The Texas Medical Center Library DigitalCommons@TMC
UT SON Dissertations (Open Access) School of Nursing
Spring 5-2019
Impact of Adult Attachment Style on Outcome Trajectories in a Parent-Focused Early Language Development Intervention
Cary Cain University of Texas Health Science Center at Houston-Cizik School of Nursing
Follow this and additional works at: https://digitalcommons.library.tmc.edu/uthson_etd
Part of the Nursing Commons
Recommended Citation Cain, Cary, "Impact of Adult Attachment Style on Outcome Trajectories in a Parent-Focused Early Language Development Intervention" (2019). UT SON Dissertations (Open Access). 33. https://digitalcommons.library.tmc.edu/uthson_etd/33
This is brought to you for free and open access by the School of Nursing at DigitalCommons@TMC. It has been accepted for inclusion in UT SON Dissertations (Open Access) by an authorized administrator of DigitalCommons@TMC. For more information, please contact [email protected].
IMPACT OF ADULT ATTACHMENT STYLE ON OUTCOME TRAJECTORIES IN
A PARENT-FOCUSED EARLY LANGUAGE DEVELOPMENT INTERVENTION
A DISSERTATION
SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN NURSING
THE UNIVERSITY OF TEXAS HEALTH SCIENCE CENTER AT HOUSTON
CIZIK SCHOOL OF NURSING
BY
CARY M. CAIN, MPH, BSN, RN
MAY, 2019
Dedication
In memory of my father, William David Cain and grandparents, William Maurice and Lona Nadine Cain (Papa and Memaw) and James Edwin and Doris Gayle Jones (Daddy Ed and Mama Gayle)
Therefore, since we are surrounded by so great a cloud of witnesses, let us also lay aside every weight, and sin which clings so closely, and let us run with endurance the race that is set before us, looking to Jesus, the founder and perfecter of our faith.
To the only God, our Savior, through Jesus Christ our Lord, be glory, majesty, dominion, and authority, before all time and now and forever. Amen.
iii
Acknowledgements
This work would not have been possible without the generosity and guidance of the Robert Wood Johnson Foundation Future of Nursing Scholars Program, secured at
Cizik School of Nursing by Dr. Geri Wood. I am also grateful to the McGovern
Foundation for generously funding the Cizik School of Nursing Accelerated PhD
Program McGovern Scholarship.
I owe much gratitude to my parents for their love, support, and sacrifices so that I can be at this point in my life. I thank them for instilling Christian values in me and for teaching me the importance of diligence and hard work. I am grateful for my mom’s continual prayers and encouragement. I thank my step-father for his support and especially his advice to me as I began this PhD journey. I am grateful for such amazing friends and family – Jana, Kyler, Ganesh, Orhue, Audrey, Marisa, Chad, Meredith, Jeff, and even Libby (my dog) – who have supported me through laughter and tears. I’m thankful for the prayers and support of my Christ the King community group and friends.
I have been blessed to work with an incredible team at Texas Children’s and have learned so much from my colleagues – Dr. Greeley, Kim, Beth, Nancy, Maura, Yen, Angie, Lisa,
John, and Richard – I appreciate their patience with me and allowing me to learn while working. I couldn’t have made it through this program without my sidekick (and unofficial advisor) Hilary and classmates Mary and Lai – the four of us are carrying each other through the finish line and I know we will continue to support each other as we continue our journeys.
I thank my advisor and dissertation chairperson, Dr. Cathy Rozmus, for her guidance, patience, and support through the doctoral program. I would also like to thank
iv
Dr. Nikhil Padhye for his participation on my dissertation committee and invaluable contributions and advice. Thanks to Dr. Diane Santa Maria and Dr. Geri Wood for serving on my candidacy committee and for their valuable mentorship throughout my time in this program. In addition, I am grateful to the dedicated Cizik School of Nursing
PhD faculty and staff, who are too numerous to count.
Special gratitude is owed to my external advisor, Dr. Dorothy Mandell, for her willingness to mentor me, graciously giving of her time and patience to provide wisdom, guidance, and instruction. I admire how she applies her knowledge, and even her commitment in her personal life, to improve the lives of young children and families. I also would not have been able to complete the analysis in this dissertation without the guidance and expertise of Dr. Katharine Buek – I am so grateful that we crossed paths.
Lastly, I wish to acknowledge the upWORDS team and families who contributed to this dissertation by participating in the upWORDS program. My hope is that by better understanding relationships in aggregate, we will be able to improve the lives of children and their families.
v
Cary Cain
Impact of Adult Attachment Style on Outcome Trajectories in a Parent-focused
Early Language Development Intervention
May, 2019
Abstract
Background: Disparities in early language development have been noted for children from low-income homes, which can impact their educational and health trajectory. The quantity and quality of language exposure in the home environment influences a child’s language development. Adult attachment style is a predictor of parenting behavior and could impact critical interactions between the parent and child. Evidence demonstrates that differences in adult attachment style impact program participation and outcomes in early childhood parenting programs delivered through home visitation.
Aims: The aims of this study were to 1) explore the association of adult attachment style on intervention outcome trajectories from a group based, parent-focused early language development intervention and 2) evaluate the association of adult attachment style on intervention attendance.
Methods: This exploratory study used data from a program evaluation of a group-based, parent-focused early language development intervention to examine the associations of adult attachment style on intervention attendance and parent-child interaction outcomes measured longitudinally. Parents of children (ages 0-24 months) were recruited to participate in a 13-week early language development program. A sample of program participants completed questionnaire data that included adult attachment style, depression symptoms, and sociodemographic information. Confirmatory and exploratory factor
vi
analyses were used to find three-factors of adult attachment style; relationship anxiety, discomfort with closeness, and relationships as secondary to achievements. Parent-child interaction was assessed by digital language processors recording child-directed speech and reciprocal interactions in the home environment weekly and parents reported average shared book reading weekly. Latent growth trajectories of parent-child interaction variables were identified through latent class growth analyses and regressed onto adult attachment style factors and covariates in order to determine the association of adult attachment style on the outcome trajectories. In addition, intervention attendance was regressed onto adult attachment style factors and covariates to understand the impact of adult attachment style on intervention attendance.
Results: This study found three latent growth trajectories (high, middle, and low) for child-directed speech and two latent growth trajectories (high and low) for reciprocal interaction and shared book reading. This study found that discomfort with closeness had strong odds (OR: 2.602; p < .05) of attending 10 or more intervention sessions; however, discomfort with closeness was significantly associated with lower baseline and lower growth trajectory of reciprocal interactions between the parent and child (OR: .336, p <
.001).
Conclusions: Interventions supporting early language development are often targeted towards parents, as healthy development is fostered by parents’ provision of a supportive and cognitively stimulating home environment. Findings from this study suggest that adult attachment style does impact the effects of parent-focused language development programs and thus additional supportive measures may be needed for parents with insecure adult attachment styles to fully benefit from these types of programs.
vii
Keywords: Early Language Development, Adult Attachment Style, Latent Class Growth
Analysis, Parent-Child Interaction, Child Development
viii
Table of Contents
APPROVAL PAGE ...... ii
DEDICATION ...... iii
ACKNOLEDGEMENTS ...... iv
ABSTRACT ...... vi
SUMMARY OF STUDY ...... 1
PROPOSAL ...... 3
Specific Aims ...... 4
Background and Significance ...... 7
Research Design and Methods ...... 12
Research Subject Risk and Protection ...... 21
Literature Cited ...... 22
DISSERTATION MANUSCRIPT
Manuscript #1
Impact of Adult Attachment Style on Outcome Trajectories in a Parent-focused
Early Language Development Intervention ...... 33
Letter to the Editor ...... 91
Manuscript #2
Impact of Adult Attachment Style on Outcome Trajectories in a Parent-focused
Early Language Development Intervention ...... 92
APPENDIXES
A. University of Texas Health Science Center CPHS Approval ...... 151
B. Baylor College of Medicine Approval ...... 154
C. Data Dictionary ...... 161
D. Adult Attachment Style Questionnaire ...... 170
E. Edinburgh Postnatal Depression Scale ...... 173
F. Collaborative Institutional Training Initiative Certificates ...... 176
CURRICULUM VITAE ...... 185 1
Summary of the Study
The proposal entitled, “Impact of Adult Attachment Style on Outcome
Trajectories in a Parent-focused Early Language Development Intervention” was a study conducted to investigate the associations of adult attachment style on outcome trajectories of child-directed speech, reciprocal interaction, shared book reading, and intervention attendance in an early language development intervention. Approval for this study was sought from the Committee for the Protection of Human Subjects (CPHS) at the University of Texas Health Science Center at Houston’s Institutional Review Board and exempt status was granted in July, 2018. The parent study was approved by Baylor
College of Medicine’s Institutional Review Board in July, 2017.
This exploratory study used data from a program evaluation of a parent-focused early language development intervention (parent study). The present study utilized deidentified program data on longitudinal outcomes measured by digital language processors, parent-report data, and program attendance data and deidentified data on adult attachment style, depression, and sociodemographic characteristics from a sample of program participants that participated in the parent study. The specific aims of the study were to: 1) Examine the association of adult attachment style on the growth trajectories of parent-child interaction measured longitudinally in the home environment and 2) Evaluate the association of adult attachment style on intervention attendance.
Descriptive statistics were used to analyze demographic characteristics and evaluate the data for completeness. Confirmatory and exploratory factor analyses were used to determine the factor structure of the Attachment Style Questionnaire among parents of young children in Houston, Texas using data from parents in this study, a 2
sample of parents of young children participating in an early head start program, and a sample of parents of young children waiting in an emergency center. The factor scores were retained and used in this study for the analyses. Correlation analyses were conducted of the adult attachment style factor scores, depression symptoms, parent age, educational attainment, income level, child age, and intervention attendance. Latent class growth analyses with predictors was conducted using the 3-step Vermunt method.
Outcome trajectories of parent-child interaction variables were determined and regressed onto adult attachment style factors and covariates to determine the association of adult attachment style on outcome trajectories. Intervention attendance was regressed onto adult attachment style factors and covariates to understand the impact of adult attachment style on intervention attendance.
Two manuscripts were written on topics that were pertinent to the dissertation study. Manuscript #1 presented and discussed the methods and results of this exploratory study. The primary changes from the dissertation proposal was the phrasing of the aims with the Attachment Style Questionnaire factors and the method of analysis for the second aim from multiple regression to logistic regression, which was changed due to the distribution of the data.
Manuscript #2 was a systematic review of the efficacy of early language development interventions on language, cognitive, and social outcomes. Results from the systematic review are presented along with a discussion on the findings from the literature.
Appendices A-F contain supplemental information related to the studies.
3
IMPACT OF ADULT ATTACHMENT STYLE IN A PARENT-FOCUSED EARLY LANGUAGE DEVELOPMENT INTERVENTION
A PROPOSAL
SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN NURSING
THE UNIVERSITY OF TEXAS HEALTH SCIENCE CENTER AT HOUSTON
CIZIK SCHOOL OF NURSING
BY
CARY M. CAIN, MPH, BSN, RN
JULY, 2018 4
Proposal and Specific Aims Significant disparities in early language development are noted for children growing up in low-income homes, which can impact their educational and health trajectory throughout the life course (Duncan & Magnuson, 2013; Hart & Risley, 1995;
Hoff, 2013). It is widely recognized that the first few years of life are critical for brain development, which is facilitated through stable and responsive caregiving (Casey,
Giedd, & Thomas, 2000; Hurt & Betancourt, 2017; Knudsen, 2004). Conversely, adverse parent-child interactions, including low stimulation and neglect, lead to poor cognitive, behavioral, language, and health outcomes (Shonkoff & Phillips, 2000). The goal of this research is to investigate the association of adult attachment style on parent-child interaction outcomes and parent participation in a parent-focused early language intervention.
Studies have shown adult attachment style is a predictor of parenting behavior; parents with secure attachment style are likely to contribute to parent-child engagement because of consistent and responsive parenting behaviors (Haley & Stansbury, 2003;
Rholes, Simpson, Blakely, Lanigan, & Allen, 1997). Parent-focused early language interventions have promising evidence in promoting child development; however, the interventions have varied effects (Roberts & Kaiser, 2011). The impact of adult attachment style in parent-focused early language interventions, to date, has not been published in the literature. Studies investigating the association of adult attachment style on early childhood home visitation programs found that parents have varied engagement and program benefit due to attachment style differences (Duggan, Berlin, Cassidy,
Burrell, & Tandon, 2009). Understanding parental influences in interventions to foster healthy child development, which are targeted primarily toward the parent, are critical to 5
eliminate educational and health disparities throughout the child’s life course (National
Scientific Council on the Developing Child, 2004).
The long-term goal of this research is to understand how to decrease early language disparities through promotion of stable and responsive home language environments. The overall objective for this study is to explore the association of adult attachment style in a parent-focused early language intervention. This intervention uses a group-based parent support model to deliver curriculum and feedback to improve the home language environment and to foster early language development. To attain the overall objective for this study, the following two specific aims are proposed:
Specific Aim 1. Evaluate the association of adult attachment style on the intervention effects on parent-child interaction measured longitudinally in the home language environment.
Hypothesis 1.1a. Participants with secure adult attachment style will have greater increases in child-directed speech (adult word count) than participants with insecure attachment style.
Hypothesis 1.1b. Of the participants with insecure attachment styles, those with anxious attachment style will show greater increases in child-directed speech compared to participants with avoidant attachment style.
Hypothesis 1.2a. Participants with secure adult attachment style will have greater increases in serve and return interaction (conversational turns) than participants with insecure attachment style. 6
Hypothesis 1.2b. Of the participants with insecure attachment styles, those with anxious attachment style will show greater increases in serve and return interaction compared to participants with avoidant attachment style.
Hypothesis 1.3a. Participants with secure adult attachment style will have greater increases in the amount of time spent in shared book reading than participants with insecure attachment style.
Hypothesis 1.3b. Of the participants with insecure attachment styles, those with anxious attachment style will show greater increases the amount of time spent in shared book reading compared to participants with avoidant attachment style.
Specific Aim 2. Evaluate the association of adult attachment style on intervention participation.
Hypothesis 2a. Participants with a secure adult attachment style will have higher intervention participation than participants with insecure adult attachment style.
Hypothesis 2b. Of the participants with insecure attachment styles, those with anxious attachment style will have higher intervention participation than participants with avoidant attachment style.
Most early language experiences occur in the context of the home environment
(Kalil, Ziol-Guest, Ryan, & Markowitz, 2016). Understanding the associations of adult attachment style on the home language environment outcomes are vital to inform future targeted interventions which foster early childhood development (Bowman, Donovan, &
Burns, 2001).
7
Significance. Early language development. Disparities are noted among children from differing socioeconomic status (SES) backgrounds in the quantity and quality of their exposure to language during the critical time of early cognitive development (Gilkerson,
Richards, Warren, et al., 2017; Hart & Risley, 1995; Hirsh-Pasek et al., 2015; Rowe, 2008).
Hart and Risely (1995) found that before reaching the age of 4, children from lower-income households heard 32 million fewer words and more negative or harsh language compared with children from households of higher-income levels. A recent study confirms these earlier findings and found significant cognitive and language differences among infants from low-income households compared with those from relatively higher-income households, even at 1 year of age (Hurt & Betancourt, 2017). These deficits can impact a child’s educational trajectory and attainment, which is considered a social determinant of health (Kuhl, 2011; Marmot, 2005; Sénéchal & Lefevre, 2002). Language has a critical role in a child’s cognitive and social development, as it provides a means of communication, methods of obtaining knowledge, and fosters future inquiry (Connell &
Prinz, 2002; Song, Spier, & Tamis-Lemonda, 2014).
Parental involvement. Parental involvement and engagement in their child’s language attainment is recognized as critical for healthy development (Raikes et al., 2006;
Sénéchal & Lefevre, 2002). Kuhl (2011) demonstrated that infant language development is facilitated by social interaction, such as reading and talking. Studies have shown that frequency of reading to a child regularly, as well as having accessible children's books in the household, were associated with positive child outcomes, such as early academic success (Pati, Hashim, Brown, Fiks, & Forrest, 2011; Zuckerman & Augustyn, 2011). A nurturing home environment facilitates early language development through positive 8
interactions between the primary caregiver and the child and stimulates the child’s cognitive and social development (Siddiqi, Irwin, & Hertzman, 2007). Another influence on child development is social support. Parents with sufficient social support show greater increases in parent-child interaction compared to parents that lack supportive and cohesive relationships (Green, Furrer, & McAllister, 2007). Conversely, negative experiences early in life can impair development. If a young child is exposed to prolonged adversity in the absence of an adequate buffer, such as a supportive and responsive caregiver, disruptions are noted in brain development and the child’s stress regulating responses (Jutte, Miller, &
Erickson, 2015; Rushton & Kraft, 2014; Shonkoff et al., 2012).
Efforts to foster early language development have been targeted toward parents and are part of the larger effort to reduce toxic stress through strengthening the parent-child relationship through knowledge of child development and intentional interactions
(Greenwood et al., 2017). Facilitating an environment that promotes parent-child interaction through parental understanding of the importance of language in child development can potentially prevent or reduce the effects of stress on the child, foster healthy brain development, and prevent the proliferation of health and social problems later in life. Evidence has shown that targeted interventions for families facing risks to engage with their young children significantly increase interactions between the parent and child, child language production, and the amount of vocalization response of the child toward the parent, as well as enhance the diversity and breadth of the parents’ vocabulary toward the child (Cronan, Cruz, Arriaga, & Sarkin, 1996; Leffel & Suskind, 2013). Parent-focused language interventions show promising evidence of significant positive effects on children’s language skills, though the effects differ across interventions, so it is important 9
to understand parent characteristics that might impact these effects (Roberts & Kaiser,
2011).
Attachment theory. This study uses attachment theory to understand the interactions of the parent and child, and parent engagement with the intervention group.
Bowlby’s (1978) theory of attachment focuses on the need of an infant to securely explore their environment through consistent and responsive interactions with the primary caregiver. These patterns of interaction between the infant and the primary caregiver organize the child’s stress response system which is foundational for the social and emotional development of the child. When interaction between the infant and primary caregiver has been disrupted due to disengagement or lack of proximity, this causes stress.
