273.
doi:10.1207/s15327965pli0904_1
Eisenberg, N., Guthrie, I. K., Fabes, R. A., Shepard, S., Losoya, S., Murphy, B. C., . . .
Reiser, M. (2000). Prediction of elementary school children's externalizing
problem behaviors from attentional and behavioral regulation and negative
emotionality. Child Development, 71(5), 1367-1382. doi:10.1111/1467-
8624.00233
Eisenberg, N., Valiente, C., Morris, A. S., Fabes, R. A., Cumberland, A., Reiser, M., . . .
Losoya, S. (2003). Longitudinal relations among parental emotional
expressivity, children's regulation, and quality of socio-emotional functioning.
Developmental Psychology, 39(1), 3-19. doi:10.1037//0012-1649.39.1.3
Eyberg, S. M. (2005). Tailoring and adapting parent-child interaction therapy to new
populations. Education and Treatment of Children, 197-201.
Falkus, G., Tilley, C., Thomas, C., Hockey, H., Kennedy, A., Arnold, T., . . . Pring, T.
(2016). Assessing the effectiveness of parent-child interaction therapy with
language delayed children: A clinical investigation. Child Language Teaching
and Therapy, 32(1), 7-17. doi:10.1177/0265659015574918
Fanning, J. L. (2007). Parent training for caregivers of typically developing,
economically disadvantaged pre-schoolers: An initial study in enhancing
language development, avoiding behaviour problems, and regulating family
stress. (Ph.D), University of Oregon, Ann Arbor. (3276047)
Farran, D. C., & Kasari, C. (1985). Rating maternal interactions with handicapped and
ecologically at-risk infants in 3 different samples. Cahiers De Psychologie
Cognitive-Current Psychology of Cognition, 5(3-4), 273-274.
281
Farrant, B. M., Devine, T. A. J., Maybery, M. T., & Fletcher, J. (2012). Empathy,
perspective-taking and prosocial behaviour: The importance of parenting
practices. Infant and Child Development, 21(2), 175-188. doi:10.1002/icd.740
Farrant, B. M., & Zubrick, S. R. (2012). Early vocabulary development: The importance
of joint attention and parent-child book reading. First Language, 32(3), 343-364.
doi:10.1177/0142723710393795
Feldman, H. M., Dale, P. S., Campbell, T. F., Colborn, D. K., Kurs-Lasky, M.,
Rockette, H. E., & Paradise, J. L. (2005). Concurrent and predictive validity of
parent reports of child language at ages two and three years. Child Development,
76(4), 856-868. doi:10.1111/j.1467-8624.2005.00882.x
Fenson, L., Bates, E., Dale, P. S., Marchman, V. A., Reznick, J. S., & Thal, D. J.
(2007). MacArthur-Bates Communicative Development Inventories: Paul H.
Brookes Publishing Company.
Fenson, L., Dale, P. S., Reznick, J. S., Bates, E., Thal, D. J., Pethick, S. J., . . . Stiles, J.
(1994). Variability in early communicative development. Monographs of the
Society for Research in Child Development, 59(5), v-173.
doi:http://dx.doi.org/10.2307/1166093
Fenson, L., Pethick, S., Renda, C., Cox, J. L., Dale, P. S., & Reznick, J. S. (2000).
Short-form versions of the MacArthur Communicative Development
Inventories. Applied Psycholinguistics, 21(1), 95-115.
Fernald, A., Marchman, V. A., & Weisleder, A. (2013). SES differences in language
processing skill and vocabulary are evident at 18 months. Developmental
Science, 16(2), 234-248.
282
Forgatch, M. S., & Patterson, G. R. (2010). Parent Management Training Oregon
Model: An intervention for antisocial behavior in children and adolescents.
Evidence-based psychotherapies for children and adolescents, 2, 159-178.
Funamoto, A., & Rinaldi, C. M. (2015). Measuring parent-child mutuality: A review of
current observational coding systems. Infant Mental Health Journal, 36(1), 3-11.
doi:10.1002/imhj.21481
Furlong, M., McGilloway, S., Bywater, T., Hutchings, J., Smith, S. M., & Donnelly, M.
(2012). Behavioural and cognitive-behavioural group-based parenting
programmes for early-onset conduct problems in children aged 3 to 12 years.
Cochrane Database of Systematic Reviews(2), 357.
doi:10.1002/14651858.CD008225.pub2
Gaertner, B. M., Spinrad, T. L., & Eisenberg, N. (2008). Focused attention in toddlers:
Measurement, stability, and relations to negative emotion and parenting. Infant
and Child Development, 17(4), 339-363. doi:10.1002/icd.580
Garcia, D., Bagner, D. M., Pruden, S. M., & Nichols-Lopez, K. (2015). Language
production in children with and at-risk for delay: Mediating role of parenting
skills. Journal of Clinical Child and Adolescent Psychology, 44(5), 814-825.
doi:10.1080/15374416.2014.900718
Gardner, F. (1994). The quality of joint activity between mothers and their children with
behaviour problems. Journal of Child Psychology and Psychiatry and Allied
Disciplines, 35(5), 935-948. doi:10.1111/j.1469-7610.1994.tb02303.x
Gardner, F. (2000). Parent-child interaction: Do observational findings reflect the
natural behavior of participants? Clinical Child and Family Psychology Review,
3(3), 185-199. doi:10.1023/A:1009503409699
283
Gardner, F., Dishion, T. J., Shaw, D. S., Burton, J., & Supplee, L. (2007). Randomized
prevention trial for early conduct problems: Effects on pro-active parenting and
links to toddler disruptive behavior. Journal of Family Psychology, 21(3), 398-
406. doi:10.1037/0893-3200.21.3.398
Gardner, F., & Leijten, P. (2017). Incredible Years parenting interventions: Current
effectiveness research and future directions. Current Opinion in Psychology, 15,
99-104. doi:10.1016/j.copsyc.2017.02.023
Gardner, F., Ward, S., Burton, J., & Wilson, C. (2003). The role of mother-child joint
play in the early development of children's conduct problems: A longitudinal
observational study. Social Development, 12(3), 361-378. doi:10.1111/1467-
9507.00238
Garner, P. W., Jones, D. C., Gaddy, G., & Rennie, K. M. (1997). Low-income mothers'
conversations about emotions and their children's emotional competence. Social
Development, 6(1), 37-52. doi:10.1111/j.1467-9507.1997.tb00093.x
Giles, J. W., & Heyman, G. D. (2003). Preschoolers' beliefs about the stability of
antisocial behavior: Implications for navigating social challenges. Social
Development, 12(2), 182-197. doi:10.1111/1467-9507.00228
Girard, L., Pingault, J. B., Doyle, O., Falissard, B., & Tremblay, R. E. (2016).
Developmental associations between conduct problems and expressive language
in early childhood: A population-based study. Journal of Abnormal Child
Psychology, 44(6), 1033-1043. doi:10.1007/s10802-015-0094-8
Girard, L., Pingault, J. B., Falissard, B., Boivin, M., Dionne, G., & Tremblay, R. E.
(2014). Physical aggression and language ability from 17 to 72 months: Cross-
lagged effects in a population sample. Plos One, 9(11), 8.
doi:10.1371/journal.pone.0112185
284
Girolametto, L. E., Bonafacio, S., Visini, C., Weitzman, E., Zocconi, E., & Pearce, P. S.
(2002). Mother-child interactions in Canada and Italy: Linguistic responsiveness
to late-talking toddlers. International Journal of Language and Communication
Disorders, 37(2), 151-171. doi:http://dx.doi.org/10.1080/13682820110116794
Girolametto, L. E., Bonifacio, S., Visini, C., Weitzman, E., Zocconi, E., & Pearce, P. S.
(2002). Mother-child interactions in Canada and Italy: linguistic responsiveness
to late-talking toddlers. International Journal of Language and Communication
Disorders, 37(2), 153-171. doi:10.1080/13682820110116794
Girolametto, L. E., Greenberg, J., & Manolson, H. A. (1986). Developing dialogue
skills: The Hanen Early Language Parent Program. Seminars in Speech and
Language, 7(4), 367-382. doi:10.1055/s-0028-1085235
Girolametto, L. E., Pearce, F. S., & Weitzman, E. (1997). Effects of lexical intervention
on the phonology of late talkers. Journal of Speech Language, and Hearing
Research, 40(2), 338-348.