If engagement is quickly recovered through sensitivity of the caregiver towards the infant, the infant develops trust in the caregiver and can adapt and overcome the stressor (Albers,
Marianne Riksen-Walraven, Sweep, & Weerth, 2008; Tronick, 2006). If engagement is not recovered and the primary caregiver is not readily available and supportive, the infant will be overwhelmed by stress and limited in its capacity to cope, causing maladaptation
(Schore, 2000; Tronick, 2006; van IJzendoorn, Goldberg, Kroonenberg, & Frenkel, 1992).
This maladaptation can result in either hyperactivation or deactivation in the organization of the response system, indicating anxious or avoidant attachment (Schore, 2000). All humans cultivate internal working models of attachment based on their attachment with their primary caregiver, in addition to other significant relationships, which develop throughout the life span (Bowlby, 1978).
Adult attachment style. Research suggests that an individual’s attachment style is influenced by patterns with figures of attachment throughout life that yield relatively stable 10
representations of attachment perceptions in adulthood (Brennan, Clark, & Shaver, 1998;
Hazan & Shaver, 1987). According to the literature on adult attachment style, the two patterns of insecure attachment are avoidance and anxiety (Brennan et al., 1998; Green et al., 2007). Individuals that score high on avoidance measures tend to have difficulty opening up to others and suppress desire for intimacy. Individuals that score high on the anxious dimension tend to have stress and insecurity in relationships with desires for intimacy, but concern about availability and responsiveness. Conversely, people that score low on both dimensions are secure with intimacy, depending on and offering support to others.
Conceptual framework. The conceptual framework for this study is shown in Figure 1. Studies have shown adult attachment style is a predictor of parenting behavior; parents with secure adult attachment style are likely to contribute to parent-child interaction because of consistent and responsive parenting behavior (Haley & Stansbury, 2003; Figure 1. Conceptual framework
Rholes et al., 1997). A study of mother-child dyads of 3-year-olds found that shared book reading occurred less frequently among insecurely attached dyads (Bus, van Ijzendoorn, &
Bus Marinus H van IJzendoorn, 1995). Green et al. (2007) observed that attachment style impacts how low-resourced parents access social support and that families with greater perceived stress were more likely to have insecure attachment and less cohesive relationships. Evaluations of early childhood home visitation programs have investigated the impact of adult attachment style on program outcomes with mixed results (Cluxton-
Keller et al., 2014; Duggan et al., 2009; Korfmacher et al., 2008; McFarlane et al., 2010). 11
Duggan et al. (2009) found that the likelihood for substantiated child maltreatment was higher among mothers that were depressed and scored high on attachment avoidance and low on attachment anxiety. Additionally, they found that non-depressed mothers with secure attachment style did not benefit from the home visitation program; benefits were more apparent in those with low to moderate attachment insecurity. Anxious attachment enhanced home visitation impacts for non-depressed mothers; and for depressed mothers, home visiting benefits were more apparent in those with low to moderate insecure attachment style. Berlin et al. (2011) reported that parents which scored high in attachment avoidance were less supportive of their child and also participated less in the parenting intervention compared to mothers that scored low on attachment avoidance. A systematic review of adult attachment and parenting found that attachment security was related to positive parenting behaviors; however, the authors noted some inconsistencies in findings across studies and called for further research on the interaction of adult attachment style and parenting behaviors (Jones, Cassidy, & Shaver, 2015).
Innovation. Studies on adult attachment style among parent-focused implemented early language programs that intervene in the first few years of life have not been published and so this study will add to the empirical literature increasing understanding of the associations between adult attachment style and program outcomes. This study builds on previous research on the impact of adult attachment style in early childhood home visitation programs. Adult attachment style remains understudied in terms of the relationship between attachment styles and parenting intervention outcomes (Alhusen, Hayat, & Gross,
2013). 12
This study is innovative because most early language development studies focus on characteristics and outcomes related to the child. In this study, the parent is the primary target of intervention and so it is critical to understand the impact of parent characteristics, namely adult attachment style, on the intervention effects on parent-child interaction outcomes in the home language environment. In addition, this program is delivered via a group parent support model, so it is important to understand the impact of adult attachment style on parent participation in the intervention.
The study also utilizes an innovative approach to measure parent-child interaction in the home language environment. This interaction is usually measured through transcription of parent and child utterances or through the use of observational survey tools
(Gilkerson, Richards, & Topping, 2017). This study uses a digital language processor
(DLP) to measure the parent-child interaction outcomes in the home language environment and provide quantitative feedback to the individual parent, which is an innovative way to observe the effects of the program and provide critical information to the parent on their progress in fostering their child’s language development (Gilkerson, Richards, & Topping,
2017; Wang et al., 2017).
Approach. Design. The proposed study is a secondary data analysis of data from an ongoing parent study on a parent-focused early language development intervention. The parent study employs a pre/post quasi-experimental design to evaluate the effectiveness of a parent-focused early language development intervention and to assess the associations of acute and chronic stressors on child early language development and protective factors that potentially mitigate the impact of adversity on child development. The parent study was approved by the institutional review board at Baylor College of Medicine/Texas Children’s 13
Hospital. The current study will seek approval as a secondary analysis through expedited review by the institutional review board at the University of Texas Health Science Center at Houston.
Sample, Setting, and Power. Parents of children between the ages of 0 to 24 months of age were recruited to participate in a 13-week early language program conducted in community settings in Houston, Texas. Participants were recruited through WIC Centers, community libraries, community shelters, the children’s museum, and community pediatric clinics. Inclusion criteria for study participants included parent-child dyads with at least one child (study child) between the ages of 0-24 months, parents were 18 years or older and English or Spanish speaking. Exclusion criteria included study child with known developmental or language delay per parent report, study child enrolled in early head start, home visitation programs, or other early intervention programs, and family primary language other than English or Spanish. Informed consent was obtained for each participating parent. If more than one parent per child participated in the program, the family chose one primary parent to complete the study questionnaires and report shared book reading minutes. Incentives for program participation included a mini-library of 16 age-appropriate books for their child at the conclusion of the program. Data will be deidentified by the research team from the parent study and will be transmitted for analysis via a secure email server or transferred via a password protected encrypted hard drive.
An a priori sample size estimate using G*Power 3.1.9.2 (Faul, Erdfelder, Lang, &
Buchner, 2007) for multiple linear regression was conducted for f2 = 0.15, with α = 0.05, and power (1-) = 0.8; 55 subjects are needed to detect statistical differences. Power estimates for latent class growth curve models using the “Vermunt 3-step method” 14
(Vermunt, 2010) require Monte Carlo simulation studies to determine sufficient sample size to detect statistically significant differences. Curran, et al. (2010) state that growth models have been successfully fitted to small samples, but samples close to 100 are preferable. This is an exploratory study using secondary data and so the sample size is limited to the sample amount in the dataset.
Intervention Description. The parent-focused early language intervention uses a group support education model, in which up to 15 families meet weekly over the course of
13 weeks. The intervention uses the Language Environment Analysis (LENA) Research
Foundation LENA Start™ curriculum to train families on how to engage their infants in language development. In addition to the curriculum, each week families take home a DLP to record up to 16 hours of parent-child interaction in the home language environment one day during the week. The device measures the number of adult words spoken toward or near the child, the number of child vocalizations, the number of conversational turn interactions (time-adjacent adult-child language interaction occurring within 5 seconds of one another), and the amount of time the child is exposed to television or other audio electronics (Zimmerman et al., 2009). At the weekly sessions, parents are provided feedback from the data collected on the DLP so they are able to track changes and work to improve in the quantity of their interactions with their children. The intervention contains behavioral change theoretical methods from Ajzen’s (1991) Theory of Planned Behavior and Bandura’s (1986) Social Cognitive Theory, including knowledge, beliefs, and outcome expectations.
Data collection. Questionnaire data for the parent study were collected and managed electronically using REDCap electronic data capture tools (Harris et al., 2009), a 15
secure web-based application hosted at Baylor College of Medicine. Baseline data were collected prior to the intervention and post-questionnaire data was collected at the conclusion of the program. The DLP recording data was collected twice prior to intervention and once weekly between intervention sessions. The means of the two preintervention recordings were used to create the baseline measures used in this study.
This was performed to reduce the potential of the “Hawthorne effect,” that being observed can alter behavior, of the initial recording (McCambridge, Witton, & Elbourne, 2014).
Deidentified study data will be transferred via a secure email server or a password protected encrypted drive.
Measurements. Adult attachment style was measured at baseline by the Attachment
Style Questionnaire (ASQ) (Feeney, Noller, & Hanrahan, 1994). The ASQ is a 40-item multidimensional, self-report instrument which has respondents rate their agreement with statements about their perceptions of themselves and others using a 6-point Likert type scale ranging from “totally disagree” to “totally agree” (Feeney et al., 1994; Karantzas,
Feeney, & Wilkinson, 2010). In previous studies the ASQ demonstrated acceptable evidence of internal consistency (Cronbach's alpha > 0.80) and stability (test-retest reliability) over a ten-week period (r=0.76) (Feeney et al., 1994). Convergent validity was established with correlations with the Hazan and Shaver paragraphs of adult attachment
(Feeney, Alexander, Noller, & Hohaush, 2003; Hazan & Shaver, 1987). The original study on the ASQ demonstrated construct validity through principal component analysis with a three and five factor solution (Feeney et al., 1994). Other studies have used the ASQ to measure two broad adult attachment dimensions: anxiety and avoidance (Alexander,
Feeney, Hohaus, & Noller, 2001; Feeney, Alexander, Noller, & Hohaus, 2003; McMahon, 16
Barnett, Kowalenko, & Tennant, 2005; Meredith, Strong, & Feeney, 2006). The factor structure of the ASQ varies across populations and so recent studies have used the factor scores from factor analyses to operationalize adult attachment style (Cluxton-Keller et al.,
2014; Duggan et al., 2009; McFarlane et al., 2010, 2013). Factor scores will be obtained from a separate study conducting an exploratory factor analysis among parents of young children in Houston, Texas to understand the factor structure of the ASQ in this population.
The individual factor scores for each participant will be used in the data analysis for this study.
Weekly measurements of parent-child interaction in the home language environment was measured using a secure text messaging system and the LENA DLP and analysis software. This construct includes the average weekly time spent in shared book reading, the number of child-directed speech (adult words spoken to the study child), and the number of serve and return interactions (conversational turns). The parent study includes averaged baseline recordings and 12 weekly recordings that are recorded between each weekly intervention meeting. Participating parents were instructed to collect full-day recordings on a typical day when the parent would be with the study child. On the recording day of choice for the parent, as the study child woke up, the parent pressed record on the DLP and placed it in a snap pocket in a vest. The parent dressed the study child with the vest outside of the child’s clothing with the DLP facing on the front of the study child’s chest. The DLP automatically turned off after recording for approximately 16 hours.
Parents brought the DLP with them when they attended the intervention sessions for the data to be processed. The DLP data was uploaded to the LENA analysis software and analyzed with preprogrammed speech algorithms. In a validation study, the Pearson 17
product-moment correlation was high (r=.92, p<.01) between the monitor and human estimates of distinguishing word counts (Xu, Yapanel, & Gray, 2009; Zimmerman et al.,
2009). Average weekly time spent in shared book reading was collected each week in between intervention meetings through self-report from participating parents through a secure text messaging system.
Intervention participation The sum score of intervention participation was determined by the participants attendance in the group intervention. The intervention includes 13 weekly meetings and so this value ranges from 1 to 13 indicating that the participant attended between 1 to 13 intervention meetings.
Sociodemographic data was collected at baseline. Variables that will be controlled for as confounding factors include parental race/ethnicity, age, educational attainment, marital status, income by poverty threshold (based on the ratio of income to federal poverty thresholds established by the U.S. Census Bureau), and child age.
Postnatal depression symptomology is important to control for as a confounding factor because it is estimated that between 10% and 29% of women and 1.2% - 25.5% of men report depressive symptoms in the first year after the birth of a child (Chabrol &
Callahan, 2007; Chaudron, Szilagyi, Kitzman, Wadkins, & Conwell, 2004; Goodman,
2004). Reported estimates of depression are even higher for low-income parents of young children (Hall, Williams, & Greenberg, 1985; Heneghan, Johnson Silver, Bauman,
Westbrook, & Stein, 1998). This variable was measured at baseline using the Edinburgh
Postnatal Depression Scale (EPDS) (Cox, Holden, & Sagovsky, 1987). This is a 10-item questionnaire measuring depressive symptomology. A cut off score of 10 or more was used 18
to identify depressive symptoms. The scale has acceptable psychometric properties of 86% sensitivity, 78% specificity, and positive predictive value of 73% (Cox et al., 1987).
Analysis. Descriptive statistics and correlation analyses (using either Pearson’s correlation coefficient or Spearman’s rank-order correlation coefficient) will be conducted with all variables included in this secondary analysis.
Aim 1. This study will use the mean baseline recording measure and the first recorded data collected in between each intervention meeting. The analysis will be conducted using the
Vermunt three-step model (Vermunt, 2010). The initial step includes classifying individual growth trajectories of each of the home language environment repeated measures (adult word count, conversational turns, and shared book reading) using latent class growth models in MPlus version 8.1 (Muthén & Muthén, 2017) Latent class growth models assess change in repeated measures of linear and nonlinear data and allow for differences in the amount of repeated measures and the trajectory of the measures among the individual participants (Llabre, Spitzer, Spiegel, Saab, & Schneiderman, 2004). The second step includes retaining the posterior probabilities of growth trajectory membership for the individuals in each model. Instead of classifying each individual to the growth trajectory with the highest probability, posterior probabilities are retained to prevent misclassification
(Vermunt, 2010). The third step of modeling includes the predictor variables: factor scores from the ASQ, sum scores of intervention participation, depressive symptomology scores, and the sociodemographic variables. Multinomial logistic regression models are used to assess the associations of the predictor variables on the growth trajectory of the home language environment measures. Figures 2.1-2.3 depict the models for this analysis. 19
Figure 2.1. Child-directed speech (Adult Word Count) growth model with predictor variables
Figure 2.2. Serve and return interaction (Conversational Turns) growth model with predictor variables
20
Figure 2.3. Shared book reading growth model with predictor variables
Aim 2. Multiple linear regression models will be used, with the sum scores of intervention participation as the dependent variable and the factor scores on the ASQ as the independent variables. Covariates of parental race/ethnicity, age, educational attainment, marital status, income by poverty threshold, child’s age, and postnatal depression scores will be controlled for in the model.
Potential problems and alternative strategies. It is possible that a selection bias may be present among parents that have voluntarily chosen to participate in an early language intervention and participate in the parent research study by potentially overrepresenting parents who have a secure attachment style. Meaning, those that are more avoidant or anxious may not participate in a group intervention. The potential presence of this bias may influence the analysis of parents with insecure attachment styles in this study.
Studies have found that among a nationally representative sample that 59% of the respondents were classified as having a secure attachment style, 25.2% with an avoidant attachment style, 11.3% with an anxious attachment style, and 4.5% were unclassified due to equal scores on the dimensions (Mickelson, Kessler, & Shaver, 1997). 21
Strengths and limitations. This study requires the use of secondary data that has already been collected and so the investigator is limited to the constraints of the data set.
Another potential limitation is that social desirability bias may make some parents overestimate their attachment in relationships; which may be inherent in the self-reported assessment of adult attachment style, though the ASQ has shown evidence of acceptable reliability and validity. Social desirability bias may also be found in the measures of the home language environment, as the parent knows that they are recording their interactions on that specific day. Additionally, the DLP measures all child-directed language interaction in the home environment and does not distinguish between specific adults that are interacting with the child. The measures reported in this study are a reflection of the interaction that occurred toward the study child during the recording time in the home environment and not specific to the participating parent. The strengths of this study, such as the use of an innovative approach to measure the home language environment and the potential to understand how parent characteristics of adult attachment style impact parent- focused early language development programs outweigh any potential limitations. The results of this study have the promise to inform future targeted interventions to foster early childhood development.
Risks and benefits. There are no known risks associated with this secondary data analysis. Data obtained for this study will be deidentified and so there is no risk of loss of confidentiality. The study has the potential to inform future targeted interventions that foster early childhood development and so the risk to benefit ratio is favorable.
22
References
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human
Decision Processes, 50(2), 179–211.
Albers, E. M., Marianne Riksen-Walraven, J., Sweep, F. C. G. J., & Weerth, C. D.
(2008). Maternal behavior predicts infant cortisol recovery from a mild everyday
stressor. Journal of Child Psychology and Psychiatry and Allied Disciplines,
49(1), 97–103.
Alexander, R., Feeney, J. A., Hohaus, L., & Noller, P. (2001). Attachment style and
coping resources as predictors of coping strategies in the transition to parenthood.
Personal Relationships, 8(2), 137–152.
Alhusen, J. L., Hayat, M. J., & Gross, D. (2013). A longitudinal study of maternal
attachment and infant developmental outcomes. Arch Womens Ment Health,
16(6), 521–529.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory.
Englewood Cliffs, N.J: Prentice-Hall.
Berlin, L. J., Whiteside-Mansell, L., Roggman, L. A., Green, B. L., Robinson, J., &
Spieker, S. (2011). Testing maternal depression and attachment style as
moderators of Early Head Start’s effects on parenting. Attachment & Human
Development, 13(1), 49–67.
Bowlby, J. (1978). Attachment theory and its therapeutic implications. Adolescent
Psychiatry, 6, 5–33.
Bowman, B. T., Donovan, S., & Burns, M. (2001). Eager to learn: Educating our 23
preschoolers. (B. T. Bowman, S. Donovan, & M. Burns, Eds.). Washington, DC:
National Academy Press.
Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-Report measures of adult
attachment: An integrative overview. Attachment Theory and Close Relationships,
(January 1998), 46–72.
Bus, A. G., van Ijzendoorn, M. H., & Bus Marinus H van IJzendoorn, A. G. (1995).
Mothers reading to their 3-year-olds: The role of mother-child attachment security
in becoming literate. Reading Research Quarterly, 30(4), 998–1015.
Casey, B. J., Giedd, J. N., & Thomas, K. M. (2000). Structural and functional brain
development and its relation to cognitive development. Biological Psychology,
54(1–3), 241–257.
Chabrol, H., & Callahan, S. (2007). Prevention and treatment of postnatal depression.
Expert Review of Neurotherapeutics, 7(5), 557–576.