Girolametto, L. E., Pearce, P. S., & Weitzman, E. (1995). The effects of focused
stimulation for promoting vocabulary in young children with delays: A pilot
study. Journal of Children's Communication Development, 17(2), 39-49.
Girolametto, L. E., Pearce, P. S., & Weitzman, E. (1996). Interactive focused
stimulation for toddlers with expressive vocabulary delays. Journal of Speech
and Hearing Research, 39(6), 1274-1283.
Girolametto, L. E., Weitzman, E., Wiigs, M., & Pearce, P. S. (1999). The relationship
between maternal language measures and language development in toddlers with
expressive vocabulary delays. American Journal of Speech-Language
Pathology, 8(4), 364-374.
285
Goodman, R. (2001). Psychometric properties of the Strengths and Difficulties
Questionnaire. Journal of the American Academy of Child and Adolescent
Psychiatry, 40(11), 1337-1345. doi:10.1097/00004583-200111000-00015
Grady, J. S., & Karraker, K. (2014). Do maternal warm and encouraging statements
reduce shy toddlers' social reticence? Infant and Child Development, 23(3), 295-
303. doi:10.1002/icd.1850
Gridley, N., Hutchings, J., & Baker-Henningham, H. (2015). The Incredible Years
Parent-Toddler Programme and parental language: A randomised controlled
trial. Child: Care, Health, and Development, 41(1), 103-111.
doi:10.1111/cch.12153
Gross, D., Garvey, C., Julion, W., Fogg, L., Tucker, S., & Mokros, H. (2009). Efficacy
of the Chicago Parent Program with low-income African American and Latino
parents of young children. Prevention Science, 10(1), 54-65.
doi:10.1007/s11121-008-0116-7
Gunderson, E. A., Gripshover, S. J., Romero, C., Dweck, C. S., Goldin-Meadow, S., &
Levine, S. C. (2013). Parent Praise to one to three year olds predicts children's
motivational frameworks five years later. Child Development, 84(5), 1526-1541.
doi:10.1111/cdev.12064
Guralnick, M. J. (2011). Why early intervention works: A systems perspective. Infants
and Young Children, 24(1), 6-28. doi:10.1097/IYC.0b013e3182002cfe
Guralnick, M. J. (2017). Early intervention for children with intellectual disabilities: An
update. Journal of Applied Research in Intellectual Disabilities, 30(2), 211-229.
doi:10.1111/jar.12233
286
Hack, M., Klein, N. K., & Taylor, H. G. (1995). Long term developmental outcomes of
low birth-weight infants. Future of Children, 5(1), 176-196.
doi:10.2307/1602514
Halfon, N., & Hochstein, M. (2002). Life course health development: An integrated
framework for developing health, policy, and research. Milbank Quarterly, 80,
433-479.
Halle, T., Forry, N., Hair, E., Perper, K., Wandner, L., Wessel, J., & Vick, J. (2009).
Disparities in early learning and development: Lessons from the Early
Childhood Longitudinal Study Birth Cohort. Retrieved from Washington, DC:
Hancock, T. B., Kaiser, A. P., & Delaney, E. M. (2002). Teaching parents of pre-
schoolers at high risk: Strategies to support language and positive behavior.
Topics in Early Childhood Special Education, 22(4), 191-212.
doi:10.1177/027112140202200402
Hart, B., & Risley, T. T. (1995). Meaningful differences in the everyday experiences of
young American children. Baltimore, MD: Paul H. Brooks.
Hartas, D. (2011). Families’ social backgrounds matter: Socio-economic factors, home
learning and young children’s language, literacy and social outcomes. British
Educational Research Journal, 37(6), 893-914.
doi:10.1080/01411926.2010.506945
Hayden, E. P., Durbin, C. E., Klein, D. N., & Olino, T. M. (2010). Maternal personality
influences the relationship between maternal reports and laboratory measures of
child temperament. Journal of Personality Assessment, 92(6), 586-593.
Hebert, H. M., Swank, P., Smith, K., & Landry, S. H. (2004). Maternal support for play
and language across early childhood. Early Education and Development, 15(1),
93-113. doi:http://dx.doi.org/10.1207/s15566935eed1501_6
287
Heineichs, N., Hahlweg, K., Bertram, H., Kuschel, A., Naumann, S., & Harstick, S.
(2006). The one year efficacy of a parent-training in the universal prevention of
child-behavior problems: Results from mothers and fathers. Zeitschrift Fur
Klinische Psychologie Und Psychotherapie, 35(2), 82-96. doi:10.1026/1616-
3443.35.2.82
Helland, W. A., Biringer, E., Helland, T., & Heimann, M. (2012). Exploring language
profiles for children with ADHD and children with Asperger Syndrome. Journal
of Attention Disorders, 16(1), 34-43. doi:10.1177/1087054710378233
Henderlong, J., & Lepper, M. R. (2002). The effects of praise on children's intrinsic
motivation: A review and synthesis. Psychological Bulletin, 128(5), 774-795.
doi:10.1037//0033-2909.128.5.774
Hirsh-Pasek, K., Adamson, L. B., Bakeman, R., Owen, M. T., Golinkoff, R. M., Pace,
A., . . . Suma, K. (2015). The contribution of early communication quality to
low-income children's language success. Psychological Science, 26(7), 1071-
1083. doi:10.1177/0956797615581493
Hoff-Ginsberg, E. (1998). The relation of birth order and socioeconomic status to
children's language experience and language development. Applied
Psycholinguistics, 19(4), 603-629. doi:10.1017/s0142716400010389
Hoff, E. (2003). The specificity of environmental influence: Socioeconomic status
affects early vocabulary development via maternal speech. Child Development,
74(5), 1368-1378. doi:http://dx.doi.org/10.1111/1467-8624.00612
Hoff, E., & Naigles, L. (2002). How children use input to acquire a lexicon. Child
Development, 73(2), 418-433. doi:10.1111/1467-8624.00415
Horwitz, S. M., Irwin, J. R., Briggs-Gowan, M. J., Heenan, J. M. B., Mendoza, J., &
Carter, A. S. (2003). Language delay in a community cohort of young children.
288
Journal of the American Academy of Child and Adolescent Psychiatry, 42(8),
932-940. doi:10.1097/01.ch1.0000046889.27264.5e
Huber, P. J. (1967). The behavior of maximum likelihood estimates under non-standard
conditions. Paper presented at the Fifth Berkeley Symposium on Mathematical
Statistics and Probability.
Hudry, K., Aldred, C., Wigham, S., Green, J., Leadbitter, K., Temple, K., . . .
Consortium, P. (2013). Predictors of parent-child interaction style in dyads with
autism. Research in Developmental Disabilities, 34(10), 3400-3410.
doi:10.1016/j.ridd.2013.07.015
Hughes, N., Sciberras, E., & Goldfeld, S. (2016). Family and community predictors of
comorbid language, socioemotional and behavior problems at school entry. Plos
One, 11(7), e0158802.
Hurwitz, S., & Watson, L. R. (2016). Joint attention revisited: Finding strengths among
children with autism. Autism, 20(5), 538-550. doi:10.1177/1362361315593536
Hutchings, J., Pearson-Blunt, R., & Pasteur, M. (2016). A pilot trial of the Incredible
Years Autism Spectrum and Language Delays programme. Good Autism
Practise, 17, 15-22.
Huttenlocher, J., Haight, W., Bryk, A., Seltzer, M., & Lyons, T. (1991). Early
vocabulary growth: Relation to language input and gender. Developmental
Psychology, 27(2), 236-248. doi:10.1037/0012-1649.27.2.236
Huttenlocher, J., Waterfall, H. R., Vasilyeva, M., Vevea, J. L., & Hedges, L. V. (2010).
Sources of variability in children's language growth. Cognitive Psychology,
61(4), 343-365. doi:10.1016/j.cogpsych.2010.08.002
Irwin, J. R., Carter, A. S., & Briggs-Gowan, M. J. (2002). The social-emotional
development of "late-talking" toddlers. Journal of the American Academy of
289
Child and Adolescent Psychiatry, 41(11), 1324-1332.
doi:10.1097/01.chi.0000024842.60748.41
Ivanova, M. Y., Achenbach, T. M., Rescorla, L. A., Harder, V. S., Ang, R. P.,
Bilenberg, N., . . . Verhulst, F. C. (2010). Pre-school psychopathology reported
by parents in 23 societies: Testing the Seven-Syndrome Model of the Child
Behavior Checklist for ages 1.5 to five. Journal of the American Academy of
Child and Adolescent Psychiatry, 49(12), 1215-1224.
doi:10.1016/j.jaac.2010.08.019
Jenkins, J. M., Turrell, S. L., Kogushi, Y., Lollis, S., & Ross, H. S. (2003). A
longitudinal investigation of the dynamics of mental state talk in families. Child
Development, 74(3), 905-920. doi:10.1111/1467-8624.00575
Jessee, A., McElwain, N. L., & Booth-LaForce, C. (2016). Maternal supportive
behavior, cognitive talk, and desire/emotion talk at 24 months: Distinct factors
and differential antecedents. Parenting: Science and Practice, 16(2), 63-86.
doi:10.1080/15295192.2016.1134988
Johnson, A. M., Hawes, D. J., Eisenberg, N., Kohlhoff, J., & Dudeney, J. (2017).
Emotion socialization and child conduct problems: A comprehensive review and
meta-analysis. Clinical Psychology Review, 54, 65-80.
doi:10.1016/j.cpr.2017.04.001
Johnson, C. J., Beitchman, J. H., & Brownlie, E. B. (2010). Twenty-year follow-up of
children with and without speech-language impairments: Family, educational,
occupational, and quality of life outcomes. American Journal of Speech-
Language Pathology, 19(1), 51-65. doi:10.1044/1058-0360(2009/08-0083)
290
Jones, T. L., & Prinz, R. J. (2005). Potential roles of parental self-efficacy in parent and
child adjustment: A review. Clinical Psychology Review, 25(3), 341-363.
doi:10.1016/j.cpr.2004.12.004
Kaiser, A. (1993). Parent-implemented language intervention: An environmental system
perspective. In A. Kaiser & D. Gray (Eds.), Enhancing Children's
Communication: Research Foundations for Intervention (pp. 63-84). Baltimore,
MD: P.H.Brookes.
Kaminski, J. W., Valle, L. A., Filene, J. H., & Boyle, C. L. (2008). A meta-analytic
review of components associated with parent training program effectiveness.
Journal of Abnormal Child Psychology, 36(4), 567-589. doi:10.1007/s10802-
007-9201-9
Kannass, K. N., & Oakes, L. M. (2008). The development of attention and its relations
to language in infancy and toddlerhood. Journal of Cognition and Development,
9(2), 222-246.
Kasari, C., Gulsrud, A., Freeman, S., Paparella, T., & Hellemann, G. (2012).
Longitudinal follow-up of children with autism receiving targeted interventions
on joint attention and play. Journal of the American Academy of Child and
Adolescent Psychiatry, 51(5), 487-495.
Keown, L. J., Woodward, L. J., & Field, J. (2001). Language development of pre-school
children born to teenage mothers. Infant and Child Development, 10(3), 129-
145. doi:10.1002/icd.282
Kiernan, K. E., & Huerta, M. C. (2008). Economic deprivation, maternal depression,
parenting and children's cognitive and emotional development in early
childhood. The British Journal of Sociology, 59(4), 783-806.
doi:10.1111/j.1468-4446.2008.00219.x
291
Kiernan, K. E., & Mensah, F. K. (2009). Poverty, maternal depression, family status and
children’s cognitive and behavioural development in early childhood: A
longitudinal study. Journal of Social Policy, 38(4), 569–588.
Kim, J. M., & Mahoney, G. (2005). The effects of relationship focused intervention on
Korean parents and their young children with disabilities. Research in
Developmental Disabilities, 26(2), 117-130. doi:10.1016/j.ridd.2004.08.001
Kirkwood, B. R., & Sterne, J. A. C. (2013). Essential Medical Statistics (2nd ed.).
Malden, Mass: Blackwell Science.
Klee, T., Carson, D. K., Gavin, W. J., Hall, L., Kent, A., & Reece, S. (1998).
Concurrent and predictive validity of an early language screening program.
Journal of Speech, Language, and Hearing Research, 41(3), 627-641.
Klimes-Dougan, B. (2007). Introduction to the special issue of social development:
Emotion socialization in childhood and adolescence. Social Development, 16(2),
203-209. doi:10.1111/j.1467-9507.2007.00380.x
Kong, N. Y., & Carta, J. J. (2013). Responsive interaction interventions for children
with or at-risk for developmental delays: A research synthesis. Topics in Early
Childhood Special Education, 33(1), 4-17. doi:10.1177/0271121411426486
Kostelnik, M., Whiren, A., Soderman, A. K., & Gregory, K. (2006). Guiding Children's
Social Development: Theory to Practice (5th ed.). NY, USA: Thomson Delmar
Learning.
La Paro, K. M., Justice, L., Skibbe, L. E., & Pianta, R. C. (2004). Relations among
maternal, child, and demographic factors and the persistence of pre-school
language impairment. American Journal of Speech-Language Pathology, 13(4),
291-303.
292
Laible, D. J. (2004). Mother-child discourse surrounding a child's past behavior at 30
months: Links to emotional understanding and early conscience development at
36 months. Merrill-Palmer Quarterly-Journal of Developmental Psychology,
50(2), 159-180. doi:10.1353/mpq.2004.0013
Landis, J. R., & Koch, G. G. (1977). Application of hierarchical kappa-type statistics in
assessment of majority agreement among multiple observers. Biometrics, 33(2),
363-374. doi:10.2307/2529786
Landry, S. H., Miller-Loncar, C. L., Smith, K. E., & Swank, P. R. (2002). The role of
early parenting in children's development of executive processes. Developmental
Neuropsychology, 21(1), 15-41.
Landry, S. H., Smith, K. E., & Swank, P. R. (2006). Responsive parenting: Establishing
early foundations for social, communication, and independent problem-solving
skills. Developmental Psychology, 42(4), 627-642.
Landry, S. H., Smith, K. E., Swank, P. R., & Denson, S. E. (2000). Is responsive
parenting during infancy enough to promote more optimal development for pre-
term children or is consistency across childhood necessary? Pediatric Research,
47(4), 316A-316A.
Landry, S. H., Smith, K. E., Swank, P. R., & Guttentag, C. (2008). A responsive
parenting intervention: The optimal timing across early childhood for impacting
maternal behaviors and child outcomes. Developmental Psychology, 44, 1335-
1353. doi:10.1037/a0013030. Document Type: journal.Citation. DOI:
10.1037/a0013030.
Law, J., Plunkett, C. C., & Stringer, H. (2012). Communication interventions and their
impact on behaviour in the young child: A systematic review. Child Language
Teaching & Therapy, 28(1), 7-23. doi:10.1177/0265659011414214
293
Law, J., Reilly, S., & Snow, P. C. (2013). Child speech, language and communication
need re-examined in a public health context: A new direction for the speech and
language therapy profession. International Journal of Language and
Communication Disorders, 48(5), 486-496. doi:10.1111/1460-6984.12027
Law, J., Rush, R., & McBean, K. (2014). The relative roles played by structural and
pragmatic language skills in relation to behaviour in a population of primary
school children from socially disadvantaged backgrounds. Emotional and
Behavioural Difficulties, 19(1), 28-40. doi:10.1080/13632752.2013.854960
Law, J., & Stringer, H. (2014). The overlap between behaviour and communication and
its implications for mental health in childhood: The elephant in the room.
Emotional and Behavioural Difficulties, 19(1), 2-6.
doi:10.1080/13632752.2013.854959
Leffel, K., & Suskind, D. (2013). Parent-directed approaches to enrich the early
language environments of children living in poverty. Seminars in Speech and
Language, 34(4), 267-277. doi:10.1055/s-0033-1353443
Levickis, P. (2013). Understanding parent behaviours that promote language in slow-
to-talk toddlers (PhD PhD), The University of Melbourne, Melbourne.
Levickis, P., Reilly, S., Girolametto, L. E., Ukoumunne, O. C., & Wake, M. (2014).