Chaudron, L. H., Szilagyi, P. G., Kitzman, H. J., Wadkins, H. I. M., & Conwell, Y.
(2004). Detection of postpartum depressive symptoms by screening at well-child
visits. Pediatrics, 113, 551–558.
Cluxton-Keller, F., Burrell, L., Crowne, S. S., McFarlane, E., Tandon, S. D., Leaf, P. J.,
& Duggan, A. K. (2014). Maternal relationship insecurity and depressive
symptoms as moderators of home visiting impacts on child outcomes. Journal of
Child and Family Studies, 23(8), 1430–1443.
Connell, C. M., & Prinz, R. J. (2002). The impact of childcare and parent-child
interactions on school readiness and social skills development for low-income
African American children. Journal of School Psychology, 40(2), 177–193. 24
Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Edinburgh Postnatal Depression Scale
(EPDS): Instructions for using the Edinburgh Postnatal Depression Scale. British
Journal of Psychiatry, 150(3), 782–786.
Cronan, T. A., Cruz, S. G., Arriaga, R. I., & Sarkin, A. J. (1996). The effects of a
community-based literacy program on young children’s language and conceptual
development. American Journal of Community Psychology (24).
Curran, P. J., Obeidat, K., & Losardo, D. (2010). Twelve frequently asked questions
about growth curve modeling. Journal of Cognitive Development, 11(2), 121–
136.
Duggan, A. K., Berlin, L. J., Cassidy, J., Burrell, L., & Tandon, S. D. (2009). Examining
maternal depression and attachment insecurity as moderators of the impacts of
home visiting for at-risk mothers and infants. J Consult Clin Psychol, 77(4), 788–
799.
Duncan, G. J., & Magnuson, K. (2013). Investing in preschool programs. Journal of
Economic Perspectives, 27(2), 109–132.
Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible
statistical power analysis program for the social, behavioral, and biomedical
sciences. Behavior Research Methods, 39(2), 175–191.
Feeney, J. A., Alexander, R., Noller, P., & Hohaus, L. (2003). Attachment insecurity,
depression, and the transition to parenthood. Personal Relationships, 10, 475–
493.
Feeney, J. A., Alexander, R., Noller, P., & Hohaush, L. (2003). Attachment insecurity, 25
depression, and the transition to parenthood. Personal Relationships, 10(4), 475–
493.
Feeney, J. A., Noller, P., & Hanrahan, M. (1994). Assessing adult attachment:
Developments in conceptualization of security and insecurity. In M. B. Sperling
& W. H. Berrman (Eds.), Attachment in adults: Theory, assessment, and
treatment (pp. 128–152). New York, NY: Guilford Publications.
Gilkerson, J., Richards, J. A., & Topping, K. J. (2017). The impact of book reading in the
early years on parent–child language interaction. Journal of Early Childhood
Literacy, 17(1), 92–110.
Gilkerson, J., Richards, J. A., Warren, S. F., Montgomery, J. K., Greenwood, C. R., Oller,
D. K., … Paul, T. D. (2017). Mapping the early language environment using all-
day recordings and automated analysis. American Journal of Speech-Language
Pathology, 26(2), 248–265.
Goodman, J. H. (2004). Paternal postpartum depression, its relationship to maternal
postpartum depression, and implications for family health. Journal of Advanced
Nursing, 45(1), 26–35.
Green, B. L., Furrer, C., & McAllister, C. (2007). How do relationships support
parenting? Effects of attachment style and social support on parenting behavior in
an at-risk population. American Journal of Community Psychology, 40, 96–108.
Greenwood, C. R., Carta, J. J., Walker, D., Watson-Thompson, J., Gilkerson, J., Larson,
A. L., & Schnitz, A. (2017). Conceptualizing a Public Health Prevention
Intervention for Bridging the 30 Million Word Gap. Clinical Child and Family
Psychology Review, 20(1), 1–22. 26
Haley, D. W., & Stansbury, K. (2003). Infant stress and parent responsiveness:
Regulation of physiology and behavior during still-face and reunion. Child
Development, 74(5), 1534–1546.
Hall, L. A., Williams, C. A., & Greenberg, R. S. (1985). Supports, stressors, and
depressive symptoms in low-income mothers of young children. American
Journal of Public Health, 75, 518–522.
Harris, P. A., Taylor, R., Thielke, R., Payne, J., Gonzalez, N., & Conde, J. G. (2009).
Research electronic data capture (REDCap)-A metadata-driven methodology and
workflow process for providing translational research informatics support.
Journal of Biomedical Informatics, 42(2), 377–381.
Hart, B., & Risley, T. R. (1995). Meaningful Differences in the Everyday Experience of
Young American Children. Baltimore, MD: Paul H. Brookes Publishing Co.
Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process.
Journal of Personality and Social Psychology, 52(3), 511–524.
Heneghan, A. M., Johnson Silver, E., Bauman, L. J., Westbrook, L. E., & Stein, R. E.
(1998). Depressive symptoms in inner-city mothers of young children: Who is at
risk? Pediatrics, 102(6), 1394–1400.
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.
Hoff, E. (2013). Interpreting the early language trajectories of children from low-SES and
language minority homes: Implications for closing achievement gaps.
Developmental Psychology, 49(1), 4–14. 27
Hurt, H., & Betancourt, L. M. (2017). Turning 1 year of age in a low socioeconomic
environment: A portrait of disadvantage. Journal of Developmental & Behavioral
Pediatrics, 38(7), 493–500.
Jones, J. D., Cassidy, J., & Shaver, P. R. (2015). Parents’ self-reported attachment styles:
a review of links with parenting behaviors, emotions, and cognitions. Personality
and Social Psychology Review, 19(1), 44–76.
Jutte, D. P., Miller, J. L., & Erickson, D. J. (2015). Neighborhood adversity, child health,
and the role for community development. Pediatrics, 135(Supplement), S48–S57.
Kalil, A., Ziol-Guest, K. M., Ryan, R. M., & Markowitz, A. J. (2016). Changes in
income-based gaps in parent activities with young children from 1988 to 2012.
AERA Open, 2(3), 1–17.
Karantzas, G. C., Feeney, J. a., & Wilkinson, R. (2010). Is less more? Confirmatory
factor analysis of the Attachment Style Questionnaires. Journal of Social and
Personal Relationships, 27(6), 749–780.
Knudsen, E. I. (2004). Sensitive periods in the development of the brain and behavior.
Journal of Cognitive Neuroscience, 16(8), 1412–1425.
Korfmacher, J., Green, B., Staerkel, F., Peterson, C., Cook, G., Roggman, L., …
Schiffman, R. (2008). Parent involvement in early childhood home visiting. Child
& Youth Care Forum, 37(4), 171–196.
Kuhl, P. K. (2011). Early language learning and literacy: Neuroscience implications for
education. Mind, Brain and Education, 5(3), 128–142.
Leffel, K., & Suskind, D. (2013). Parent-directed approaches to enrich the early language 28
environments of children living in poverty. Seminars in Speech and Language,
34(4), 267–277.
Llabre, M. M., Spitzer, S., Spiegel, S., Saab, P. G., & Schneiderman, N. (2004). Applying
latent growth curve modeling to the investigation of individual differences in
cardiovascular recovery from stress. Psychosomatic Medicine, 66(1), 29–41.
Marmot, M. (2005). Social determinants of health inequalities. Lancet, 365(9464), 1099–
1104.
McCambridge, J., Witton, J., & Elbourne, D. R. (2014). Systematic review of the
Hawthorne effect: New concepts are needed to study research participation
effects. Journal of Clinical Epidemiology, 67(3), 267–277.
McFarlane, E., Burrell, L., Crowne, S., Cluxton-Keller, F., Fuddy, L., Leaf, P. J., &
Duggan, A. (2013). Maternal relationship security as a moderator of home visiting
impacts on maternal psychosocial functioning. Prevention Science, 14(1), 25–39.
McFarlane, E., Burrell, L., Fuddy, L., Tandon, D., Derauf, D. C., Leaf, P., & Duggan, A.
(2010). Association of home visitors’ and mothers’ attachment style with family
engagement. Journal of Community Psychology, 38(5), 541–556.
McMahon, C., Barnett, B., Kowalenko, N., & Tennant, C. (2005). Psychological factors
associated with persistent postnatal depression: Past and current relationships,
defense styles and the mediating role of insecure attachment style. Journal of
Affective Disorders, 84(1), 15–24.
Meredith, P., Strong, J., & Feeney, J. A. (2006). Adult attachment, anxiety, and pain self-
efficacy as predictors of pain intensity and disability. Pain, 123, 146–154.
Mickelson, K. D., Kessler, R. C., & Shaver, P. R. (1997). Adult attachment in a 29
nationally representative sample. Journal of Personality and Social Psychology,
73(5), 1092–1106.
Muthén, L., & Muthén, B. (2017). Mplus User’s Guide (8th ed.). Los Angeles, CA:
Muthen & Muthen.
National Scientific Council on the Developing Child. (2004). Young children develop in
an environment of relationships: Working paper No. 1. Retrieved from
http://www.developingchild.net.
Pati, S., Hashim, K., Brown, B., Fiks, A. G., & Forrest, C. B. (2011). Early identification
of young children at risk for poor academic achievement: Preliminary
development of a parent-support prediction tool. BMC Health Services Research,
11, 197.
Raikes, H., Luze, G., Brooks-Gunn, J., Abigail Raikes, H., Alexander Pan, B., Tamis-
LeMonda, C. S., … Rodriguez, E. T. (2006). Mother-child bookreading in low-
income families: Correlates and outcomes during the first three years of life.
Child Development,77(4), 924–953.
Rholes, W. R., Simpson, J. A., Blakely, B. S., Lanigan, L., & Allen, E. A. (1997). Adult
attachment styles, the desire to have children, and working models of parenthood.
Journal of Personality, 65(2), 357–385.
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.
Rowe, M. L. (2008). Child-directed speech: relation to socioeconomic status, knowledge 30
of child development and child vocabulary skill. Journal of Child Language,
35(1), 185–205.
Rushton, F. E., & Kraft, C. (2014). Building brains, forging futures: the pediatrician’s
role. International Journal of Pediatrics and Adolescent Medicine, 1(1), 3–7.
Schore, A. N. (2000). Parent-infant communication and the neurobiology of emotional
development. Head Start National Research Conference (Washington, DC, June
28-July 1, 2000).
Sénéchal, M., & Lefevre, J.A. (2002). Parental involvement in the development of
children’s reading skill: A five-year longitudinal study. Child Development, 73(2),
445–460.
Shonkoff, J. P., Garner, A. S., The Committee on Psychosocial Aspects of Child and
Family Health, Committee on Early Childhood, Adoption and Dependent Care
Section on Developmental and Behavioral Pediatrics. (2012). The lifelong effects
of early childhood adversity and toxic stress. Pediatrics, 129(1), e232–e246.
Shonkoff, J. P., & Phillips, D. A. (2000). From neurons to neighborhoods: The science of
early childhood development. (J. P. Shonkoff & D. A. Phillips, Eds.), National
Academy of Sciences (Vol. 23). Washington, DC: National Academy Press.
Siddiqi, A., Irwin, L. G., & Hertzman, C. (2007). Total environment assessment model
for early child development: Evidence report. Vancouver, BC. Retrieved from
https://pdfs.semanticscholar.org/722d/f13405b23fb902e524ed8dc7c2b198fc2b89.
Song, L., Spier, E. T., & Tamis-Lemonda, C. S. (2014). Reciprocal influences between 31
maternal language and children’s language and cognitive development in low-
income families. Journal of Child Language, 41(2), 305–326.
Tronick, E. (2006). The inherent stress of normal daily life and social interaction leads to
the development of coping and resilience, and variation in resilience in infants
and young children: Comments on the papers of Suomi and Klebanov & Brooks-
Gunn. Annals of the New York Academy of Sciences, 1094, 83–104. van IJzendoorn, M. H., Goldberg, S., Kroonenberg, P. M., & Frenkel, J. (1992). The
relative effects of maternal and child problems on the quality of attachment: A
meta-analysis of attachment in clinical samples. Child Development, 63(4), 840–
858.
Vermunt, J. K. (2010). Latent class modeling with covariates: Two improved three-step
approaches. Political Analysis, 18(4), 450–469.
Wang, Y., Hartman, M., Abdul Aziz, N. A., Arora, S., Shi, L., & Tunison, E. (2017). A
systematic review of the use of LENA technology. American Annals of the Deaf,
162(3), 296–311.
Xu, D., Yapanel, U., & Gray, S. (2009). Reliability of the LENA TM Language
Environment Analysis System in Young Children’s Natural Home Environment.
Retrieved from http://www.lenafoundation.org/wp-content/uploads/2014/10/LTR-
05-2_Reliability.pdf
Zimmerman, F. J., Gilkerson, J., Richards, J. A., Christakis, D. A., Xu, D., Gray, S., &
Yapanel, U. (2009). Teaching by listening: the importance of adult-child
conversations to language development. Pediatrics, 124(1), 342–9.
Zuckerman, B., & Augustyn, M. (2011). Books and reading: Evidence-based standard of 32
care whose time has come. Academic Pediatrics, 11(1), 11–17.
33
Manuscript #1
Impact of Adult Attachment Style on Outcome Trajectories in a Parent-focused
Early Language Development Intervention 34
Background
Significant disparities in early language development have been noted for children growing up in low-income homes, which can impact their educational and health trajectory throughout the life course (Duncan & Magnuson, 2013; Fernald, Marchman, &
Weisleder, 2013; Hart & Risley, 1995; Hurt & Betancourt, 2017; Huttenlocher, Waterfall,
Vasilyeva, Vevea, & Hedges, 2010; Rowe, 2012, 2018). It is widely recognized that the first few years of life are critical for brain development, which is facilitated through stable and responsive caregiving (Casey, Giedd, & Thomas, 2000; Hurt & Betancourt,
2017; Knudsen, 2004). A nurturing environment can facilitate language, cognitive, and social development through positive and responsive interactions between adults and the child and potentially mediate some of the effects of poverty (Hoff, 2003; Siddiqi, Irwin,
& Hertzman, 2007; Smith, Landry, & Swank, 2006). Conversely, adverse environments and adverse parent-child interactions, including low stimulation and neglect, lead to poor cognitive, behavioral, language, and health outcomes (Shonkoff & Phillips, 2000). If a young child is exposed to prolonged adversity in the absence of an adequate buffer, such as a supportive and responsive caregiver, disruptions are noted in brain development and the child’s stress regulating responses (Jutte, Miller, & Erickson, 2015; Rushton & Kraft,
2014; Shonkoff, Garner, The Committee on Psychosocial Aspects of Child and Family,
Committee on Early Childhood Adoption and Dependent Care, & Section on
Developmental and Behavioral Pediatrics, 2012).
Efforts to foster early language development have been targeted toward parents and are part of the larger effort to reduce toxic stress through strengthening the parent-child relationship through knowledge of child development and intentional interactions 35
(Greenwood et al., 2017). Parental involvement and engagement in their child’s language attainment is recognized as critical for healthy development (Raikes et al., 2006; Sénéchal
& Lefevre, 2002). Kuhl (2011) demonstrated that infant language development is facilitated by social interaction, such as reading and talking. Studies have shown that frequency of reading to a child regularly, as well as having accessible children's books in the household, were associated with positive child outcomes, such as early academic success (Pati, Hashim, Brown, Fiks, & Forrest, 2011; Zuckerman & Augustyn, 2011).
Recent studies have evaluated the associations of socioeconomic status, number of adult words spoken to the child, adult-child conversational interaction, and brain function using functional magnetic resonance imaging (MRI) (Romeo, Leonard, et al., 2018; Romeo,
Segaran, et al., 2018). These studies found that higher conversational interaction between the child and the caregiver was associated with greater activation of the language centers of the brain, independent of socioeconomic status and the number of adult words spoken to the child.
Adult attachment style classifications have been theoretically associated with perceptions, emotions, and behavior in the context of close relationships (Edelstein et al.,
2004). Research has demonstrated associations of adult attachment style as a predictor of parenting behavior (Haley & Stansbury, 2003). Parents with secure attachment style are likely to contribute to parent-child engagement because of consistent and responsive parenting behaviors (Haley & Stansbury, 2003; Rholes, Simpson, Blakely, Lanigan, &
Allen, 1997). A systematic review of adult attachment and parenting found that attachment security was related to positive parenting behaviors; however, the authors noted some inconsistencies in findings across studies and called for further research on 36
the interaction of adult attachment style and parenting behaviors (Jones, Cassidy, &
Shaver, 2015). Parent-focused early language development interventions have promising evidence in promoting child development; however, these interventions have varied effects (Cain, Wood, Santa Maria, Mandell, & Rozmus, n.d.; Roberts & Kaiser, 2011).
Due to the theoretical associations with parental behavior, understanding the impact of adult attachment style in a parent-focused early language intervention may explain some of the varied effects. To date, no study has examined the associations of adult attachment style in parent-focused early language development interventions. Studies investigating the association of adult attachment style on early childhood home visitation programs found that parents have varied engagement and program benefit due to attachment style differences (L. J. Berlin et al., 2011; Duggan, Berlin, Cassidy, Burrell, & Tandon, 2009).
Understanding parental influences in interventions to promote healthy child development are critical to eliminate educational and health disparities throughout the child’s life course (National Scientific Council on the Developing Child, 2004).
Therefore, the specific aims of this study were to investigate: 1) the association of adult attachment style on the growth trajectories of parent-child interaction measured longitudinally in the home environment and 2) the association of adult attachment style on intervention attendance. It was hypothesized that adult attachment style would predict individual trajectories over-time in child-directed speech (number of adult words spoken to the child), reciprocal interaction (number of conversational turns), and shared book reading (average minutes spent reading with child), with scores that indicate a more insecure attachment style being associated with least favorable growth trajectories and lower attendance compared to those with scores on the opposite end of the scale. 37
Theoretical Framework
This study uses attachment theory to understand the interactions of the parent and child, and parent engagement with the intervention group. All humans cultivate internal working models of attachment based on their attachment with their primary caregiver, as well as to other significant relationships, which develop throughout the life span
(Bowlby, 1978). Research suggests that an individual’s attachment style is influenced by patterns with figures of attachment throughout life that yield relatively stable representations of attachment perceptions in adulthood (Brennan, Clark, & Shaver, 1998;
Hazan & Shaver, 1987). According to the literature on adult attachment style, the two patterns of insecure attachment are avoidance and anxiety (Brennan et al., 1998; Green,
Furrer, & McAllister, 2007). Individuals that score high on avoidance measures tend to have difficulty opening up to others and suppress desire for intimacy. Individuals that score high on the anxious dimension tend to have stress and insecurity in relationships with desires for intimacy, but concern about availability and responsiveness. Conversely, as attachment is measured along a continuum, people that score low on both dimensions are secure with intimacy, depending on and offering support to others.