Maternal behaviors promoting language acquisition in slow-to-talk toddlers:
Prospective community-based study. Journal of Developmental and Behavioral
Pediatrics, 35(4), 274-281. doi:10.1097/dbp.0000000000000056
Lindsay, G., & Dockrell, J. E. (2012). Longitudinal patterns of behavioral, emotional,
and social difficulties and self-concepts in adolescents with a history of Specific
Language Impairment. Language, Speech, and Hearing Services in Schools,
43(4), 445-460. doi:10.1044/0161-1461(2012/11-0069)
294
Lougheed, J. P., Hollenstein, T., Lichtwarck-Aschoff, A., & Granic, I. (2015). Maternal
regulation of child affect in externalizing and typically-developing children.
Journal of Family Psychology, 29(1), 10-19. doi:10.1037/a0038429
Lunkenheimer, E. S., Dishion, T. J., Shaw, D. S., Connell, A. M., Gardner, F., Wilson,
M. N., & Skuban, E. M. (2008). Collateral benefits of the Family Check-Up on
early childhood school readiness: Indirect effects of parents' positive behavior
support. Developmental Psychology, 44(6), 1737-1752. doi:10.1037/a0013858
Magill-Evans, J., & Harrison, M. J. (2001). Parent–child interactions, parenting stress,
and developmental outcomes at 4 years. Children's Health Care, 30(2), 135-150.
doi:http://dx.doi.org/10.1207/S15326888CHC3002_4
Mahoney, G., & Perales, F. (2003). Using relationship-focused intervention to enhance
the social-emotional functioning of young children with autism spectrum
disorders. Topics in Early Childhood Special Education, 23(2), 77-89.
Mahoney, G., & Perales, F. (2005). Relationship-focused early intervention with
children with pervasive developmental disorders and other disabilities: A
comparative study. Journal of Developmental and Behavioral Pediatrics, 26(2),
77-85. doi:10.1097/00004703-200504000-00002
Mahoney, G., Powell, A., & Finger, I. (1986). The maternal behavior rating scale.
Topics in Early Childhood Special Education, 6(2), 44-56.
Manolson, H. A., Ward, B., & Doddington, N. (1995). You make the difference in
helping your child learn: The Hanen Centre.
Marfo, K. (1990). Maternal directiveness in interactions with mentally handicapped
children: An analytical commentary. Journal of Child Psychology and
Psychiatry, 31(4), 531-549.
295
Martin, R. M., & Green, J. A. (2005). The use of emotion explanations by mothers:
Relation to pre-schoolers' gender and understanding of emotions. Social
Development, 14(2), 229-249.
Martinez, C. R., & Forgatch, M. S. (2001). Preventing problems with boys'
noncompliance: Effects of a parent training intervention for divorcing mothers.
Journal of Consulting and Clinical Psychology, 69(3), 416-428.
doi:10.1037//0022-006x.69.3.416
Masten, A. S., & Cicchetti, D. (2010). Developmental cascades. Development and
Psychopathology, 22(3), 491-495. doi:10.1017/s0954579410000222
Masur, E. F., Flynn, V., & Eichorst, D. L. (2005). Maternal responsive and directive
behaviours and utterances as predictors of children's lexical development.
Journal of Child Language, 32(1), 63-91.
Masur, E. F., Flynn, V., & Lloyd, C. A. (2013). Infants' behaviors as antecedents and
consequents of mothers' responsive and directive utterances. First Language,
33(4), 354-371. doi:10.1177/0142723713490603
Matthews, D. (2014). Pragmatic Development in First Language Acquisition.
Amsterdam: John Benjamins Publishing Company.
McDonald, L., & Pien, D. (1982). Mother conversational behaviour as a function of
interactional intent. Journal of Child Language, 9(2), 337-358.
McKean, C., Reilly, S., Bavin, E. L., Bretherton, L., Cini, E., Conway, L., Cook F.,
Eadie P., Prior, M., Wake M., & Mensah, F. (2017). Language outcomes at
seven years: Early predictors and co-occurring difficulties. Pediatrics, 139(3),
10. doi:10.1542/peds.2016-1684
McGillion, M., Pine, J. M., Herbert, J. S., & Matthews, D. (2017). A randomised
controlled trial to test the effect of promoting caregiver contingent talk on
296
language development in infants from diverse socioeconomic status
backgrounds. Journal of Child Psychology and Psychiatry, 58(10), 1122-1131.
doi:10.1111/jcpp.12725
McLoyd, V. C. (1998). Socioeconomic disadvantage and child development. American
Psychologist, 53(2), 185-204. doi:10.1037/0003-066x.53.2.185
Meins, E., Fernyhough, C., Fradley, E., & Tuckey, M. (2001). Re-thinking maternal
sensitivity: Mothers' comments on infants' mental processes predict security of
attachment at 12 months. Journal of Child Psychology and Psychiatry and
Allied Disciplines, 42(5), 637-648. doi:10.1017/s0021963001007302
Michiels, D., Grietens, H., Onghena, P., & Kuppens, S. (2008). Parent-child interactions
and relational aggression in peer relationships. Developmental Review, 28(4),
522-540. doi:10.1016/j.dr.2008.08.002
Mills, R. S. L., & Rubin, K. H. (1998). Are behavioural and psychological control both
differentially associated with childhood aggression and social withdrawal?
Canadian Journal of Behavioural Science-Revue Canadienne Des Sciences Du
Comportement, 30(2), 132-136. doi:10.1037/h0085803
Möller, E. L., Nikolić, M., Majdandžić, M., & Bögels, S. M. (2016). Associations
between maternal and paternal parenting behaviors, anxiety and its precursors in
early childhood: A meta-analysis. Clinical Psychology Review, 45, 17-33.
Moore, C., & Dunham, P. J. (1995). Joint attention: Its origins and role in development.
Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
Moore, C., Furrow, D., Chiasson, L., & Patriquin, M. (1994). Developmental
relationships between production and comprehension of mental terms. First
Language, 14(40, Pt 1), 1-17.
297
Moran, G., Pederson, D. R., & Krupka, A. (2005). Maternal unresolved attachment
status impedes the effectiveness of interventions with adolescent mothers. Infant
Mental Health Journal, 26(3), 231-249. doi:10.1002/imhj.20045
Morgan, P. L., Farkas, G., Hillemeier, M. M., Hammer, C. S., & Maczuga, S. (2015). 24
month old children with larger oral vocabularies display greater academic and
behavioral functioning at kindergarten entry. Child Development, 86(5), 1351-
1370. doi:10.1111/cdev.12398
Mueller, C., Rowe, M. L., & Zuckerman, B. (2017). Mindset matters for parents and
adolescents. JAMA Pediatrics, 171(5), 415-416.
10.1001/jamapediatrics.2016.5160
Mundy, P., Block, J., Delgado, C., Pomares, Y., Van Hecke, A. V., & Parlade, M. V.
(2007). Individual differences and the development of joint attention in infancy.
Child Development, 78(3), 938-954. doi:10.1111/j.1467-8624.2007.01042.x
Mundy, P., Delgado, C., Block, J., Venezia, M., Hogan, A., & Seibert, J. (2003). Early
social communication scales (ESCS). Coral Gables, FL: University of Miami.
Mundy, P., Kasari, C., Sigman, M., & Ruskin, E. (1995). Non-verbal communication
and early language acquisition in children with Down's Syndrome and in
normally developing children. Journal of Speech, Language, and Hearing
Research, 38(1), 157-167.
Nelson, H. D., Nygren, P., Walker, M., & Panoscha, R. (2006). Screening for speech
and language delay in pre-school children: Systematic evidence review for the
US preventive services task force. Pediatrics, 117(2), E298-E319.
doi:10.1154/peds.2005-1467
Nelson, K. (2007). Young minds in social worlds: Experience, meaning, and memory.
Cambridge, Mass.: Harvard University Press.
298
Newport, E. L., Gleitman, H., & Gleitman, L. R. (1977). Mother, I'd rather do it myself:
Some effects and non-effects of maternal speech style. In Snow C. E. &
Ferguson C. (eds.), Talking to children. Cambridge: C.U.P.
Ninio, A. (1980). Picture book reading in mother-infant dyads belonging to two
subgroups in Israel. Child Development, 51(2), 587-590. doi:10.2307/1129299
Ninio, A. (1983). Joint book reading as a multiple vocabulary acquisition device.
Developmental Psychology, 19(3), 445-451. doi:10.1037//0012-1649.19.3.445
Ninio, A., Snow, C. E., Pan, B. A., & Rollins, P. R. (1994). Classifying communicative
acts in children's interactions. Journal of Communication Disorders, 27(2), 157-
187. doi:10.1016/0021-9924(94)90039-6
Nixon, R. D. V., Sweeney, L., Erickson, D. B., & Touyz, S. W. (2003). Parent-child
interaction therapy: A comparison of standard and abbreviated treatments for
oppositional defiant pre-schoolers. Journal of Consulting and Clinical
Psychology, 71(2), 251-260. doi:10.1037/0022-006x.71.2.251
Noel, M., Peterson, C., & Jesso, B. (2008). The relationship of parenting stress and
child temperament to language development among economically disadvantaged
preschoolers. Journal of Child Language, 35(4), 823-843.
doi:10.1017/s0305000908008805
Noldus, I. (2008). The Observer XT (Version 8.0).
Norbury, C. F., Gooch, D., Baird, G., Charman, T., Simonoff, E., & Pickles, A. (2016).
Younger children experience lower levels of language competence and academic
progress in the first year of school: Evidence from a population study. Journal of
Child Psychology and Psychiatry, 57(1), 65-73. doi:10.1111/jcpp.12431
Nowak, C., & Heinrichs, N. (2008). A comprehensive meta-analysis of Triple-P
Positive Parenting Program using hierarchical linear modeling: Effectiveness
299
and moderating variables. Clinical Child and Family Psychology Review, 11(3),
114-144. doi:10.1007/s10567-008-0033-0
Nozadi, S. S., Spinrad, T. L., Eisenberg, N., Bolnick, R., Eggum-Wilkens, N. D., Smith,
C. L., . . . Sallquist, J. (2013). Prediction of toddlers' expressive language from
maternal sensitivity and toddlers' anger expressions: A developmental
perspective. Infant Behavior and Development, 36(4), 650-661.
doi:10.1016/j.infbeh.2013.06.002
O'Connor, T. G. (2002). Annotation: The "effects" of parenting re-considered: Findings,
challenges, and applications. Journal of Child Psychology and Psychiatry and
Allied Disciplines, 43(5), 555-572. doi:10.1111/1469-7610.00046
Pachter, L. M., Auinger, P., Palmer, R., & Weitzman, M. (2006). Do parenting and the
home environment, maternal depression, neighborhood, and chronic poverty
affect child behavioral problems differently in different racial-ethnic groups?
Pediatrics, 117(4), 1329-1338. doi:10.1542/peds.2005-1784
Pan, B. A., Rowe, M. L., Singer, J. D., & Snow, C. E. (2005). Maternal correlates of
growth in toddler vocabulary production in low-income families. Child
Development, 76(4), 763-782.
Parker, G. (1983). Parental affectionless control as an antecedent to adult depression: A
risk factor delineated. Archives of General Psychiatry, 40(9), 956-960.
Patterson, G. R. (1982). Coercive Family Process (Vol. 3): Castalia Publishing
Company.
Patterson, S. Y., Elder, L., Gulsrud, A., & Kasari, C. (2014). The association between
parental interaction style and children’s joint engagement in families with
toddlers with autism. Autism, 18(5), 511-518.
300
Paul, R., & Shiffer, M. E. (1991). Communicative initiations in normal and late-talking
toddlers. Applied Psycholinguistics, 12(4), 419-431.
doi:http://dx.doi.org/10.1017/S0142716400005853
Perneger, T. V. (1998). What's wrong with Bonferroni adjustments? BMJ, 316, 1236.
Petersen, I. T., Bates, J. E., D'Onofrio, B. M., Coyne, C. A., Lansford, J. E., Dodge, K.
A., . . . Van Hulle, C. A. (2013). Language ability predicts the development of
behavior problems in children. Journal of Abnormal Psychology, 122(2), 542-
557. doi:10.1037/a0031963
Pettit, G. S., & Bates, J. E. (1989). Family interaction patterns and children's behaviour
problems from infancy to 4 years. Developmental Psychology, 25(3), 413-420.
doi:10.1037//0012-1649.25.3.413
Pine, J. M. (1992). Maternal style at the early one-word stage: Re-evaluating the
stereotype of the directive mother. First Language, 12(35, Pt 2), 169-186.
doi:http://dx.doi.org/10.1177/014272379201203504
Pinquart, M. (2017). Associations of parenting dimensions and styles with externalizing
problems of children and adolescents: An updated meta-analysis. Developmental
Psychology, 53(5), 873-932. doi:10.1037/dev0000295
Plomin, R., Price, T. S., Eley, T. C., Dale, P. S., & Stevenson, J. (2002). Associations
between behaviour problems and verbal and non-verbal cognitive abilities and
disabilities in early childhood. Journal of Child Psychology and Psychiatry and
Allied Disciplines, 43(5), 619-633. doi:10.1111/1469-7610.00051
Pontoppidan, M., Niss, N. K., Pejtersen, J. H., Julian, M. M., & Vaever, M. S. (2017).
Parent report measures of infant and toddler social-emotional development: A
systematic review. Family Practice, 34(2), 127-137.
doi:10.1093/fampra/cmx003
301
Pungello, E. P., Iruka, I. U., Dotterer, A. M., Mills-Koonce, R., & Reznick, J. S. (2009).
The effects of socioeconomic status, race, and parenting on language
development in early childhood. Developmental Psychology, 45(2), 544-557.
doi:http://dx.doi.org/10.1037/a0013917
Raver, C. C. (1996). Relations between social contingency in mother-child interaction
and 2 year olds' social competence. Developmental Psychology, 32(5), 850-859.
doi:10.1037/0012-1649.32.5.850
Raviv, T., Kessenich, M., & Morrison, F. J. (2004). A mediational model of the
association between socioeconomic status and three year old language abilities:
The role of parenting factors. Early Childhood Research Quarterly, 19(4), 528-
547. doi:10.1016/j.ecresq.2004.10.007
Reese, E., Leyva, D., Sparks, A., & Grolnick, W. (2010). Maternal elaborative
reminiscing increases low-income children's narrative skills relative to dialogic
reading. Early Education and Development, 21(3), 318-342.
doi:10.1080/10409289.2010.481552
Reilly, S., Wake, M., Ukoumunne, O. C., Bavin, E. L., Prior, M., Cini, E., . . .
Bretherton, L. (2010). Predicting language outcomes at four years of age:
Findings from Early Language in Victoria Study. Pediatrics, 126(6), e1530-
e1537. doi:10.1542/peds.2010-0254
Rescorla, L. (1989). The language development survey screening tool for delayed
language in toddlers. Journal of Speech and Hearing Disorders, 54(4), 587-599.
Rescorla, L., Bascome, A., Lampard, J., & Feeny, N. (2001). Conversational patterns in
late-talkers at age 3. Applied Psycholinguistics, 22(2), 235-251.
doi:http://dx.doi.org/10.1017/S0142716401002053
302
Rescorla, L., & Fechnay, T. (1996). Mother-child synchrony and communicative
reciprocity in late-talking toddlers. Journal of Speech and Hearing Research,
39(1), 200-208.
Rhoades, E. A. (2003). Lexical-semantic and morphosyntactic language assessment in
auditory-verbal intervention: A position paper. Volta Review, 103(3), 169-184.
Rice, M. L., Sell, M. A., & Hadley, P. A. (1990). The social interactive coding system
(SICS): An on-line clinicaly relevant descriptive tool. Language, Speech, and
Hearing Services in Schools, 21, 1-20.
Rieffe, C., & Wiefferink, C. H. (2017). Happy faces, sad faces: Emotion understanding
in toddlers and pre-schoolers with language impairments. Research in
Developmental Disabilities, 62, 40-49. doi:10.1016/j.ridd.2016.12.018
Ripley, K., & Yuill, N. (2005). Patterns of language impairment and behaviour in boys
excluded from school. British Journal of Educational Psychology, 75, 37-50.
doi:10.1348/000709905x27696
Roben, C. K. P., Cole, P. M., & Armstrong, L. M. (2013). Longitudinal relations among
language skills, anger expression, and regulatory strategies in early childhood.