Studies have shown adult attachment style is a predictor of parenting behavior; parents with secure adult attachment style are likely to contribute to parent-child interaction because of consistent and responsive parenting behavior (Haley & Stansbury, 2003; Rholes et al., 1997). A study of mother-child dyads of 3-year-olds found that shared book reading occurred less frequently among insecurely attached dyads (Bus & van Ijzendoorn, 1995;
Bus, van IJzendoorn, & Pellegrini, 1995). Green et al. (2007) observed that adult attachment style impacts how low-resourced parents access social support and that families 38
with greater perceived stress were more likely to have insecure attachment and less cohesive relationships.
Evaluations of early childhood home visitation programs have investigated associations of adult attachment style on program outcomes with mixed results (Cluxton-
Keller et al., 2014; Duggan et al., 2009; Korfmacher et al., 2008; McFarlane et al., 2010).
Duggan et al. (2009) found that non-depressed mothers with secure attachment style did not benefit from the home visitation program; benefits were more apparent in those with low to moderate attachment insecurity. Anxious attachment enhanced home visitation impacts for non-depressed mothers; and for depressed mothers, home visiting benefits were more apparent in those with low to moderate insecure attachment style. Berlin et al. (2011) reported that parents scoring high in attachment avoidance were less supportive of their child and also participated less in the parenting intervention compared to mothers that scored low on attachment avoidance.
Methods
Design
This exploratory study used data from a program evaluation of a parent-focused early language development intervention (parent study). The present study utilizes deidentified program data on longitudinal outcomes measured by digital language processors, parent-report data, and program attendance data and deidentified data on adult attachment style, depression, and sociodemographic characteristics from a sample of program participants that participated in the parent study (Figure 1). The parent study was approved by the institutional review board at Baylor College of Medicine/Texas Children’s
Hospital. The present study was determined to qualify for exempt status for utilizing 39
existing deidentified data by the institutional review board at the University of Texas
Health Science Center at Houston.
Sample
Parents/guardians (hereafter referred to as parent) of children between the ages of
0 to 24 months of age were recruited to participate in a universally delivered 13-week early language development program conducted in community settings. Participants were recruited through WIC Centers, community libraries, the children’s museum, and community pediatric clinics. Inclusion criteria for study participants included parent-child dyads with at least one child (study child) between the ages of 0-24 months, parents were
18 years or older and English or Spanish speaking. Exclusion criteria included study child with known developmental or language delay per parent report, and study child enrolled in early head start, home visitation programs, or other early intervention programs. Program and study informed consent was obtained for each participating parent. Figure 1 depicts the sample for this study and the attachment style questionnaire factor analysis study (Cain, in preparation).
Power estimation conducted in G*Power 3.1.9.2. (Faul, Erdfelder, Lang, &
Buchner, 2007) for a logistic regression analysis with an α = .05 demonstrated that with a moderate effect (R2 = .15; OR 1.7), detection of statistical differences would be achieved with 80% power from 153 participants. After the program data used to determine latent growth trajectories was matched with survey data, the study had 87 participants for the logistic regression analyses and so the results should be considered exploratory and may underestimate the true relationships in the population.
40
Intervention Description
The parent-focused early language development intervention used a group support education model, in which up to 15 families met weekly over the course of 13 weeks. The intervention used the Language Environment Analysis (LENA™) Research Foundation
LENA Start™ curriculum to train families on how to engage their infants in language development. A complete list of topics covered in each session can be provided by request to the first author. In addition to the curriculum, each week families took home a digital language processor to record up to 16 hours of parent-child interaction in the home environment one day during the week (LENA Research Foundation, 2018). The device is worn by the child throughout the day and measured the number of adult words spoken to or near the child, the number of child vocalizations, the number of conversational interactions (time-adjacent adult-child language interaction occurring within 5 seconds of one another), and the amount of time the child is exposed to television or other audio electronics (Zimmerman et al., 2009). During the weekly sessions, parents were provided feedback from the data collected with the digital language processors so they could track their individual changes and work to improve the quantity of their interactions with their children. The intervention contains behavioral change theoretical constructs from
Bandura’s (1986) Social Cognitive Theory, including knowledge, beliefs, and outcome expectations. Incentives for program participation included a mini-library of up to 16 age- appropriate books for their child. One book was given to participating caregivers at recruitment and at each intervention session. Additionally, after enrollment, study participants were provided with a resource guide for obtaining additional information and referrals, if needed, for depression or familial violence. 41
Data collection
Questionnaire data for the parent study were collected and managed electronically using REDCap electronic data capture tools (Harris et al., 2009), a secure web-based application hosted at Baylor College of Medicine. Baseline data were collected from a sample of program participants prior to the intervention and post-questionnaire data were collected at the conclusion of the program. For the present study, only baseline questionnaire data were utilized. The digital language processor recording data were collected at baseline and once weekly between intervention sessions. Fifty-eight percent of the participants with baseline recording data had two valid baseline recordings. For these analyses, these were averaged to create a single baseline. The practice of collecting two baseline recordings was performed to reduce the potential of the “Hawthorne effect,” that being observed can alter behavior, of the initial recording (McCambridge, Witton, &
Elbourne, 2014).
Measures
Adult attachment style. Adult attachment style was measured at baseline by the
Attachment Style Questionnaire (ASQ) (Feeney, Noller, & Hanrahan, 1994). The original
ASQ was a 40-item multidimensional, self-report instrument, which has respondents rate their agreement with statements about their perceptions of themselves and others using a
6-point Likert type scale ranging from “totally disagree” to “totally agree” (Feeney et al.,
1994; Karantzas, Feeney, & Wilkinson, 2010). In previous studies the ASQ demonstrated acceptable evidence of internal consistency (Cronbach's alpha > 0.80) and stability over a ten-week period (r=0.76) (Feeney et al., 1994). Convergent validity was established with correlations with the Hazan and Shaver paragraphs of adult attachment (Feeney, 42
Alexander, Noller, & Hohaus, 2003; Hazan & Shaver, 1987). The original study on the
ASQ, conducted among Anglo-Australian university students, demonstrated construct validity through principal component analysis with three- and five-factor solutions (Feeney et al., 1994). Additional studies used the ASQ to measure the two broad adult attachment style dimensions: attachment anxiety and avoidance (Alexander, Feeney, Hohaus, &
Noller, 2001; Feeney et al., 2003; McMahon, Barnett, Kowalenko, & Tennant, 2005;
Meredith, Strong, & Feeney, 2006).
Karantzas, et al (2010) validated the ASQ short form (ASQ-SF), which reduced the original ASQ from 40-items to 29-items. The population that the ASQ-SF was validated on included Australian university students and community members. The authors found five-factors which were nested in the broad attachment dimensions (2 factors) of attachment avoidance and anxiety. The five-factor solution /included 1) confidence (“I am confident about relating to others” and “I feel confident that other people will be there for me when I need them”), 2) discomfort with closeness (“I find it hard to trust other people”), 3) need for approval (“I wonder why people would want to be involved with me”), 4) preoccupation with relationships (“I worry a lot about my relationships”), and 5) relationships as secondary (“achieving things is more important than building relationships” and “if you’ve got a job to do, you should do it no matter who gets hurt”). The broad dimensions and five factors demonstrated acceptable evidence of internal consistency (avoidance: α = .84, anxiety: α = .85, confidence: α = .82, discomfort with closeness: α = .85, need for approval: α = .75, preoccupation with relationships: α = .72, and relationships as secondary: α = .84). Karantzas, et al (2010) 43
state that compared to the ASQ, the ASQ-SF is optimal for measuring adult attachment style.
The ASQ and ASQ-SF have previously been developed and validated on populations representing relatively racially homogenous samples (Duggan et al., 2009;
Feeney et al., 1994; Karantzas et al., 2010; McFarlane et al., 2010). These populations are racially and culturally different from the current study population and so a validation study on the factor structure of the 29-item ASQ-SF was conducted. The sample for the factor analysis study included participants with survey data from this study, data from an early head start program evaluation, and data from parents of young children (ages 0-3 years of age) waiting in a children’s hospital emergency department (Figure 1). The purpose of this factor analysis was to investigate the factor structure among a representative sample of parents of young children in Houston, Texas and to determine factor scores for the study participants (Cain, in preparation). Confirmatory factor analyses, conducted using MPlus version 8.1 (Muthén & Muthén, 2017), revealed that the
ASQ-SF subscales from the Karantzas et al. study (2010) were not a good fit for this ethnically diverse sample (2 Factor Model: RMSEA = .102, CFI = .39; 5 Factor Model:
RMSEA = .11, CFI = .33). Maximum likelihood exploratory factor analysis using both
Varimax (orthogonal) and Promax (oblique) rotations was conducted using SPSS
Statistical Package (version 25) for Windows (IBM Corporation, 2017). Factors with eigenvalues greater than 1.0 and at least three items with factor loadings greater than .4 were considered for interpretation of the factor solutions (Costello & Osborne, 2005;
Ferguson & Cox, 1993). The factor structures for both Varimax and Promax rotations included a three-factor solution. The promax solution was chosen for the final factor 44
structure because it was expected that the factors would be correlated. Factor loadings can be found in Table 1, including comparison of the current study’s factor loadings with the broad adult attachment style constructs from Karantzas et al (2010), Duggan et al.
(2009), and McFarlane et al. (2010).
The interpretation of the factor structure for this study was consistent with previous research on adult attachment style (Duggan et al., 2009; Feeney et al., 1994; Karantzas et al., 2010; McFarlane et al., 2010). Six items related to “relationship anxiety” loaded on factor 1 (i.e. “I worry that others won’t care about me as much as I care about them” and
“I worry that I won’t measure up to other people”). Compared to the original five-factor model, these items relate to both the “need for approval” and “preoccupation with relationships” factors which represent attachment anxiety. Factor 2 represents the
“discomfort with closeness.” Four items loaded on factor 2 (i.e. “I find it hard to trust people” and “I find it difficult to depend on others”). The third factor represents
“relationships as secondary to achievements.” Three items loaded on factor 3 (i.e.
“achieving things is more important than building relationships” and “if you’ve got a job to do, you should do it no matter who gets hurt”). Adult attachment style is measured along a continuum of the same scale and so for each factor, those at the lower end of the continuum have more secure adult attachment styles and those at the higher end of the continuum have more insecure adult attachment styles.
Bivariate correlations revealed that the factors are all significantly correlated, with the strongest correlation between factor 1 and factor 3 (r =.52, p<.001). Reliability analyses revealed adequate internal consistency of all the factors (Factor 1 α = .81; Factor 2 α = .75;
Factor 3 α = .73). Least squares regression approach was used to predict factor scores which 45
were retained for use in the current study (Thurstone, 1934). The Cronbach’s α for the entire ASQ-SF in this current study was .82 (Factor 1: α = .73; Factor 2: α = .74; Factor 3:
α = .65).
Weekly measurements of parent-child interaction in the home environment.
These data were measured using a secure text messaging system and the LENA™ digital language processor analysis software. These data include repeated measures of the number of child-directed speech (adult words spoken to/near the child), the number of reciprocal interactions (conversational turns), and the average weekly time spent in shared book reading.
Child-directed speech and reciprocal interactions. Data included baseline recordings and 12 weekly recordings that were recorded between each weekly intervention meeting. Parents were instructed to collect full-day recordings on a typical day when the parent would be with their child. On the recording day of choice for the parent, the parent pressed power and record on the digital language processor when the child woke up in the morning and placed it in a snap pocket of a special a vest the child wore throughout the day. The digital language processor automatically shuts off after recording for approximately 16 hours. Parents brought the digital language processor with them when they attended the intervention sessions for the data to be processed. The encrypted audio data were transferred from the digital language processor and uploaded to a secure computer via USB port. After uploading, the audio data were automatically erased from the digital language processor (Oller, 2010). The computer contains LENA™ digital language processor software which segments the acoustical features, such as frequencies and proximity, of the audio data into subcategories (adult male, adult female, key child, 46
other child, overlapping speech, electronic media, noise, and silence) through a pattern recognition and speech processing algorithm which was developed using a large sample of professionally transcribed natural language recordings (Gilkerson, Coulter, & Richards,
2008; Xu, Yapanel, & Gray, 2009; Xu et al., 2008). The adult subcategories are segmented to achieve syllabification of speech to estimate adult word counts. In a validation study, the Pearson product-moment correlation was high (r=.92, p<.01) between the digital language processor analysis and human estimates of distinguishing adult word counts (Xu et al., 2009; Zimmerman et al., 2009). Child subcategories are analyzed to determine the number of child vocalizations and then conversational turns are counted when adult and child subcategories occur within five seconds of one another and without other intervening speech activity (Gilkerson et al., 2008; Xu et al., 2009). The program recording data obtained for the current study included at least three valid recordings for each participant in order to determine a growth trajectory (n=224). Due to non-normality of the distributions, these data were log-transformed. A large number of parameter estimates can result in difficulty in fitting a latent class growth model (Wickrama, Lee, Walker O’Neal,
& Lorenz, 2016). Therefore, the log-transformed variables were reduced from 13 time- points to 5 time-points (baseline; mean of time 2, 3, and 4; mean of time 5, 6, and 7; mean of time 8, 9, and 10; and mean of time 11, 12, and 13). This reduction is consistent with recommendations by the LENA research foundation to have a baseline, midpoint of the mean of recordings between times 2-9 (the current study breaks these in to three midpoint assessments), and the final assessment of the mean of the final three recordings (LENA
Research Foundation, 2017). The means of the log-transformed 5 time-point data can be found in Table 2. 47
Shared book reading. Average weekly time spent in shared book reading was collected each week through self-report from participating parents through a secure text messaging system. Due to non-normality of the distributions, these data were log- transformed and reduced from 13 time-points to 5 time-points, shown in Table 2.
Intervention attendance. The sum score of intervention attendance was determined by the participants’ attendance in the program (range 1 to 13). The sum score of intervention attendance had a non-normal distribution, even after transformation (log and fractional rank methods). The variable was transformed to a binary variable for the analysis of the second aim based on the median value of 11 classes (attended < 11 classes
= 0; attended > 11 classes = 1).
Covariates. Sociodemographic information. These data were collected at baseline. Variables used in the analyses included: parental race/ethnicity, age, educational attainment, marital status, income by poverty threshold (based on the ratio of income to federal poverty thresholds established by the U.S. Census Bureau), and child age and gender. Dummy variables were created for the categorical variables in the models (White, non-Hispanic = 1, all other races = 0; Black, non-Hispanic = 1, all other races = 0; Hispanic
= 1, non-Hispanic = 0; Other race not previously classified = 1, White/Black/Hispanic = 0; married/living with a partner = 1, single/separated/divorced/widowed = 0; income above the federal poverty level = 1, income below the federal poverty level = 0; child gender female = 1, child gender male = 0).
Postnatal depression. This variable was measured at baseline using the Edinburgh
Postnatal Depression Scale (EPDS) (Cox, Holden, & Sagovsky, 1987). This is a 10-item questionnaire measuring depressive symptomology. The scale is widely used and has 48
acceptable psychometric properties of 86% sensitivity, 78% specificity, and positive predictive value of 73% (Cox et al., 1987). The Cronbach’s α for the EPDS in this current study was .81.
Analysis
Statistical analyses were performed with IBM SPSS Statistical Package (version
25) for Windows (IBM Corporation, 2017) and MPlus version 8.1 (Muthén & Muthén,
2017). Descriptive statistics were calculated to evaluate data for completeness and assess the distributions of the data. Missing parent-child interaction data was handled using
Markov Chain Monte Carlo multiple imputation in MPlus, where parameter estimates were averaged over 20 sets of analyses and standard errors were computed using the average of the squared standard errors over 20 sets of analyses and the between analysis parameter estimate variation (Muthén & Muthén, 2017; Rubin, 1987; Schafer, 1997). Correlation analyses of the adult attachment style factor scores, depression symptoms, parent age, educational attainment, income level, child age, and intervention attendance were analyzed using Spearman’s rho correlation coefficient.
Latent class growth analyses with predictors. To examine the association of adult attachment style on the growth trajectories among the individual participants on parent-child interaction repeated measures, latent class growth analysis (LCGA) with predictors and covariates was used following the procedures outlined by Jung and
Wickrama (2008), Asparouhou and Muthén (2014), and Vermunt (2010). The advantage of using LCGA to assess change in repeated measures are that these models 1) can classify whether a trajectory is linear or nonlinear, 2) determine the presence of a mixture of latent growth trajectories (assessment of individual differences in change), and 3) allow 49
investigation of predictors (such as adult attachment style) on the growth trajectories
(Harrison & Wang, 2018; Preacher, Winchman, MacCallum, & Briggs, 2008). Individual differences are noted in maturational measures and so LCGA is a useful method to classify trajectories of change in individual behavior, development, and learning (Harrison &
Wang, 2018). Latent class growth models were performed on the LENA digital language processor data (adult word count and conversational turns) (n = 224) and self-reported shared reading minutes (n = 223) in order to identify latent trajectories.
In these models (the same process for adult word count, conversational turns, and reading minutes), Time 1–Time 5 (T1-T5) were entered as observable variables with intercept (i) and slope (s) modelled as latent variables describing the trajectory of growth
(c) over time (Figure 2). The linearity of the single-class growth curve models without covariates was assessed using fixed basis linear models (linear and logarithmic), polynomial models (quadratic and cubic), and latent basis models where the model is freely estimated. Linearity was determined by evaluating the significance of the variance of the slope and model fit was assessed by RMSEA (root mean squared error of approximation),
CFI (comparative fit index), TLI (Tucker-Lewis index), and SRMR (standardized root mean squared residual).
Individuals were classified according to their growth trajectory for each parent- child interaction outcome using LCGA. Covariance estimates within classes were constrained to zero in the latent class growth model (Jung & Wickrama, 2008). The classification of growth trajectories was determined using maximum likelihood estimation, model fit indices, and selection criteria: Akaike Information Criterion (AIC), Bayesian
Information Criterion (BIC), Entropy, and Integrated Classification Likelihood with 50
Bayesian-type Approximation (ICL-BIC). Likelihood ratio tests (LRC) (Vuong-Lo-
Mendell-Rubin likelihood ratio test (VLMR) and bootstrap likelihood ratio tests via 100 draws) were used to assess model fit of the k class model with the k-1 class model (Nylund,
Asparouhov, & Muthén, 2007). Additional considerations were the sample size of the smallest class and theoretical coherence (K.S. Berlin, Williams, & Parra, 2014; Lubke &
Neale, 2006; Ram & Grimm, 2009).
The growth trajectory data were matched to the study survey data (n = 108); 87 of the 108 had classified growth trajectories from the LCGA (based upon valid recording and shared book reading data). To examine the associations of adult attachment style on the latent growth trajectories, the predictor (individual ASQ factors) was regressed individually (bivariate unadjusted) and with covariate variables (sum scores of intervention attendance, depressive symptomology scores, parent age, race/ethnicity, marital status, income level, child gender, and child age) onto the latent growth trajectories. This analysis was conducted using the three-step approach developed by Vermunt (2010) and described by Asparouhou and Muthén (2014). For the first step, growth models were estimated without predictors or covariates (as previously described). During this step, the classification probabilities of being in k class were saved. In addition, the “growth trajectory” signaling the classes with the steepest slopes were identified. The second step involved specifying the classification probabilities of each class using the logits for the classification probabilities that were saved from step one. For the third step, logistic regression was used by regressing the predictor and covariate variables onto the latent trajectories, using the logit values to account for potential classification error. For each class-based regression, bivariate unadjusted (individual ASQ factors) and adjusted models 51
were run. Regression logits and odds ratios were evaluated to determine the associations of being classified into the growth trajectory with the steepest slope. Some covariates were removed from the models due to singularities or estimation producing negative covariances. Regression coefficients were converted to odds ratios for interpretation.
Predictors of intervention attendance. Backward logistic regression was conducted to identify the most parsimonious combination of adult attachment style factors, depression scores, parent age, race/ethnicity, marital status, income level, child gender, and child age in predicting program attendance to greater than 10 classes. The predictor (individual ASQ factors) were regressed individually (unadjusted) and with the covariate variables, intervention attendance, depressive symptomology scores, and the sociodemographic variables (adjusted). Regression coefficients and odds ratios were evaluated to determine the association of attachment style on intervention attendance.
Results
Descriptive statistics on demographic variables of the program (n=224) and survey samples (n=108) are reported in Table 3. Limited demographic information was collected for the program sample. No significant differences were noted in the child characteristics among the program sample and survey sample. The survey sample population was primarily female with an average age of 32. The survey sample was predominantly composed of a racial and ethnic minority population with over 80% of the population classified as non-White (54% Hispanic).
Missing data were evident at the item level in all variables in the sociodemographic, depression, and adult attachment style variables. The nonmissing percentages for these variables are included in Table 3. Missing data were evident at the 52
item level at all time points for all the parent-child interaction variables. The mechanisms for missing recording data were errors in recordings resulting in unreliable data and program attrition. Table 2 provides information on the distributions of the 5 time-point parent-child interaction variables before and after transformation and the percent nonmissing from each variable at each time point. Markov Chain Monte Carlo multiple imputation was performed for all parent-child interaction variables. Table 4 provides estimated sample statistics of these variables after multiple imputation.
Results from the correlation analyses of the adult attachment style factors and the continuous covariate variables can be found in Table 5. Educational attainment was excluded from all the analyses due to the significant correlation > .5 with income level.
Latent Class Growth Modeling
Growth curve model for optimal form. The optimal form of the models were determined using slope variances and model fit indices for the single-class growth curve models. For adult word count over time, the fixed basis logarithmic model was the best fitting model (X2(df) = 71.05(10), p<.001; RMSEA = .16; CFI = .91; TLI = .91; SRMR =
.31). The fixed basis logarithmic model was the best fitting model for conversational turns over time (X2(df) = 65.77(10), p<.001; RMSEA = .16; CFI = .95; TLI = .95;
SRMR = .18). The logarithmic curvature of these models indicated acceleration and deceleration over time. The latent basis linear model was the best fitting model for reading minute growth over time (X2(df) = 45.11(7), p<.001; RMSEA = .16; CFI =.96;
TLI = .94; SRMR = .13). For all of these single-class growth curve models, the poor model fit and significant slope variances provided evidence of heterogeneity of growth. 53
Latent class growth analyses. Child-directed speech (adult word count). Class membership, fit indices, and parameter estimates for the adult word count LCGA without covariates are provided in Table 6. According to fit indices that indicate parsimony (AIC and BIC), the 4-class model had the lowest values. However, entropy, indicating clustering accuracy, favored the 3-class model. The ICL-BIC that indicate both parsimony and clustering accuracy also favored the 3-class model. The likelihood ratio tests noted significant improvement in the 3-class model compared to the 2-class model.
In the 3-class model (Figure 4), the individuals that started with the lowest baseline
(intercept = 8.62, p<.001) had the highest growth trajectory (slope = .30, p<.001). While individuals in this group had the highest growth, the mean estimate of the last observation
(high time 5: 9.10) remained below the other individuals in the other 2 trajectories (low time 5: 10.02; mid time 5: 9.52).
Reciprocal interactions (conversational turns). Class membership, fit indices, and parameter estimates for the adult word count LCGA without covariates are provided in Table 7. The 3-class model was favored over the other models according to fit indices that indicate parsimony and clustering accuracy. The 2-class model was chosen as the optimal model due to class proportions and likelihood ratio tests (VLMR, p = .065). In the 2-class model (Figure 5), individuals belonging to the highest growth trajectory class demonstrated the highest baseline (intercept = 6.11, p < .001) and steepest slope (.204, p
< .001).
Shared book reading (reading minutes). Class membership, fit indices, and parameter estimates for the adult word count LCGA without covariates are provided in
Table 8. According to fit indices that indicate parsimony and quality of the clustering 54
favored the 3-class model. The 2-class model was chosen as the more optimal model due to the distribution of the class proportions. The individuals belonging to the highest growth trajectory class in the 2-class model demonstrated a lower baseline (intercept =
2.171, p < .001) but had the steepest slope (.631, p < .001).
Class based regression analyses with predictors. Child-directed speech (adult word count). Shown in Table 9 are the coefficients and odds ratios from the multinomial logistic regression of the likelihood of adult word count latent class trajectory categorization in the highest and mid trajectory classes, compared with the lowest trajectory. Adult attachment style was not significantly associated with the number of adult words spoken to a child in all the models (unadjusted and adjusted with covariates).
Reciprocal interactions (conversational turns). The coefficients and odds ratios from the binary logistic regression of the likelihood of conversational turn latent class trajectory categorization in the lowest class, compared with the highest trajectory are shown in Table 10. Discomfort with closeness made a significant contribution to the unadjusted and adjusted models, recording an odds ratio of .531 in the unadjusted model and .336 in the model with covariates. This indicates that the odds of being in the highest growth trajectory decrease as discomfort with closeness increases (a decrease of 66% with each unit increase of discomfort with closeness when controlling for covariates).
One other covariate (child age) made significant contributions in predicting membership in the highest growth trajectory. The odds of being in the higher growth trajectory class for reciprocal interactions are increasingly greater as the child ages (by months).
Shared book reading (reading minutes). The coefficients and odds ratios from the binary logistic regression of the likelihood of shared book reading latent class 55
trajectory categorization in the lowest class, compared with the highest trajectory are shown in Table 11. Adult attachment style was not significantly associated with the estimated number of minutes spent reading to the child in any of the models (unadjusted and adjusted). Child age was the only significant covariate in the model with the odds of being in the highest growth trajectory decreasing as the child gets older.
Logistic Regression Predicting Intervention Attendance
Backward step-wise logistic regression was conducted to identify the most parsimonious combination of adult attachment style, depression, and demographic variables in predicting program attendance to greater than 10 classes. Regression results indicated that the overall model including all the predictors was the optimum model in comparison to more parsimonious models in determining associations with intervention attendance (-2 Log Likelihood = 88.450, X2(14) = 31.903, p = .004). Regression coefficients and odds ratios are presented in Table 12. Adult attachment style was significantly associated with intervention attendance. The odds of attending more than 10 classes increase by more than 2x as parents’ discomfort with closeness increases.
Additional significant covariates were income between 200-300% FPL, child gender, and child age.
Discussion
This exploratory study examined the associations of adult attachment style in a parent-focused early language development intervention. The LCGA revealed that there are latent trajectories in the observable parent-child interaction variables. Detecting variation from the mean population on these outcomes by identifying the latent trajectories enhanced understanding of the associations of adult attachment style on the 56
longitudinal parent-child interaction variables. The findings from the regression analyses indicate that adult attachment style was associated with reciprocal interaction and intervention attendance. It was expected that adult attachment style would also be associated with child-directed speech and shared book reading; however, these associations were not noted. This study demonstrates that the odds of intervention attendance increased more than twofold as discomfort with closeness increased; however, for the reciprocal interaction outcome, parents with higher discomfort with closeness had higher odds of having a lower baseline and lower growth trajectory on program measures throughout the program.
Impact on Reciprocal Interaction
Efforts to improve early language development are a part of the larger movement to foster healthy brain development in infants (Golinkoff, Hoff, Rowe, Tamis-Lemonda,
& Hirsh-Pasek, 2018; Greenwood et al., 2017; Hindman, Wasik, & Snell, 2016). Our findings suggest that parents with higher discomfort with closeness are less verbally responsive to their child in the home environment, which is consistent with the literature on adult attachment style. Berlin et al. (2011) found that attachment avoidance predicted lower maternal supportiveness for their children. Mills-Koonce et al. (2011) found that avoidant adult attachment style was associated with less sensitivity in parenting when the parent had increased stress. The finding that associates an insecure adult attachment style
(discomfort with closeness) with a lower baseline and lower growth trajectory for reciprocal interaction is concerning, as studies have shown that increases in this reciprocal interaction are significantly correlated with increases in child receptive and expressive language development (Gilkerson & Topping, 2017; Zimmerman et al., 2009). 57
Impact on Intervention Attendance
The findings on the association of adult attachment style on intervention attendance are divergent with the literature and study hypotheses. Our study found that the odds of attending more than the 10 classes increased more than two times for individuals with increasing discomfort with closeness. The discomfort with closeness factor represented an individuals’ perception of depending on others and trusting others
(i.e., “I find it difficult to depend on others” and “I find it hard to trust people”).
McFarlane et al. (2010) did not find significant associations with program engagement and attachment avoidance in the home visitation context and L.J. Berlin et al, (2011) found that mothers with attachment avoidance were associated with receiving less services than intended in the Early Head Start program. It should be noted that the duration of the early language development program is significantly less than home visitation and Early Head Start programs.
Feedback was a key component of this early language development intervention.
Participants received weekly reports of their recording data which include the number of adult words spoken to the child, number of conversational turns, and the amount of time the child is exposed to television or electronics. The participants with discomfort with closeness were more likely to be in the lower baseline and lower trajectory for reciprocal interaction and so they may have continued to participate because they had not achieved their goals according to the feedback. Furthermore, this finding suggests that attendance in the intervention may not equate to benefit for individuals with insecure adult attachment style. This association in our study lends itself to further research to understand which components of this early language intervention increased the 58
attendance of participants that have attachment avoidance and how to help them with behavioral changes to improve their interaction with their child.
In addition, our study did not find significant associations of relationship anxiety with intervention attendance. McFarlane et al. (2010) found that attachment anxiety was associated with a high dose of engagement with home visitation programs.
Secondary Outcomes
Though not a main variable of interest in this study, it was surprising that income, when controlled with the other variables, was not found to have significant effects on the parent-child interaction models. In addition, unlike previous research on home visitation programs, depression, which was moderately correlated with relationship anxiety, did not have significant effects in any of the analyses (Cluxton-Keller et al., 2014; Duggan et al.,
2009).
Of note, the latent class growth analysis for child-directed speech (adult word count) showed that individuals that started with the lowest baseline had the highest growth trajectory. This indicates that these individuals had higher increases in the amount of words spoken towards the child throughout their time in the intervention, compared to other participants that were classified in the other two growth trajectories. Once learning about the importance of speaking to their child, the individuals with the lower baseline may have been more motivated to increase the amount of speech toward their child during the program or because they started with the lower baseline, they had more room to grow in this behavior toward their child compared to those with a higher baseline.
Child age was a significant predictor with modest odds ratios for reciprocal interaction, shared book reading, and program attendance. The associations with the 59
increases in odds of being in the highest trajectory for reciprocal interaction and higher attendance were expected as the child’s age increased. However, this study also found that as the child increased in age, the odds of being in the highest growth trajectory for shared book reading decreased by 12.5%. This finding may be due to parents of older children already participating in shared book reading prior to the intervention. Parents of younger children may have not been spending as much time in shared book reading and so their increases in this behavior were greater than those with older children.
Another significant covariate in the intervention attendance model was child gender. Parents with male children were over 4 times more likely to attend more than 10 classes, compared to parents with female children. Studies have found gender differences in early language acquisition favoring girls (Bauer, Goldfield, & Reznick, 2002; Fenson et al., 1994; Roulstone, Loader, Northstone, & Beveridge, 2002). It is not clear why parents with boys were more likely to attend 10 or more sessions, but gender differences in language acquisition potentially influenced intervention attendance.
Limitations
This study has several limitations that should be considered when interpreting the findings. The first of these is the relatively small sample size. Analyses were limited by the existing program data. Power to detect moderate effects was low and so this study should be considered exploratory. A larger dataset would have possibly allowed for more robust statistical methods to be utilized. With a larger sample size, latent growth mixture modeling with differences in class variances may provide more insight in understanding the predictors of these differences (Preacher et al., 2008). Secondly, all the measures in this study had missing data. The missingness of the baseline recording measure was 60
relatively high. Multiple imputation was utilized for the recording and reported reading data, this method reduced potential bias introduced into the results due to the imputation process (Rubin, 1987). The measures for shared book reading, adult attachment style, and depression were all self-report measures. While it is assumed that participants reported these measures in a forthright and honest manner, given social desirability, these measures may not completely accurately reflect these constructs. Social desirability may also have influenced the recordings (especially the initial recordings) which may, in some part, explain the logarithmic single-class growth curve for adult word count and conversational turns. This curve indicates acceleration and then deceleration and so there is potential that over time in the program, as the parents get more accustomed to using the recorders, that social desirability bias decreases. This curve also might indicate that participants reach a certain ceiling and are not able to show improvement after reaching this point. This phenomenon needs further exploration. While this study adds to the literature on racial and ethnic minority populations, the results may not be generalizable to the general population. In addition, the participants in this study were primarily female
(92%). Future studies should seek to increase the attendance and inclusion of fathers.
Future Direction and Research
Despite these limitations, this study adds to the literature concerning the influence of adult attachment style in a group-based parent-focused early language development intervention. The LCGA and subsequent regression analyses indicate that a one size fits all approach to early language development may not be the most efficacious. Within the population are subpopulations of individuals that have different characteristics, resources, strengths, and weaknesses. This study suggests that additional supportive measures to 61
foster parent-child reciprocal interaction and various means of engagement may be needed for parents with insecure adult attachment style. This research would be enhanced by increased understanding of the nature of the attachment insecurity. A social ecological approach to investigate not only the intrapersonal and interpersonal characteristics of the parent and child, but additional community and societal factors may enhance understanding how to intervene to foster critical reciprocal interaction between the parent and child (Brofenbrenner, 1979).
Conclusions
Language has a significant role in a child’s development by providing a means of communication and methods for obtaining knowledge (Song, Spier, & Tamis-Lemonda,
2014). Interventions supporting early language development are often targeted towards parents, as healthy development is fostered by parents’ time and attention toward their child through a supportive and cognitively stimulating environment (Kalil, Ryan, & Chor,
2014). This study explored associations of adult attachment style on outcomes collected in the home environment through a parent-focused early language development intervention. The results suggest that insecure adult attachment style is associated with intervention attendance and a lower baseline and lower trajectory of critical reciprocal interaction between the parent and the child. These study findings highlight the importance of distinguishing subgroups that may need different or enhanced models to realize program benefit. More research is needed to understand the nature of insecure adult attachment styles and potential contributing factors that impact a parents’ engagement with their child, which is critical for child development.
62
References
Alexander, R., Feeney, J. A., Hohaus, L., & Noller, P. (2001). Attachment style and
coping resources as predictors of coping strategies in the transition to parenthood.
Personal Relationships, 8(2), 137–152.
Asparouhov, T., & Muthén, B. (2014). Auxiliary variables in mixture modeling: Three-
step approaches using MPlus. Structural Equation Modeling.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory.
Englewood Cliffs, N.J: Prentice-Hall, 1986. xiii, 617.
Bauer, D. J., Goldfield, B. A., & Reznick, J. S. (2002). Alternative approaches to
analyzing individual differences in the rate of early vocabulary development.
Applied Psycholinguistics, 23(03), 313–335.
Berlin, K. S., Williams, N. A., & Parra, G. R. (2014). An introduction to latent variable
mixture modeling (Part 1): Overview and cross-sectional latent class and latent
profile analyses. Journal of Pediatric Psychology.
Berlin, L. J., Whiteside-Mansell, L., Roggman, L. A., Green, B. L., Robinson, J., &
Spieker, S. (2011). Testing maternal depression and attachment style as
moderators of Early Head Start’s effects on parenting. Attachment & Human
Development, 13(1), 49–67.
Bowlby, J. (1978). Attachment theory and its therapeutic implications. Adolescent
Psychiatry, 6, 5–33.
Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-Report measures of adult
attachment: An integrative overview. Attachment Theory and Close Relationships,
(January 1998), 46–72. 63
Brofenbrenner, U. (1979). The Ecology of Human Development: Experiments by Nature
and Design. Cambridge, MA and London, England: Harvard University Press.