Child Development, 84(3), 891-905. doi:10.1111/cdev.12027
Roberts, M. Y., & Kaiser, A. P. (2011). The effectiveness of parent-implemented
language interventions: A meta-analysis. American Journal of Speech-Language
Pathology, 20(3), 180-199.
Roberts, M. Y., & Kaiser, A. P. (2015). Early intervention for toddlers with language
delays: A randomized controlled trial. Pediatrics, 135(4), 686-693.
doi:10.1542/peds.2014-2134
303
Robinson, E. A., & Eyberg, S. M. (1981). The dyadic parent-child interaction coding
system: Standardisation and validation. Journal of Consulting and Clinical
Psychology, 49(2), 245-250. doi:10.1037//0022-006x.49.2.245
Roggman, L. A., Cook, G. A., Innocenti, M. S., Norman, V. J., & Christiansen, K.
(2013). Parenting interactions with children: Checklist of observations linked to
outcomes (PICCOLO) in diverse ethnic groups. Infant Mental Health Journal,
34(4), 290-306. doi:10.1002/imhj.21389
Roskam, I., Brassart, E., Loop, L., Mouton, B., & Schelstraete, M. A. (2015).
Stimulating parents' self-efficacy beliefs or verbal responsiveness: Which is the
best way to decrease children's externalizing behaviors? Behaviour Research
and Therapy, 72, 38-48. doi:10.1016/j.brat.2015.06.012
Rowe, M. L. (2012). A longitudinal investigation of the role of quantity and quality of
child-directed speech in vocabulary development. Child Development, 83(5),
1762-1774. doi:10.1111/j.1467-8624.2012.01805.x
Rowe, M. L. (2017). Contributions of parent-child interactions to child language
outcomes in diverse samples: A discussion. Paper presented at the International
Association for the Study of Child Language (IASCL), Lyon, France.
Roy, P., Kersley, H., & Law, J. (2005). The Sure Start language measure
standardisation study. Retrieved from
http://tna.europarchive.org/20070101101348/http://www.dfes.gov.uk/research/
programmeofresearch/projectinformation.cfm?projectid=14628&resultspage=1
website:
Rubin, K. H., & Mills, R. S. L. (1991). Conceptualising developmental pathways to
internalising disorders in childhood. Canadian Journal of Behavioural Science-
304
Revue Canadienne Des Sciences Du Comportement, 23(3), 300-317.
doi:10.1037/h0079019
Ruffman, T., Slade, L., & Crowe, E. (2002). The relation between children's and
mothers' mental state language and theory-of-mind understanding. Child
Development, 73(3), 734-751. doi:10.1111/1467-8624.00435
Ruffman, T., Slade, L., Devitt, K., & Crowe, E. (2006). What mothers say and what
they do: The relation between parenting, theory of mind, language and
conflict/cooperation. British Journal of Developmental Psychology, 24, 105-124.
doi:10.1348/026151005x82848
Rutter, M. (1967). A children’s behaviour questionnaire for completion by teachers:
Preliminary findings. Journal of Child Psychology and Psychiatry, 8, 1-11.
Rutter, M., & Lord, C. (1987). Language disorders associated with psychiatric
disturbance. Language Development and Disorders, 206-233.
Salmon, K., O’Kearney, R., Reese, E., & Fortune, C. A. (2016). The role of language
skill in child psychopathology: Implications for intervention in the early years.
Clinical Child and Family Psychology Review, 19(4), 352-367.
Salmon, K., & Reese, E. (2016). The benefits of reminiscing with young children.
Current Directions in Psychological Science, 25(4), 233-238.
doi:10.1177/0963721416655100
Sameroff, A. J. (1998). Environmental risk factors in infancy. Pediatrics, 102(5 Suppl
E), 1287-1292.
Sameroff, A. J. (2009). The Transactional Model. In A. Sameroff (Ed.), The
Transactional Model of Development: How Children and Contexts Shape Each
Other. (pp. 3-21). Washington, DC, US: American Psychological Association.
305
Sameroff, A. J., & Chandler, M. J. (1975). Reproductive risk and the continuum of
caretaking casualty. Review of Child Development Research, 4, 187-244.
Sanders, M. R., Dadds, M. R., Johnston, B. M., & Cash, R. (1992). Childhood
depression and conduct disorder 1. Behavioural, affective, and cognitive aspects
of family problem-solving interactions. Journal of Abnormal Psychology,
101(3), 495-504. doi:10.1037/0021-843x.101.3.495
Sanders, M. R., & Markie-Dadds, C. (1992). Toward a technology of prevention of
disruptive behaviour disorders: The role of behavioural family intervention.
Behaviour Change, 9(3), 186-200.
Scaramella, L. V., & Leve, L. D. (2004). Clarifying parent-child reciprocities during
early childhood: The early childhood coercion model. Clinical Child and Family
Psychology Review, 7(2), 89-107. doi:10.1023/B:CCFP.0000030287.13160.a3
Schoon, I., Parsons, S., Rush, R., & Law, J. (2010). Children's language ability and
psychosocial development: A 29-year follow-up study. Pediatrics, 126(1), e73-
e80. doi:10.1542/peds.2009-3282
Semel, E., Wiig, E. H., & Secord, W. A. (2006). The Clinical Evaluation of Language
Fundamentals-Preschool, Second Edition (CELF-P2): Australian Standardised
Edition (2 ed.). NSW, Australia: Harcourt Assessment Inc.
Shaw, D. S., & Gross, H. E. (2008). What we have Learned about Early Childhood and
the Development of Delinquency. Long View of Crime: A Synthesis of
Longitudinal Research, 79-127. doi:10.1007/978-0-387-71165-2_3
Shimpi, P. M., & Huttenlocher, J. (2007). Re-directive labels and early vocabulary
development. Journal of Child Language, 34(4), 845-859.
doi:http://dx.doi.org/10.1017/S0305000907008112
306
Shinohara, R., Sugisawa, Y., Tong, L., Tanaka, E., Watanabe, T., Onda, Y., . . . Anme,
T. (2010). The trajectory of children's social competence from 18 months to 30
months of age and their mother's attitude towards the praise. Journal of
Epidemiology, 20 Suppl 2, S441-S446.
Sitnick, S. L., Shaw, D. S., Gill, A., Dishion, T., Winter, C., Waller, R., . . . Wilson, M.
(2015). Parenting and the Family Check-Up: Changes in observed parent-child
interaction following early childhood intervention. Journal of Clinical Child and
Adolescent Psychology, 44(6), 970-984. doi:10.1080/15374416.2014.940623
Smedler, A. C., Hjern, A., Wiklund, S., Anttila, S., & Pettersson, A. (2015). Programs
for Prevention of Externalizing Problems in Children: Limited Evidence for
Effect Beyond 6 Months Post Intervention. Child and Youth Care Forum, 44(2),
251-276. doi:10.1007/s10566-014-9281-y
Smith, J., Levickis, P., Eadie, P., Bretherton, L., Conway, L., & Goldfeld, S. (2017).
Concurrent associations between maternal behaviours and infant communication
within a cohort of women and their infants experiencing adversity. International
Journal of Speech-Language Pathology, 1-12.
doi:10.1080/17549507.2017.1329458
Snow, P. C., Woodward, M., Mathis, M., & Powell, M. B. (2016). Language
functioning, mental health and alexithymia in incarcerated young offenders.
International Journal of Speech-Language Pathology, 18(1), 20-31.
doi:10.3109/17549507.2015.1081291
Snowling, M. J., Duff, F. J., Nash, H. M., & Hulme, C. (2015). Language profiles and
literacy outcomes of children with resolving, emerging, or persisting language
impairments. Journal of Child Psychology and Psychiatry, 57(12), 1360-1369.
doi:10.1111/jcpp.12497
307
Song, L. L., Spier, E. T., & Tamis-Lemonda, C. S. (2014). Reciprocal influences
between maternal language and children's language and cognitive development
in low-income families. Journal of Child Language, 41(2), 305-326.
doi:10.1017/s0305000912000700
Spinrad, T. L., & Stifter, C. A. (2006). Toddlers' empathy-related responding to distress:
Predictions from negative emotionality and maternal behavior in infancy.