Bus, A. G., & van Ijzendoorn, M. H. (1995). Mothers reading to their 3-year-olds: The
role of mother-child attachment security in becoming literate. Reading Research
Quarterly, 30(4), 998–1015.
Bus, A. G., van IJzendoorn, M. H., & Pellegrini, A. D. (1995). Joint book reading makes
for success in learning to read: A meta-analysis on intergenerational transmission
of literacy. Review of Educational Research Spring, 65(1), 1–21.
Cain, C. M., Wood, G. L., Santa Maria, D. M., Mandell, D. J., & Rozmus, C. L. (n.d.).
Early language interventions for low socioeconomic status populations: A
systematic review of the efficacy on child language, cognitive, and social
development. Developmental Review, currently under review.
Casey, B. J., Giedd, J. N., & Thomas, K. M. (2000). Structural and functional brain
development and its relation to cognitive development. Biological Psychology,
54(1–3), 241–257.
Cluxton-Keller, F., Burrell, L., Crowne, S. S., McFarlane, E., Tandon, S. D., Leaf, P. J.,
& Duggan, A. K. (2014). Maternal relationship insecurity and depressive
symptoms as moderators of home visiting impacts on child outcomes. Journal of
Child and Family Studies, 23(8), 1430–1443.
Costello, A. B., & Osborne, J. W. (2005). Best practices in exploratory factor analysis:
Four recommendations for getting the most from your analysis. Practical
Assessment, Research & Evaluation, 10(7), 1–9.
Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Edinburgh Postnatal Depression Scale 64
(EPDS): Instructions for using the Edinburgh Postnatal Depression Scale. British
Journal of Psychiatry, 150(3), 782–786.
Duggan, A. K., Berlin, L. J., Cassidy, J., Burrell, L., & Tandon, S. D. (2009). Examining
maternal depression and attachment insecurity as moderators of the impacts of
home visiting for at-risk mothers and infants. J Consult Clin Psychol, 77(4), 788–
799.
Duncan, G. J., & Magnuson, K. (2013). Investing in preschool programs. Journal of
Economic Perspectives, 27(2), 109–132.
Edelstein, R. S., Alexander, K. W., Shaver, P. R., Schaaf, J. M., Quas, J. A., Lovas, G. S.,
& Goodman, G. S. (2004). Adult attachment style and parental responsiveness
during a stressful event. Attachment & Human Development, 6(1), 31–52.
Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible
statistical power analysis program for the social, behavioral, and biomedical
sciences. Behavior Research Methods, 39(2), 175–191.
Feeney, J. A., Alexander, R., Noller, P., & Hohaush, L. (2003). Attachment insecurity,
depression, and the transition to parenthood. Personal Relationships, 10(4), 475–
493.
Feeney, J. A., Noller, P., & Hanrahan, M. (1994). Assessing adult attachment:
Developments in conceptualization of security and insecurity. In M. B. Sperling
& W. H. Berrman (Eds.), Attachment in adults: Theory, assessment, and
treatment (pp. 128–152). New York, NY: Guilford Publications.
Fenson, L., Dale, P. S., Reznick, J. S., Bates, E., Thal, D. J., & Pethick, S. J. (1994). 65
Variability in early communicative development. Monographs of the Society for
Research in Child Development, 59(5), 174-85.
Ferguson, E., & Cox, T. (1993). Exploratory factor analysis: A user’s guide.
International Journal of Selection and Assessment, 1(2), 84–94.
Fernald, A., Marchman, V. A., & Weisleder, A. (2013). SES differences in language
processing skill and vocabulary are evident at 18 months. Developmental Science,
16, 234–248.
Gilkerson, J., Coulter, K. K., & Richards, J. A. (2008). Transcriptional analyses of the
LENA natural language corpus. Boulder, CO.
Gilkerson, J., & Topping, K. (2017). Evaluation of a LENA-based online intervention for
parents of young children. Journal of Early Intervention, 1–38.
Golinkoff, R. M., Hoff, E., Rowe, M. L., Tamis-Lemonda, C. S., & Hirsh-Pasek, K.
(2018). Language matters: Denying the existence of the 30-Million-Word Gap has
serious consequences. Child Development, 1–8.
Green, B. L., Furrer, C., & McAllister, C. (2007). How do relationships support
parenting? Effects of attachment style and social support on parenting behavior in
an at-risk population. Am J Community Psychol, 40, 96–108.
Greenwood, C. R., Carta, J. J., Walker, D., Watson-Thompson, J., Gilkerson, J., Larson,
A. L., & Schnitz, A. (2017). Conceptualizing a Public Health Prevention
Intervention for Bridging the 30 Million Word Gap. Clinical Child and Family
Psychology Review, 20(1), 1–22.
Haley, D. W., & Stansbury, K. (2003). Infant stress and parent responsiveness: 66
Regulation of physiology and behavior during still-face and reunion. Child
Development, 74(5), 1534–1546.
Harris, P. A., Taylor, R., Thielke, R., Payne, J., Gonzalez, N., & Conde, J. G. (2009).
Research electronic data capture (REDCap)-A metadata-driven methodology and
workflow process for providing translational research informatics support.
Journal of Biomedical Informatics, 42(2), 377–381.
Harrison, L. J., & Wang, C. (2018). Current approaches in quanitative reserach in early
childhood education. In M. Fleer & B. van Oers (Eds.), International Handbook
of Early Childhood Education (pp. 295–316). Dordrecht, The Netherlands.
Hart, B., & Risley, T. R. (1995). Meaningful differences in the everyday experience of
young American children. Baltimore, MD: Paul H. Brookes Publishing Co.
Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process.
Journal of Personality and Social Psychology, 52(3), 511–524.
Hindman, A. H., Wasik, B. A., & Snell, E. K. (2016). Closing the 30 Million Word Gap:
Next steps in designing research to inform practice. Child Development
Perspectives, 10(2), 135-139.
Hoff, E. (2003). The specificity of environmental influence: Socioeconomic status affects
early vocabulary development via maternal speech, Child Development, 74(5),
1368–1378.
Hurt, H., & Betancourt, L. M. (2017). Turning 1 year of age in a low socioeconomic
environment: A portrait of disadvantage. Journal of Developmental & Behavioral
Pediatrics, 38(7),493-500.
Huttenlocher, J., Waterfall, H., Vasilyeva, M., Vevea, J., & Hedges, L. V. (2010). 67
Sources of variability in children’s language growth. Cognitive Psychology, 61(4),
343–365.
IBM Corporation. (2017). SPSS Statistics for Windows. SPSS Version 25.
Jones, J. D., Cassidy, J., & Shaver, P. R. (2015). Parents’ self-reported attachment styles:
a review of links with parenting behaviors, emotions, and cognitions. Personality
and Social Psychology Review, 19(1), 44–76.
Jung, T., & Wickrama, K. A. (2008). An introduction to latent class growth analysis and
growth mixture modeling. Social and Personality Psychology Compass, 2(1),
302–317.
Jutte, D. P., Miller, J. L., & Erickson, D. J. (2015). Neighborhood adversity, child health,
and the role for community development. Pediatrics, 135(Supplement), S48–S57.
Kalil, A., Ryan, R., & Chor, E. (2014). Time investments in children across family
structures. The Annals of the American Academy of Political and Social Science,
654(1), 150–168.
Karantzas, G. C., Feeney, J. a., & Wilkinson, R. (2010). Is less more? Confirmatory
factor analysis of the Attachment Style Questionnaires. Journal of Social and
Personal Relationships, 27(6), 749–780.
Knudsen, E. I. (2004). Sensitive periods in the development of the brain and behavior.
Journal of Cognitive Neuroscience, 16(8), 1412–1425.
Korfmacher, J., Green, B., Staerkel, F., Peterson, C., Cook, G., Roggman, L., …
Schiffman, R. (2008). Parent involvement in early childhood home visiting. Child
& Youth Care Forum, 37(4), 171–196.
Kuhl, P. K. (2011). Early language learning and literacy: Neuroscience implications for 68
education. Mind, Brain and Education, 5(3), 128–142.
LENA Research Foundation. (2017). Impact Assessment Strategy Guide. Boulder, CO.
LENA Research Foundation. (2018). LENA Start. Retrieved from
www.lena.org/solutions/LENA_Start
Lubke, G., & Neale, M. C. (2006). Distinguishing between latent classes and continuous
factors: Resolution by maximum likelihood? Multivariate Behavioral Research,
41(4), 499-532.
McCambridge, J., Witton, J., & Elbourne, D. R. (2014). Systematic review of the
Hawthorne effect: New concepts are needed to study research participation
effects. Journal of Clinical Epidemiology, 67(3), 267–277.
McFarlane, E., Burrell, L., Fuddy, L., Tandon, D., Derauf, D. C., Leaf, P., & Duggan, A.
(2010). Association of home visitors’ and mothers’ attachment style with family
engagement. Journal of Community Psychology, 38(5), 541–556.
McMahon, C., Barnett, B., Kowalenko, N., & Tennant, C. (2005). Psychological factors
associated with persistent postnatal depression: Past and current relationships,
defence styles and the mediating role of insecure attachment style. Journal of
Affective Disorders, 84(1), 15–24.
Meredith, P., Strong, J., & Feeney, J. A. (2006). Adult attachment, anxiety, and pain self-
efficacy as predictors of pain intensity and disability. Pain, 123, 146–154.
Mills-Koonce, W. R., Appleyard, K., Barnett, M., Deng, M., Putallaz, M., & Cox, M.
(2011). Adult attachment style and stress as risk factors for early maternal
sensitivity and negativity. Infant Mental Health Journal, 32(3), 277–285.
Muthén, L. K., & Muthén, B. O. (2017). MPlus Statistical Analysis with Latent Variables 69
User’s Guide (Eighth Edition). Los Angeles, CA.
National Scientific Council on the Developing Child. (2004). Young children develop in
an environment of relationships: Working paper No. 1. Retrieved November 6,
2016, from http://www.developingchild.net.
Nylund, K. L., Asparouhov, T., & Muthén, B. O. (2007). Deciding on the number of
classes in latent class analysis and growth mixture modeling: A Monte Carlo
simulation study. Structural Equation Modeling.
Oller, D. K. (2010). All-day recordings to investigate vocabulary development: A case
study of a trilingual toddler. Communication Disorders Quarterly, 31(4), 213–
222.
Pati, S., Hashim, K., Brown, B., Fiks, A. G., & Forrest, C. B. (2011). Early identification
of young children at risk for poor academic achievement: Preliminary
development of a parent-support prediction tool. BMC Health Services Research,
11, 197.
Preacher, K. J., Winchman, A. L., MacCallum, R. C., & Briggs, N. E. (2008). Latent
Growth Curve Modeling. Quantitative applications in the social sciences. Los
Angeles, CA: SAGE Publications.
Raikes, H., Luze, G., Brooks-Gunn, J., Abigail Raikes, H., Alexander Pan, B., Tamis-
LeMonda, C. S., … Rodriguez, E. T. (2006). Mother-child bookreading in low-
income families: Correlates and outcomes during the first three years of life.
Child Development, 77(4), 924–953.
Ram, N., & Grimm, K. J. (2009). Methods and Measures: Growth mixture modeling: A 70
method for identifying differences in longitudinal change among unobserved
groups. International Journal of Behavioral Development. 33(6), 565-576.
Rholes, W. R., Simpson, J. A., Blakely, B. S., Lanigan, L., & Allen, E. A. (1997). Adult
attachment syles, the desire to have children, and working models of parenthood.
Journal of Personality, 65(2), 357–385.
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.
Romeo, R. R., Leonard, J. A., Robinson, S. T., West, M. R., Mackey, A. P., Rowe, M. L.,
& Gabrieli, J. D. E. (2018). Beyond the 30-Million-Word Gap: Children’s
Conversational Exposure Is Associated With Language-Related Brain Function.
Psychological Science, 29(5), 700–710.
Romeo, R. R., Segaran, J., Leonard, J. A., Robinson, S. T., West, M. R., Mackey, A. P.,
… Gabrieli, J. D. E. (2018). Language Exposure Relates to Structural Neural
Connectivity in Childhood. The Journal of Neuroscience, 38(36), 7870–7877.
Roulstone, S., Loader, S., Northstone, K., & Beveridge, M. (2002). The speech and
language of children aged 25 months: Descriptive data from the Avon
Longitudinal Study of Parents and Children. Early Child Development and Care,
172(3), 259–268.
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.
Rowe, M. L. (2018). Understanding socioeconomic differences in parents’ speech to 71
children. Child Development, 12(2), 122–127.
Rubin, D. B. (1987). Multiple imputation for nonresponse in surveys. New York, NY:
Wiley.
Rushton, F. E., & Kraft, C. (2014). Building brains, forging futures: the pediatrician’s
role. International Journal of Pediatrics and Adolescent Medicine, 1(1), 3–7.
Schafer, J. L. (1997). Analysis of incomplete multivariate data. London, England:
Chapman & Hall.
Sénéchal, M., & Lefevre, J.-A. (2002). Parental involvement in the development of
children’s reading skill: A five-year longitudinal study. Child Development, 73(2),
445–460.
Shonkoff, J. P., Garner, A. S., The Committee on Psychosocial Aspects of Child and
Family, Committee on Early Childhood, Adoption, and Dependent Care, &
Section on Developmental and Behavioral Pediatrics. (2012). The lifelong effects
of early childhood adversity and toxic stress. Pediatrics, 129(1), e232–e246
Shonkoff, J. P., & Phillips, D. A. (2000). From neurons to neighborhoods: The science of
early childhood development. (J. P. Shonkoff & D. A. Phillips, Eds.), National
Academy of Sciences (Vol. 23). Washington, DC: National Academy Press.
Siddiqi, A., Irwin, L. G., & Hertzman, C. (2007). Total environment assessment model
for early child development: Evidence report. Vancouver, BC. Retrieved from
https://pdfs.semanticscholar.org/722d/f13405b23fb902e524ed8dc7c2b198fc2b89.
Smith, K. E., Landry, S. H., & Swank, P. R. (2006). The role of early maternal 72
responsiveness in supporting school-aged cognitive development for children who
vary in birth status. Pediatrics, 117(5).
Song, 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(02), 305–326.
Thurstone, L. L. (1934). The vectors of mind. Psychological Review, 226–231.
Vermunt, J. K. (2010). Latent class modeling with covariates: Two improved three-step
approaches. Political Analysis, 18(4), 450–469.
Wickrama, K. A., Lee, T. K., Walker O’Neal, C., & Lorenz, F. O. (2016). An
introduction to growth mixture models. In Higher-Order Growth Curves and
Mixture Modeling with MPlus: A Practical Guide (pp. 191–226). New York, NY:
Routledge.
Xu, D., Yapanel, U., & Gray, S. (2009). Reliability of the LENA TM Language
Environment Analysis System in Young Children’s Natural Home Environment.
Retrieved from http://www.lenafoundation.org/wp-content/uploads/2014/10/LTR-
05-2_Reliability.pdf
Xu, D., Yapanel, U., Gray, S., Gilkerson, J., Richards, J., & Hansen, J. (2008). Signal
processing for young child speech language development. WOCCI (January
2008).
Zimmerman, F. J., Gilkerson, J., Richards, J. A., Christakis, D. A., Xu, D., Gray, S., &
Yapanel, U. (2009). Teaching by listening: the importance of adult-child
conversations to language development. Pediatrics, 124(1), 342–349.
Zuckerman, B., & Augustyn, M. (2011). Books and reading: Evidence-based standard of 73
care whose time has come. Academic Pediatrics, 11(1), 11–17.
Table 1 Attachment Style Questionnaire-Short Form (ASQ-SF) Exploratory Factor Analysis Pattern Matrix (Promax Rotation)
Factor 1 Factor 2 Factor 3 HFA HFA ASQ-SF 29-item Relationship anxiety Discomfort with closeness Relationships as secondary AKb HIc Attachment Avoidancea ASQ 3 .003 .285 .023 ASQ 4 -.128 .073 -.014 ASQ 5 -.022 .014 .103 ASQ 8 .018 -.005 .700 ASQ 9 -.209 .152 .947 ASQ 10 .160 -.155 .497 ASQ 14 .230 -.092 .245 ASQ 16 .256 .426 .081 AVD AVD ASQ 17 .319 .586 -.011 AVD AVD ASQ 19 .026 -.007 .028 ASQ 20 -.007 .529 -.047 AVD AVD ASQ 21 -.186 .830 .066 AVD AVD ASQ 23 .831 .065 -.007 ANX ASQ 25 .752 .125 .079 ANX ASQ 34 .208 -.024 -.028 ASQ 37 -.224 .078 -.076
Attachment Anxietya ASQ 11 .034 .093 .058 ASQ 13 .327 -.137 .010 ANX ANX ASQ 15 .352 -.100 .260 ANX ANX ASQ 18 .361 .224 -.102 ANX ASQ 22 .677 -.003 -.069 ANX ANX ASQ 24 .725 -.028 .007 ANX ANX ASQ 27 .631 -.110 -.117 ANX ANX ASQ 29 .257 .147 -.108 ANX ANX ASQ 30 .500 -.073 -.099 ANX ANX ASQ 31 .085 -.042 .110 ASQ 32 .144 .019 .023 ANX ANX ASQ 33 .310 -.013 .113 ANX ANX ASQ 38 -.099 .003 -.098 ANX
Note. HFA = Healthy Families America; AK = Alaska; HI = Hawaii; AVD = Attachment Avoidance Factor; ANX = Attachment Anxiety Factor a. Represent items mapped to broad attachment avoidance and attachment anxiety constructs from Karantzas et al., 2010. b. Represent items mapped to broad attachment avoidance and attachment anxiety constructs from Duggan et al., 2009. c. Represent items mapped to broad attachment avoidance and attachment anxiety constructs rom McFarlane et al., 2010.