Infancy, 10(2), 97-121. doi:10.1207/s15327078in1002_1
St Clair, M. C., Pickles, A., Durkin, K., & Conti-Ramsden, G. (2011). A longitudinal
study of behavioral, emotional and social difficulties in individuals with a
history of specific language impairment. Journal of Communication Disorders,
44(2), 186-199. doi:10.1016/j.jcomdis.2010.09.004
Stein, A., Malmberg, L. E., Sylva, K., Barnes, J., Leach, P., & Team, F. (2008). The
influence of maternal depression, caregiving, and socioeconomic status in the
post-natal year on children's language development. Child: Care, Health, and
Development, 34(5), 603-612. doi:10.1111/j.1365-2214.2008.00837.x
Stolt, S., Haataja, L., Lapinleimu, H., & Lehtonen, L. (2009). The early lexical
development and its predictive value to language skills at two years in very low
birthweight children. Journal of Communication Disorders, 42(2), 107-123.
doi:10.1016/j.jcomdis.2008.10.002
Stone, L. L., Otten, R., Engels, R., Vermulst, A. A., & Janssens, J. (2010). Psychometric
properties of the parent and teacher versions of the Strengths and Difficulties
Questionnaire for four to 12 year olds: A review. Clinical Child and Family
Psychology Review, 13(3), 254-274. doi:10.1007/s10567-010-0071-2
308
Supplee, L. H., Shaw, D. S., Hailstones, K., & Hartman, K. (2004). Family and child
influences on early academic and emotion regulatory behaviors. Journal of
School Psychology, 42(3), 221-242. doi:10.1016/j.jsp.2004.02.001
Swenson, S., Ho, G. W. K., Budhathoki, C., Belcher, H. M. E., Tucker, S., Miller, K., &
Gross, D. (2016). Parents' use of praise and criticism in a sample of young
children seeking mental health services. Journal of Pediatric Health Care,
30(1), 49-56. doi:10.1016/j.pedhc.2015.09.010
Tamis-LeMonda, C. S., Baumwell, L., & Cristofaro, T. (2012). Parent-child
conversations during play. First Language, 32(4), 413-438.
doi:10.1177/0142723711419321
Tamis-LeMonda, C. S., & Bornstein, M. H. (1991). Individual variation,
correspondence, stability, and change in mother and toddler play. Infant
Behavior and Development, 14(2), 143-162.
Tamis-LeMonda, C. S., Bornstein, M. H., & Baumwell, L. (2001). Maternal
responsiveness and children's achievement of language milestones. Child
Development, 72(3), 748-767.
Tamis-LeMonda, C. S., Shannon, J. D., Cabrera, N. J., & Lamb, M. E. (2004). Fathers
and mothers at play with their two and three year olds: Contributions to
language and cognitive development. Child Development, 75(6), 1806-1820.
doi:10.1111/j.1467-8624.2004.00818.x
Taylor, C. L., Zubrick, S. R., & Rice, M. L. (2013). Population and public health
perspectives on late language emergence at 24 months as a risk indicator for
language impairment at 7 years. In L. Rescorla & P. Dale (Eds.), Late talkers:
language development, interventions and outcomes. Baltimore, USA: Brooks
Publishing Co.
309
Taylor, N., Donovan, W., Miles, S., & Leavitt, L. (2009). Maternal control strategies,
maternal language usage and children's language usage at two years. Journal of
Child Language, 36(2), 381-404.
doi:http://dx.doi.org/10.1017/S0305000908008969
Tempel, A. B., Wagner, S. M., & McNeil, C. B. (2009). Parent-child interaction therapy
and language facilitation: The role of parent-training on language development.
The Journal of Speech and Language Pathology–Applied Behavior Analysis,
3(2-3), 216.
Thomas, R., & Zimmer-Gembeck, M. J. (2007). Behavioral outcomes of parent-child
interaction therapy and Triple-P Positive Parenting Program: A review and
meta-analysis. Journal of Abnormal Child Psychology, 35(3), 475-495.
doi:10.1007/s10802-007-9104-9
Tomasello, M. (1992). The social bases of language acquisition. Social Development,
1(1), 67-87.
Tomasello, M., & Akhtar, N. (1995). Two-year-olds use pragmatic cues to differentiate
reference to objects and actions. Cognitive Development, 10(2), 201-224.
Tomasello, M., & Farrar, M. J. (1986). Joint attention and early language. Child
Development, 57(6), 1454-1463. doi:10.1111/j.1467-8624.1986.tb00470.x
Tomasello, M., Kruger, A., & Ratner, H. (1993). Cultural learning. Behavioral and
Brain Sciences, 16(3), 495-511. Doi: 10.1017/S0140525X0003123X
Tomasello, M., & Todd, J. (1983). Joint attention and lexical acquisition style. First
Language, 4, 197-212.
Tomblin, J. B., Records, N. L., Buckwalter, P., Zhang, X. Y., Smith, E., & O'Brien, M.
(1997). Prevalence of specific language impairment in kindergarten children.
Journal of Speech Language, and Hearing Research, 40(6), 1245-1260.
310
Tomblin, J. B., Zhang, X., Buckwalter, P., & Catts, H. (2000). The association of
reading disability, behavioral disorders, and language impairment among
second-grade children. Journal of Child Psychology and Psychiatry, and Allied
Disciplines, 41(4), 473-482.
Toppelberg, C. O., & Shapiro, T. (2000). Language disorders: A 10 year research
update review. Journal of The American Academy of Child and Adolescent
Psychiatry, 39(2), 143-152.
Treyvaud, K., Anderson, V. A., Howard, K., Bear, M., Hunt, R. W., Doyle, L. W., . . .
Anderson, P. J. (2009). Parenting behavior is associated with the early neuro-
behavioral development of very pre-term children. Pediatrics, 123(2), 555-561.
doi:10.1542/peds.2008-0477
Trueswell, J. C., Lin, Y., Armstrong, B., Cartmill, E. A., Goldin-Meadow, S., &
Gleitman, L. R. (2016). Perceiving referential intent: Dynamics of reference in
natural parent-child interactions. Cognition, 148, 117-135.
doi:10.1016/j.cognition.2015.11.002
UCLA. (2015). How can I explain a continuous by continuous interaction? Stata FAQ.
Uziel, L. (2010). Look at me, I'm happy and creative: The effect of impression
management on behavior in social presence. Personality and Social Psychology
Bulletin, 36(12), 1591-1602. doi:10.1177/0146167210386239
Vallotton, C., & Ayoub, C. (2011). Use your words: The role of language in the
development of toddlers’ self-regulation. Early Childhood Research Quarterly,
26(2), 169-181. doi:10.1016/j.ecresq.2010.09.002 van Balkom, H., Verhoeven, L., & van Weerdenburg, M. (2010). Conversational
behaviour of children with Developmental Language Delay and their caretakers.
311
International Journal of Language and Communication Disorders, 45(3), 295-
319. doi:10.3109/13682820902994226 van Daal, J., Verhoeven, L., & van Balkom, H. (2007). Behaviour problems in children
with language impairment. Journal of Child Psychology and Psychiatry, 48(11),
1139-1147. doi:10.1111/j.1469-7610.2007.01790.x van Zeijl, J., Mesman, J., Stolk, M. N., Alink, L. R. A., van Ijzendoorn, M. H.,
Bakermans-Kranenburg, M. J., . . . Koot, H. M. (2006). Terrible ones?
Assessment of externalizing behaviors in infancy with the Child Behavior
Checklist. Journal of Child Psychology and Psychiatry, 47(8), 801-810.
doi:10.1111/j.1469-7610.2006.01616.x
Vella, L. R., Butterworth, R. E., Johnson, R., & Law, G. U. (2015). Parents' experiences
of being in the Solihull Approach parenting group, 'Understanding Your Child's
Behaviour': an interpretative phenomenological analysis. Child Care Health and
Development, 41(6), 882-894. doi:10.1111/cch.12284
Vigil, D. C., Hodges, J., & Klee, T. (2005). Quantity and quality of parental language
input to late-talking toddlers during play. Child Language Teaching and
Therapy, 21(2), 107-122.