74
75
Table 2 Non-transformed and log-transformed 5 time-point parent-child interaction variables and percent nonmissing
Non-transformed variable Log-transformed Variable Mean Skewness Kurtosis Mean Skewness Kurtosis % (+ SD) (+ SE) (+ SE) (+ SD) (+ SE) (+ SE) Nonmissing Adult Word Count (per 10,000 units) Log-Transformed Adult Word Count (n = 224) (n = 224) Time 1* 1.23 (.61) 1.43 (.19) 3.67 (.38) 9.31 (.50) -.43 (.19) .97 (.38) 74.1 Time 2 1.44 (.64) 1.23 (.16) 2.75 (.33) 9.45 (.46) -.50 (.16) 1.39 (.33) 97.8 Time 3 1.61 (.68) .93 (.17) 1.31 (.33) 9.58 (.43) -.28 (.17) -.03 (.33) 95.1 Time 4 1.65 (.68) .72 (.17) .94 (.35) 9.60 (.46) -.91 (.17) 1.92 (.34) 88.8 Time 5 1.56 (.61) 1.09 (.18) 1.21 (.34) 9.56 (.39) .02 (.18) .00 (.35) 86.6
Conversational Turns per 100 units Log-Transformed Conversational Turns (n = 224) (n = 224) Time 1* 3.41 (2.08) 2.15 (.19) 8.65 (.37) 5.67 (.57) -.08 (.19) .09 (.37) 74.1 Time 2 4.02 (2.45) 1.87 (.16) 5.0 (.33) 5.81 (.58) -.23 (.16) .71 (.33) 97.8 Time 3 4.54 (2.74) 2.07 (.17) 7.26 (.33) 5.94 (.56) .04 (.17) .01 (.33) 95.1 Time 4 4.77 (3.07) 2.58 (.17) 11.63 (.34) 5.98 (.58) -.11 (.17) .62 (.34) 88.8 Time 5 4.52 (2.60) 1.83 (.17) 4.52 (.35) 5.95 (.53) .07 (.17) .42 (.35) 86.6
Reading Minutes Log-Transformed Reading Minutes (n = 223) (n = 223) Time 1 13.63 (15.36) 1.52 (.18) 1.93 (.36) 1.92 (1.35) -.28 (.18) -1.21 (.36) 80.3 Time 2 19.46 (13.21) 1.60 (.20) 3.70 (.41) 2.72 (.79) -1.32 (.20) 3.32 (.41) 63.2 Time 3 21.13 (12.46) 1.85 (.17) 5.15 (.33) 2.86 (.58) -0.575 (.17) 2.26 (.33) 95.1 Time 4 24.16 (13.22) 1.98 (.17) 5.48 (.34) 3.04 (.50) -.04 (.17) .44 (.34) 91.5 Time 5 26.31 (14.06) 2.01 (.17) 6.79 (.35) 3.14 (.49) -.04 (.17) .40 (.35) 88.3
Note.
*87 participants had two valid baseline measurements which were averaged together; 79 participants had 1 valid baseline measurement.
75
76
Table 3
Sample Characteristics Program Sample Study Survey Differences Sample n = 224 n = 108 n (%) or Mean (+ SD) n (%) or Mean (+ SD) Child Age (m)a 14.35 (8.28) 13.98 (7.76) p = .467
Child Gendera Male 136 (60.7) 58 (53.7) p = .465 Female 87 (38.8) 48 (44.4)
Parent Age (y)b - 32.33 (6.17)
Parent Genderc Male - 7 (6.5) Female - 99 (91.7)
Parent Race/Ethnicityc White, non-Hispanic - 15 (13.9) Black, non-Hispanic - 15 (13.9) Hispanic - 61 (56.5) Other - 15 (13.9)
Marital Statusc Married/Living with - 90 (83.3) Partner Single/Separated/Divorced/ - 16 (14.8) Widowed
Educational Attainmentc Less than 12th grade - 6 (5.6) High school graduate or GED - 29 (26.9) College graduate - 71 (65.7)
Household Incomed <100% of FPL - 32 (29.6) 100-200% of FPL - 25 (23.1) 200-300% of FPL - 13 (12.0) >300% of FPL - 35 (32.4)
Relationship anxietye - -.17 (.82)
Discomfort with closenesse - -.06 (.90) Relationships as secondarye - -.26 (.82)
Depression symptomsb - 7.72 (5.53)
Note. T-Test and Chi Square used to calculate differences from full program sample proportions a99.9% nonmissing data for the program sample and 98% in the study survey sample b95% nonmissing data for the study survey sample c98% nonmissing data for the study survey sample d97% nonmissing data for the study survey sample e91% nonmissing data for the study survey sample 77
Table 4
Estimated sample statistics of 5 time-point parent-child interaction variables after Markov Chain Monte Carlo Multiple Imputation Mean Skewness Kurtosis (+ SD) (+ SE) (+ SE)
Log-Transformed Adult Word Count (n = 224) Time 1 9.26 (.49) -.54 (.16) 1.21 (.32) Time 2 9.45 (.45) -.47 (.16) 1.38 (.32) Time 3 9.57 (.43) -.23 (.16) -.06 (.32) Time 4 9.59 (.46) -.79 (.16) 1.67 (.32) Time 5 9.55 (.38) .08 (.16) .05 (.32)
Log-Transformed Conversational Turns (n = 224) Time 1 5.63 (.56) -.19 (.16) .35 (.32) Time 2 5.82 (.58) -.23 (.16) .69 (.32) Time 3 5.93 (.56) .04 (.16) .00 (.32) Time 4 5.96 (.57) -.05 (.16) .60 (.32) Time 5 5.93 (.52) .10 (.16) .50 (.32)
Log-Transformed Reading Minutes (n = 223) Time 1 2.57 (.65) .33 (.16) -.21 (.33) Time 2 2.74 (.56) -.09 (.16) .69 (.33) Time 3 2.85 (.55) -.08 (.16) .33 (.33) Time 4 3.05 (.49) .00 (.16) .49 (.33) Time 5 3.14 (.48) -.03 (.16) .45 (.33)
78
Table 5 Intercorrelations for Continuous/Ordinal Predictor Variables
Variable 1 2 3 4 5 6 7 8 9
1. Relationship anxiety - -.06 .45** .27** -.11 -.04 .00 .03 -.12
2. Discomfort with .20* -.10 .10 -.21* -.20 .12 .22* closeness
3. Relationships as .16 -.10 -.30** .03 .12 -.02 secondary
4. Depression symptoms -.06 -.03 -.12 .03 -.10
5. Parent/guardian age .29** .14 .18 .17 (Years)
6. Educational .53** .09 .23* attainment
7. Income level (% FPL) .27** .15
8. Child age (Months) .16*
9. Intervention attendance -
Note. Spearman’s Rho *p<.05; **p<.01 Correlations > .5 in bold
78
79
Table 6 Class membership, fit indices, and parameter estimates for adult word count LCGA
Model Classification 1 Class 2 Class 3 Class 4 Class
Sample Size Class 1 224.00 84.64 139.85 98.10 Class 2 139.36 45.48 28.49 Class 3 38.67 71.67 Class 4 24.85
Model Fit Indices Free parameters 7 10 13 16 Akaike’s Information Criterion (AIC) 1366.82 131.91 828.79 777.29 Bayesian Information Criterion (BIC) 1390.70 1066.03 873.14 821.88 Entropy .79 .88 .82 Integrated Completed Likelihood Criterion with BIC (ICL-BIC) 1390.70 1130.93 931.22 945.53 Vuong-Lo-Mendell-Rubin Likelihood Ratio Test (VLMR), p .058 .004 .117 Bootstrapped Likelihood Ratio Test (via 100 draws), p <.001 <.001 <.001
Latent variable means Class 1 Intercept 9.30 (.03)* 8.90 (.11)* 9.32 (.05)* 9.50 (.14)* Class 1 Slope .19 (.02)* .24 (.03)* .16 (.02)* .14 (.03)*
Class 2 Intercept 9.55 (.07)* 9.8 (.06)* 9.90 (.19)* Class 2 Slope .16 (.02)* .15 (.04)* .19 (.04)*
Class 3 Intercept 8.62 (.10)* 9.10 (.19)* Class 3 Slope .30 (.06)* .19 (.04)*
Class 4 Intercept 8.50 (.32)* Class 4 Slope .32 (.10)*
Note. *p < .001
Estimated standard errors are shown in parenthesis. 79
80
Table 7
Class membership, fit indices, and parameter estimates for conversational turns LCGA
Model Classification 1 Class 2 Class 3 Class
Sample Size Class 1 224.00 94.87 69.71 Class 2 129.13 123.79 Class 3 30.50
Model Fit Indices Free parameters 7 10 13 Akaike’s Information Criterion (AIC) 1878.43 1379.43 1200.55 Bayesian Information Criterion (BIC) 1902.32 1413.54 1244.90 Entropy .86 .89 Integrated Completed Likelihood Criterion with BIC (ICL-BIC) 1902.32 1456.09 1298.55 Vuong-Lo-Mendell-Rubin Likelihood Ratio Test (VLMR), p .003 .065 Bootstrapped Likelihood Ratio Test (via 100 draws), p <.001 <.001
Latent variable means Class 1 Intercept 5.66 (.04)* 6.11 (.07)* 6.22 (.07)* Class 1 Slope .20 (.02)* .20 (.03)* .22 (.03)*
Class 2 Intercept 5.33 (.07)* 5.55 (.06)* Class 2 Slope .20 (.03)* .18 (.03)*
Class 3 Intercept 4.84 (.03)* Class 3 Slope .26 (.11)*
Note. *p < .001 Estimated standard errors are shown in parenthesis.
80
81
Table 8 Class membership, fit indices, and parameter estimates for reading minutes LCGA
Model Classification 1 Class 2 Class 3 Class
Sample Size Class 1 224.00 94.87 69.71 Class 2 129.13 123.79 Class 3 30.50
Model Fit Indices Free parameters 10 13 16 Akaike’s Information Criterion (AIC) 1802.22 1447.35 1243.38 Bayesian Information Criterion (BIC) 1836.16 1491.47 1297.68 Entropy .80 .89 Integrated Completed Likelihood Criterion with BIC (ICL-BIC) 1836.16 1552.16 1350.37 Vuong Lo-Mendell-Rubin Likelihood Ratio Test (VLMR), p .041 .002 Bootstrapped Likelihood Ratio Test (via 100 draws), p <.001 <.001
Latent variable means Class 1 Intercept 2.57 (.04)* 2.17 (.07)* 2.07 (.06)* Class 1 Slope .57 (.04)* .63 (.06)* .54 (.07)*
Class 2 Intercept 2.94 (.12)* 3.46 (.11)* Class 2 Slope . 51 (.08)* .43 (.12)*
Class 3 Intercept 2.64 (.06)* Class 3 Slope .62 (.06)*
Note. *p < .001 Estimated standard errors are shown in parenthesis.
81
82
Table 9
Multinomial logistic regression of the likelihood of adult word count latent class trajectory categorization in comparison to the lowest growth trajectory as the reference group Highest growth trajectory Mid growth trajectory Variable Coefficient Odds Ratio Coefficient Odds Ratio
Bivariate Unadjusted
Relationship anxiety .324 1.383 .438 1.550 Discomfort with closeness .529 1.697 .152 1.164 Relationships as secondary .534 1.706 .422 1.525
Adjusted w/ covariates†
Relationship anxiety .030 1.030 .349 1.418 Discomfort with closeness .490 1.632 .282 1.326 Relationships as secondary .226 1.254 .295 1.343 Intervention attendance -.314 .731 -.102 .903 Parent Age (y) -.151 .860 .015 1.015 Hispanic 1.009 2.743 -.454 .635 Single/Separated/ .390 1.477 .491 1.634 Divorced/ Widowed Income < FPL 1.634 5.124 -.367 .693 Child (male) -.066 .936 -.708 .493 Child Age (m) .042 1.043 -.059 .943
Note. †Race was not included in the model due to singularities in classes; Depression scores were not included in the model due to estimation producing negative covariances. No predictors or covariates reached statistical significance.
83
Table 10
Binary logistic regression of the likelihood of conversational turns latent class trajectory categorization in comparison to the lowest growth trajectory as the reference group Highest growth trajectory Variable Coefficient Odds Ratio
Bivariate Unadjusted
Relationship anxiety .012 1.012 Discomfort with closeness -.633 .531* Relationships as secondary -.245 .783
Adjusted w/ covariates†
Relationship anxiety .385 1.470 Discomfort with closeness -1.091 .336** Relationships as secondary -.158 .854 Depression -.043 .958 Intervention attendance .022 1.022 Parent Age (y) .040 1.041 Black, non-Hispanic -.929 .395 Hispanic -.488 .614 Other race, non-Hispanic 1.121 3.067 Income < FPL -1.022 .360 Child (male) 1.213 3.365 Child Age (m) .147 1.158*
Note. †Martial status was not included in the model due to singularities in classes *p<.05; **p<.001
84
Table 11
Binary logistic regression of the likelihood of shared book reading latent class trajectory categorization in comparison to the lowest growth trajectory as the reference group Highest growth trajectory Variable Coefficient Odds Ratio
Bivariate Unadjusted
Relationship anxiety .163 1.177 Discomfort with closeness .066 1.068 Relationships as secondary -.048 .953
Adjusted w/ covariates
Relationship anxiety .148 1.159 Discomfort with closeness .079 1.082 Relationships as secondary .171 1.187 Depression .016 1.016 Intervention attendance -.020 .980 Parent Age (y) .052 1.053 Black, non-Hispanic -.803 .448 Hispanic .905 2.472 Other race, non-Hispanic -.853 .426 Single/Separated/ .473 1.605 Divorced/ Widowed Income < FPL .152 1.164 Child (male) -.499 .607 Child Age (m) -.134 .875*
Note. *p<.05
85
Table 12
Binary logistic regression of the likelihood of intervention attendance (> 10 classes) Attendance at >10 classes Variable Coefficient Odds Ratio
Bivariate Unadjusted
Relationship anxiety -.242 .785 Discomfort with closeness .576 1.778* Relationships as secondary -.001 .999
Adjusted w/ covariates
Relationship anxiety -.049 .953 Discomfort with closeness .956 2.602* Relationships as secondary -.275 .760 Depression .016 1.016 Parent Age (y) .065 1.067 White, non-Hispanic† Black, non-Hispanic -.783 .457 Hispanic -.637 .529 Other race, non-Hispanic .886 2.426 Married/Living with partner† Single/Separated/ .754 2.126 Divorced/ Widowed Income < 100% FPL -.162 .851 Income 100-200% FPL -.444 .641 Income 200-300% FPL -2.846 .058* Income > 300% FPL† Child (male) 1.465 4.327* Child Age (m) .106 1.112*
Note. †Reference category *p<.05 86
Emergency Center Waiting Early Language Development Program Early Head Start Program Room Parents
*LENA Digital *Self-Report Program Survey Data Survey Data Survey Data
Language Shared Book Attendance Data (n=108) (n=68) (n=105)
Processor Data Reading Data (n=224)
(n=224) (n=223)
Matched Data for Attachment Style Regression Analyses Questionnaire Factor (n=87) Analyses (n=281)
*Data used for latent class growth analyses
Figure 1. Study sample diagram
86
87
C
i s
T1 T2 T3 T4 T5
E1 E2 E3 E4 E5
Figure 2. 5 time-point latent class growth model diagram (linear)
Predictors and C covariates
i s
T1 T2 T3 T4 T5
E1 E2 E3 E4 E5 Figure 3. 5-time point latent class growth model diagram with predictors and covariates (linear) 88
Highest Growth Trajectory (17.0%) Mid Growth Trajectory (63.4%) Lowest Growth Trajectory (19.6%) 11
10
9
TRANSFORMED)
- (LOG
8 NUMBER OF ADULT WORDS SPOKEN TO CHILDTO SPOKEN ADULT WORDS OF NUMBER
7 1 . 0 0 2 . 0 0 3 . 0 0 4 . 0 0 5 . 0 0 TIME (AVERAGED OVER 13 WEEKS)
Figure 4. Estimated latent growth trajectories for log-transformed adult word count observations over 13 weeks in an early language development program.
89
Highest Growth Trajectory (42.4%) Lowest Growth Trajectory (57.6%)
7
TRANSFORMED) - 6
5 NUMBER OF CONVERSATIONAL TURNS (LOG TURNS CONVERSATIONAL OF NUMBER 4 1 2 3 4 5 TIME (AVERAGED OVER 13 WEEKS) Figure 5. Estimated latent growth trajectories for log-transformed conversational turn observations over 13 weeks in an early language development program.
90
Highest Growth Trajectory (48.4%) Lowest Growth Trajectory (51.6%) 5
4
TRANSFORMED) - 3
2
1 ESTIMATED READING MINUTES (LOG MINUTES READING ESTIMATED
0 1 2 3 4 5 TIME (AVERAGED OVER 13 WEEKS)
Figure 6. Estimated latent growth trajectories for log-transformed shared reading minutes over 13 weeks in an early language development program.
91
Cary Cain, MPH, RN Cizik School of Nursing University of Texas Health Science Center at Houston 6901 Bertner Avenue Houston, TX 77030 Tel: 832-483-2383 E-mail: [email protected]
1/11/2019
Editorial Board Developmental Review
Dear Editor,
Enclosed please find our manuscript entitled “Early Language Interventions for Low Socioeconomic Status Populations: A Systematic Review of the Efficacy on Child Language, Cognitive, and Social Development” which we would like to submit for consideration for publication as a Review Article in Developmental Review.
Children from lower socioeconomic status (SES) backgrounds are underexposed to language during a critical period of brain development leading to language disparities. Effective language interventions have been developed and tested for low SES children; however, to date, a systematic review on the efficacy of these early language interventions on child language, cognitive, and social development has not been conducted. In this manuscript, we analyzed the articles reported between January, 1995, and October, 2018 from PubMed, Embase, PsycINFO, and CINAHL databases. A total of 15 intervention studies were identified from the literature. Our review synthesizes the efficacy of these interventions on child developmental outcomes.
On behalf of all authors, please find the required documents submitted for publication consideration. All authors have read and approved the submitted manuscript, the manuscript has not been submitted elsewhere nor published elsewhere in whole or in part. The authors have no conflicts of interest to report and have adhered to all authorship and ethical adherence best practices.
We greatly appreciate your time and look forward to hearing from you soon.