Vigil, J. M. (2009). A socio-relational framework of sex differences in the expression of
emotion. Behavioral and Brain Sciences, 32(5), 375.
doi:10.1017/s0140525x09991075
Vittinghoff, E., McCulloch, C. E., Glidden, D. V., & Shiboski, S. C. (2008). Linear and
Non-Linear Regression Methods in Epidemiology and Biostatistics. In C. R.
Rao, J. P. Miller, & D. C. Rao (Eds.), Epidemiology and Medical Statistics (Vol.
27, pp. 148-186). Amsterdam: Elsevier Science Bv.
312
Vuksanovic, J., & Bjekic, J. (2013). Developmental relationship between language and
joint attention in late-talkers. Research in Developmental Disabilities, 34(8),
2360-2368. doi:10.1016/j.ridd.2013.04.017
Wake, M., Levickis, P., Tobin, S., Zens, N., Law, J., Gold, L., . . . Reilly, S. (2012).
Improving outcomes of pre-school language delay in the community: Protocol
for the Language for Learning randomised controlled trial. Bmc Pediatrics, 12,
1-11. doi:10.1186/1471-2431-12-96
Wake, M., Tobin, S., Girolametto, L. E., Ukoumunne, O. C., Gold, L., Levickis, P., . . .
Reilly, S. (2011). Outcomes of population-based language promotion for slow-
to-talk toddlers at ages two and three years: Let's Learn Language cluster
randomised controlled trial. BMJ, 343, d4741. doi:10.1136/bmj.d4741
Wallace, I. F., Berkman, N. D., Watson, L. R., Coyne-Beasley, T., Wood, C. T., Cullen,
K., & Lohr, K. N. (2015). Screening for speech and language delay in children
five years old and younger: A systematic review. Pediatrics, 2014-3889.
Warren, S. F., & Brady, N. C. (2007). The role of maternal irresponsivity in the
development of children with intellectual disabilities. Mental Retardation and
Developmental Disabilities Research Reviews, 13(4), 330-338.
doi:10.1002/mrdd.20177
Watt, N., Wetherby, A. M., & Shumway, S. (2006). Prelinguistic predictors of language
outcome at three years of age. Journal of Speech, Language, and Hearing
Research, 49(6), 1224-1237. doi:http://dx.doi.org/10.1044/1092-
4388%282006/088%29
Webster-Stratton, C., & Reid, M. (2010). The Incredible Years Program for Children
from Infancy to Pre-adolescence: Prevention and Treatment of Behavior
Problems. New York: Springer.
313
Webster-Stratton, C., & Taylor, T. (2001). Nipping early risk factors in the bud:
Preventing substance abuse, delinquency, and violence in adolescence through
interventions targeted at young children (0–8 years). Prevention Science, 2(3),
165-192.
Wetherby, A. M., Cain, D. H., Yonclas, D. G., & Walker, V. G. (1988). Analysis of
intentional communication of normal children from the prelinguistic to the
multiword stage. Journal of Speech and Hearing Research, 31(2), 240-252.
Wetherby, A. M., Yonclas, D. G., & Bryan, A. A. (1989). Communicative profiles of
preschool-children with handicaps - Implications for early identification.
Journal of Speech and Hearing Disorders, 54(2), 148-158.
Wheeden, C., & Fewell, R. (1995). The play and learn strategies (PALS) program.
Unpublished manuscript. The Debbie Institute of the University of Miami Miller
School of Medicine. Miami, FL.
White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a
direct test for heteroskedasticity Econometrica, 48, 817–830.
Whitehouse, A. J. O., Robinson, M., & Zubrick, S. R. (2011). Late-talking and the risk
for psychosocial problems during childhood and adolescence. Pediatrics,
128(2), e324-e332. doi:10.1542/peds.2010-2782
Whitehurst, G. J., Fischel, J. E., Lonigan, C. J., Valdez-Menchaca, M. C., DeBaryshe,
B. D., & Caulfield, M. B. (1988). Verbal interaction in families of normal and
expressive-language-delayed children. Developmental Psychology, 24(5), 690-
699. doi:http://dx.doi.org/10.1037/0012-1649.24.5.690
Williams, M. E., Hastings, R., Charles, J. M., Evans, S., & Hutchings, J. (2017).
Parenting for autism, language, and communication evaluation study
314
(PALACES): Protocol for a pilot randomised controlled trial. BMJ, 7(2), 9.
doi:10.1136/bmjopen-2016-014524
Winsler, A. (2009). Still talking to ourselves after all these years: A review of current
research on private speech. Private Speech, Executive Functioning, and the
Development of Verbal Self-Regulation, 3-41.
Winsler, A., Diaz, R. M., McCarthy, E. M., Atencio, D. J., & Chabay, L. A. (1999).
Mother–child interaction, private speech, and task performance in pre-school
children with behavior problems. Journal of Child Psychology and Psychiatry,
40(6), 891-904. doi:10.1111/1469-7610.00507
Yew, S. G. K., & O'Kearney, R. (2013). Emotional and behavioural outcomes later in
childhood and adolescence for children with specific language impairments:
meta-analyses of controlled prospective studies. Journal of Child Psychology
and Psychiatry, and Allied Disciplines, 54(5), 516-524. doi:10.1111/jcpp.12009
Yew, S. G. K., & O'Kearney, R. (2015a). Early language impairments and
developmental pathways of emotional problems across childhood. International
Journal of Language and Communication Disorders, 50(3), 358-373.
Yew, S. G. K., & O'Kearney, R. (2015b). The role of early language difficulties in the
trajectories of conduct problems across childhood. Journal of Abnormal Child
Psychology, 43(8), 1515-1527. doi:10.1007/s10802-015-0040-9
Yu, C., & Ballard, D. H. (2007). A unified model of early word learning: Integrating
statistical and social cues. Neurocomputing, 70(13), 2149-2165.
Yu, C., & Smith, L. B. (2012). Embodied attention and word learning by toddlers.
Cognition, 125(2), 244-262. doi:10.1016/j.cognition.2012.06.016
Zambrana, I. M., Pons, F., Eadie, P., & Ystrom, E. (2014). Trajectories of language
delay from age three to five: Persistence, recovery and late onset. International
315
Journal of Language and Communication Disorders, 49(3), 304-316.
doi:10.1111/1460-6984.12073
Zentall, S. R., & Morris, B. J. (2010). "Good job, you're so smart": The effects of
inconsistency of praise type on young children's motivation. Journal of
Experimental Child Psychology, 107(2), 155-163.
doi:10.1016/j.jecp.2010.04.015
Zimmerman, I. L., Steiner, V. G., & Pond, R. E. (2002). The Preschool Language Scale,
4th Edition: The Psychological Corporation.
Zisser, A., & Eyberg, S. (2010). Parent-child interaction therapy and the treatment of
disruptive behavior disorders. Evidence-based psychotherapies for children and
adolescents, 2, 179-193.
Zubrick, S. R., Taylor, C. L., & Christensen, D. (2015). Patterns and predictors of
language and literacy abilities four to ten years in the Longitudinal Study of
Australian Children. Plos One, 10(9), 29. doi:10.1371/journal.pone.0135612
Zubrick, S. R., Taylor, C. L., Rice, M. L., & Slegers, D. W. (2007). Late language
emergence at 24 months: An epidemiological study of prevalence, predictors,
and covariates. Journal of Speech, Language, and Hearing Research, 50(6),
1562-1592. doi:10.1044/1092-4388(2007/106)
316
Appendices
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
Minerva Access is the Institutional Repository of The University of Melbourne
Author/s: Conway, Laura Jane
Title: What mother-child talk tells us about children’s language, social, emotional and behavioural development: a community-based study of slow-to-talk children
Date: 2017
Persistent Link: http://hdl.handle.net/11343/198290
File Description: What mother-child talk tells us about children’s language, social, emotional and behavioural development: a community-based study of slow-to-talk children
Terms and Conditions: Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the copyright owner. The work may not be altered without permission from the copyright owner. Readers may only download, print and save electronic copies of whole works for their own personal non-commercial use. Any use that exceeds these limits requires permission from the copyright owner. Attribution is essential when quoting or paraphrasing from these works.