Regards,
Cary Cain MPH, RN
92
Manuscript #2 Impact of Adult Attachment Style on Outcome Trajectories in a Parent-focused
Early Language Development Intervention
93
Impact of Adult Attachment Style on Outcome Trajectories in a Parent-focused
Early Language Development Intervention
Abstract
Approximately a quarter of all children in the United States live in households that are at or below the federal poverty level. Children raised in low socioeconomic status
(SES) households and neighborhoods are more likely to experience negative behavioral and cognitive outcomes compared to their more affluent peers. These disparities begin early in life, prior to four years of age. In particular, children from lower SES backgrounds are underexposed to language during a critical period of brain development leading to language disparities. This language disparity has been linked to a child’s academic attainment, which is a predictor of their future success and contribution to society, and so it is suggested that interventions to address this issue should begin early in life. Effective interventions have been developed and tested for low SES children preschool-age and older and for children with developmental delays, such as speech and language delays. To date, a systematic review on the efficacy of these early language interventions for young children from low SES populations on child language, cognitive, and social development has not been conducted. This article synthesizes results from 15 studies to appraise the current evidence of early language interventions for this population. It is clear from this review that no one intervention scope or modality is consistently associated with improved language, cognitive, or social-emotional developmental outcomes. Interventions showed promising evidence of intervention effects on child outcomes; however, the heterogeneity of the interventions and assessment instruments, in addition to the variability of observed differences in
94 outcomes, suggests that further research is needed to establish conclusive evidence on the efficacy of early language interventions for low SES populations on child language, cognitive, and social outcomes.
Keywords: early childhood development, early language intervention, intervention efficacy, language outcomes, cognitive outcomes, social outcomes
Abbreviations: Socioeconomic status (SES), Randomized controlled trial (RCT). Cluster randomized controlled trial (CRCT)
95
Background Experiences and environmental input during the first three years of life shape a child’s developmental trajectory (Gilmore et al., 2007; Heckman, 2006; Huttenlocher,
Haight, Bryk, Seltzer, & Lyons, 1991; MacLeod & Nelson, 2000; Nelson, 2000;
Nowakowski, 2006; Shonkoff & Phillips, 2000). Socioeconomic related differences in child development have been observed early in life (Hurt & Betancourt, 2017). In an influential study documenting socioeconomic influences on early language development,
Hart and Risley (1995) observed that children from low socioeconomic status (SES) households were exposed to approximately 1,500 fewer words per hour compared to children from higher SES households. They estimated that by the time these children reached four years of age they would be exposed to approximately 32 million fewer words than their higher SES peers, now commonly known as the “30-million Word Gap”
(Hart & Risley, 2003). While the study has been criticized for its small sample size and methodological flaws (Sperry, Sperry, & Miller, 2018), other studies have also found
SES-related disparities in quantity and quality of language exposure with significant language outcome differences noted among children, including those younger than 18 months of age (Fernald, Marchman, & Weisleder, 2013; Hoff, 2003; Hurt & Betancourt,
2017; Huttenlocher, Waterfall, Vasilyeva, Vevea, & Hedges, 2010; Rowe, 2012, 2018).
SES impacts on child development are not specific to language and also have been associated with disparities in cognitive and social development, which widen as the child ages (Bradley & Corwyn, 2002; Weisleder & Fernald, 2013). Longitudinal observational studies have shown that language skills, executive functioning, and social- emotional development are associated with SES (Matte-Gagné & Bernier, 2011; Petersen et al., 2013; Roben, Cole, & Armstrong, 2012). SES disparities early in life place a child
96 at heightened risk for gaps in educational attainment (Durham, Farkas, Hammer, Bruce
Tomblin, & Catts, 2007). Gaps in educational attainment can potentially lead to the need for special education services, school dropout, juvenile delinquency, adolescent pregnancy, increased emergency department and hospitalization visits, decreased economic productivity, unemployment, dependency on social services, and poor parenting skills (Brownell et al., 2016; Doyle, Harmon, Heckman, & Tremblay, 2009;
Ramey & Ramey, 1998).
Recently there has been widespread dissemination of information on the “30-
Million Word Gap” in the popular media, as well as substantial investments from foundations such as the Clinton Foundation and Bloomberg Philanthropies that promote implementation of early language interventions targeted towards low SES children
(Clinton Foundation, n.d.; “Provid. Talks,” n.d.; Rosenberg, 2013; Strauss, 2015; Talbot,
2015; Talk with me Baby, n.d.; University of Chicago School of Medicine, n.d.; Warren,
2015). Therefore, it is timely to synthesize the research on early language interventions targeted toward low SES populations.
Several systematic reviews and meta-analyses support the associations of SES and child brain development, including the impact of interventions that address early language learning among specific populations (Camilli, Vargas, Ryan, & Barnett, 2010;
Hurt & Betancourt, 2016; Marulis & Neuman, 2013; Needlman & Silverstein, 2004;
Roberts & Kaiser, 2011; Senechal & Young, 2008; Te Kaat-van den Os, Jongmans,
Volman, & Lauteslager, 2017). The reviews and meta-analyses on early language interventions have focused on children with developmental delays, older children that have entered preschool and kindergarten, or universal interventions that promote early
97 language development implemented through pediatric primary care. Though the literature is robust in documenting the impact of SES on child development and emphasizing the need for early intervention to reduce these SES-related disparities, no review to date has integrated the findings on the efficacy of early language interventions targeted to a population of low SES children prior to entry into preschool. The purpose of this systematic review was to examine the efficacy of early language interventions on language, cognitive, and social outcomes for children less than four years of age from low SES populations.
Methods
Search Parameters
Studies were identified using a librarian assisted systematic search of the peer- reviewed literature using PubMed, Embase, PsycINFO, and Cinahl databases and three broad search concepts: early language development, low socioeconomic status, and infant/child. Keywords and MeSH terms/subject headings were used for the following search terms: early language ("language development," "child language," "early literacy,"
"language development," and "child language"); AND low socioeconomic status
("poverty," "socioeconomic factors," "low income population," "low SES," “lowest income group,” “lower income level,” and "socioeconomic factors"); AND infant/child
(“infant," "child, preschool,” “toddler,” and “child"). A hand search of the reference lists in the studies retrieved was conducted to identify studies not retrieved through the database search.
98
Inclusion and Exclusion Criteria
Retrieved studies were screened according to specific inclusion and exclusion criteria which are listed in Table 1. This review was limited to randomized controlled trials and cluster randomized controlled trials, as the outcomes are age-dependent, and comparison of equal groups is imperative to understand if differences are attributed to the intervention or the natural maturation of the child. The studies had to focus on low SES, typically developing children younger than four years old. Hart and Risley (1995) published their findings on the early language development disparity for young children under 4 years of age in 1995; therefore, the studies reviewed included publication from
January 1995 through October 2018. Seven studies included participants from the same dataset, the Early Head Start Longitudinal Research and Evaluation Project dataset
(Ayoub, Vallotton, & Mastergeorge, 2011; Cabrera, Karberg, Malin, & Aldoney, 2017;
Chapin & Altenhofen, 2010; Love et al., 2005; Raikes et al., 2006; Spieker, Nelson,
Petras, Jolley, & Barnard, 2003; Vallotton et al., 2012). Six of these studies did not directly assess the intervention effects on child outcomes and thus these studies were excluded from this review (Ayoub et al., 2011; Cabrera et al., 2017; Chapin &
Altenhofen, 2010; Raikes et al., 2006; Spieker et al., 2003; Vallotton et al., 2012).
Data Collection and Analyses
Database searches were exported to RefWorks and duplicate studies were removed (ProQuest, 2016). Inclusion and exclusion criteria were applied to the studies obtained in the search through title, abstract, and full text review, as documented in the
PRISMA Flow Diagram (Figure 1) (Moher, Liberati, Tetzlaff, Altman, & The PRISMA
Group, 2009).
99
Data extraction was conducted using the Cochrane Collaboration data extraction checklist which included publication information, study design, sample size and demographics, specific intervention characteristics, methodological variables, and language, cognitive, and social outcomes (Tables 2-5) (Higgins & Green, 2011).
Analyses of the studies were structured to determine the efficacy of the early language interventions on language, cognitive, or social development outcomes. Cohen’s d effect sizes were calculated for studies that did not report the effect size using methodology from Thalheimer and Cook (2002).
Results
The search yielded 3,356 potential articles which were screened to determine relevance and consistency with the a priori inclusion and exclusion criteria. One hundred and fifty-six articles were identified as potentially relevant for full text review. Upon review of the full text articles, 15 articles met the inclusion and exclusion criteria to be included in this synthesis (Figure 1). Twelve studies were randomized controlled trials
(RCT) (Cronan, Cruz, Arriaga, & Sarkin, 1996; High, LaGasse, Becker, Ahlgren, &
Gardner, 2000; Love et al., 2005; Lunkenheimer et al., 2008; McGillion, Pine, Herbert, &
Matthews, 2017; Mendelsohn et al., 2005; Olds et al., 2014, 2002; Olds, Kitzman, et al.,
2004; Olds, Robinson, et al., 2004; Suskind et al., 2016; Vally, Murray, Tomlinson, &
Cooper, 2015) and three studies were cluster randomized controlled trials (CRCT)
(Goldfeld et al., 2011, 2012; Landry et al., 2014).
Methodological Rigor
The Cochrane Collaboration Risk of Bias Tool was used to assess the methodological quality of the studies (Higgins et al., 2011). This tool includes six
100 domains of bias for assessing internal validity: selection, performance, detection, attrition, reporting, and other bias, and assigns a domain score of high, low, or unclear.
All studies used randomized or cluster randomized controlled designs to evaluate intervention efficacy; however, the rigor in controlling for internal validity varied across studies. All studies provided inclusion/exclusion criteria and descriptions of the intervention and control conditions. Two studies had small sample sizes (Suskind et al.,
2016; Vally et al., 2015). The randomization sequence was reported in thirteen of the studies; five studies reported that randomization was conducted by a statistician or researcher that was not involved with the study or not involved with recruitment (Love et al., 2005; Lunkenheimer et al., 2008; McGillion et al., 2017; Olds et al., 2014, 2002;
Olds, Kitzman, et al., 2004; Olds, Robinson, et al., 2004), one study used coin toss to randomize subjects (Mendelsohn et al., 2005), one study drew allocation status from a bag (Vally et al., 2015), and one study randomized by clinic appointment day (High et al.,
2000) The studies with higher potential for bias in their randomization of subjects are the cluster randomized controlled trials, as participants were randomized by site according to certain criteria (Goldfeld et al., 2011, 2012; Landry et al., 2014). All studies assessed and reported study group comparability at baseline. One study, which was a pilot randomized controlled trial with a small sample size, reported the presence of group differences at baseline (Suskind et al., 2016).
Most studies did not report participant blinding and thus assessment of performance bias was unclear. The benefit of the cluster randomized design, in this case, is that randomization at the site level reduced risk for contamination based on awareness of allocation status. Only one study reported that families randomly allocated to the
101 control group were not allowed to enroll in the program and these families were aware of their allocation status (Love et al., 2005). Due to the nature of the interventions, most were single blinded studies with research staff blinded to allocation status and risk for detection bias was low. Attrition was reported in eleven studies with a mean of 22%
(ranging from 10% to 42%) (Goldfeld et al., 2011, 2012; High et al., 2000; Landry et al.,
2014; Love et al., 2005; Lunkenheimer et al., 2008; McGillion et al., 2017; Mendelsohn et al., 2005; Olds, Kitzman, et al., 2004; Suskind et al., 2016; Vally et al., 2015). Most studies used an intention-to-treat analysis; one study reported using a per protocol analysis, which was the study with the largest sample size and largest attrition rate (Love et al., 2005). Reporting bias was unclear for many of the studies, though most studies reported both favorable and unfavorable results. This review only included studies that were published in peer-reviewed journals and therefore there is potential risk for publication bias. Other risk of bias in interpretation of outcomes that should be noted were studies’ lack of reporting on reliability and validity of an outcome assessment instrument (Cronan et al., 1996) and baseline assessments which occurred after the start of the intervention in one study (Landry et al., 2014).
Sample
As noted in Table 2, 11 of the studies were conducted in the United States (73%)
(Cronan et al., 1996; High et al., 2000; Landry et al., 2014; Love et al., 2005;
Lunkenheimer et al., 2008; Mendelsohn et al., 2005; Olds et al., 2014, 2002; Olds,
Kitzman, et al., 2004; Olds, Robinson, et al., 2004; Suskind et al., 2016), two from
Australia (Goldfeld et al., 2011, 2012), and one study conducted in the United Kingdom
(McGillion et al., 2017) and one in South Africa (Vally et al., 2015). The classification of
102 low SES varied from study to study (each study’s low SES classification criteria is noted in Table 1 under “population”). The average number of participants in the intervention condition was N=264, ranging from 18 to 1,513 participants, and the control condition averaged N=287, ranging from 19 to 1,488 participants. Two studies did not report specific intervention and control sample sizes but gave the total of the entire sample
(Landry et al., 2014; Lunkenheimer et al., 2008). The mean age of participants at baseline was 15 months, ranging in age from 2 weeks to 2.9 years. The gender of the participants averaged 301 females and 317 males across the studies that reported child gender. Only studies conducted in the United States reported race and ethnicity of the child participants; the percentage of race/ethnicity across all reporting studies was Blacks
(40.4%), Whites (30.3%), Hispanic/Latino (23.1%), and other (6.3%).
Interventions
Information about the interventions can be found in Table 1. Eight of the fifteen studies included were evaluations of home visitation interventions (Cronan et al., 1996;
Lunkenheimer et al., 2008; McGillion et al., 2017; Olds et al., 2014, 2002; Olds,
Kitzman, et al., 2004; Olds, Robinson, et al., 2004; Suskind et al., 2016), four of the studies evaluated interventions conducted in a clinic setting (Goldfeld et al., 2011, 2012;
High et al., 2000; Mendelsohn et al., 2005), one intervention was group-based (Vally et al., 2015), one was childcare center-based (Landry et al., 2014), and one was a mixture of child-care center-based and home visitation (Love et al., 2005). Additionally, six of the interventions were primarily focused on early language development (Cronan et al.,
1996; Goldfeld et al., 2011, 2012; High et al., 2000; McGillion et al., 2017; Suskind et al., 2016; Vally et al., 2015), while six interventions were more broadly focused on child
103 development (Landry et al., 2014; Love et al., 2005; Lunkenheimer et al., 2008;
Mendelsohn et al., 2005; Olds et al., 2014, 2002; Olds, Kitzman, et al., 2004; Olds,
Robinson, et al., 2004). There was variation among the components of the interventions.
Most of the exclusively early language development focused interventions included provision of a developmentally appropriate children’s books, instruction, modeling or skills practice, and used various methods for feedback. The interventions more broadly focused on child development provided development education and assessed the needs of the family with integrated case management/resource referrals and navigation. One intervention focused on training childcare providers in responsive teaching (Landry et al.,
2014). The person delivering the intervention varied across studies with healthcare providers (physicians and nurses) being the most common (47%). The majority of the interventions (93%) were delivered to the parent as an agent to impact the development of the study child, while one intervention was delivered to childcare providers (Landry et al., 2014).
There was wide diversity in the duration and intensity of the interventions. Low intensity interventions included three 2-5 minute clinic visits over the course of a year and a half (Goldfeld et al., 2011). Another intervention included three home visits per year over a two year period (Lunkenheimer et al., 2008). Two of the interventions were moderate intensity, which were delivered weekly over the course of eight weeks (Suskind et al., 2016; Vally et al., 2015). Higher intensity interventions included 18 30-minute home visits over one academic year (Cronan et al., 1996), 900 activities conducted over one academic year (Landry et al., 2014), and programs that enroll the mother prenatally,
104 which can last up to 2-3 years with weekly home visits (Love et al., 2005; Olds et al.,
2014, 2002; Olds, Kitzman, et al., 2004; Olds, Robinson, et al., 2004).
Outcome Measurements
Tables 2-4 report the language, cognitive, and social outcomes for each study.
The majority of the studies measured outcomes immediately post-intervention (53%), one study included measurements during the intervention (Suskind et al., 2016), one study measured outcomes one-month post-intervention (High et al., 2000), while four studies measured outcomes six-months and one-year after intervention completion (Goldfeld et al., 2011, 2012; Lunkenheimer et al., 2008; McGillion et al., 2017). Three studies from the Nurse Family Partnership intervention measured outcomes longitudinally four and seven years after intervention completion (Olds et al., 2014; Olds, Kitzman, et al., 2004;
Olds, Robinson, et al., 2004).
Language development. The most common outcome measured among the studies was language; every study included at least one language development outcome.
Thirteen instruments were used to measure various language outcomes (Table 3). These language outcomes across the fifteen studies were categorized in to four domains: language comprehension, language production, conceptual knowledge, and general child language/communication. Most of the instruments used to assess language development were direct assessments administered by the examiner while observing the child complete a selected task or to observe the parent and child in play situations. One of the instruments used in two of the studies was a digital language processor which records and analyzes the number of child vocalizations and conversational turns throughout the day in the home environment (McGillion et al., 2017; Suskind et al., 2016). Five studies used
105 the MacArthur-Bates Communicative Development Inventory (CDI), which is a parent report instrument commonly used to assess language development in young children
(Fenson et al., 2000; Law & Roy, 2008). Most instruments used to measure language outcomes reported acceptable reliability and validity. The only instrument included in the synthesis that did not report reliability and validity was the PRIMER Language
Comprehension Book, which was adapted by the study authors from the CDI as an observational, rather than parent report measure (Cronan et al., 1996). As a group, the thirteen studies resulted in seventeen different measurable language outcomes, with four outcomes (expressive language, expressive vocabulary, receptive language, and language comprehension) measured by more than one instrument. Nine of the intervention studies reported at least one positive intervention effect on language development; however, the reported effect sizes were small.
Language comprehension is understanding language, which precedes language production (Gleason & Ratner, 2009). Nine studies reported language comprehension outcomes, most commonly measuring receptive language outcomes. The seven instruments used to measure language comprehension were PRIMER Language
Comprehension Book, Sutherland Phonological Awareness Test-Revised, Clinical
Evaluation of Language Fundamentals Preschool - Second Edition, MacArthur-Bates
CDI, Preschool Language Scale, Preschool Comprehensive Test of Phonological and
Print Processing, and the Peabody Picture Vocabulary Test (Cronan et al., 1996; Dunn &
Dunn, 1997; Fenson et al., 2000; Lonigan, Wagner, Torgesen, & Rashotte, 2002; Nielsen,
2003; Wiig, Secord, & Semel, 2004; Zimmerman, Steiner, & Pond, 2002). The findings were inconsistent with some studies reporting significant differences between groups on
106 language comprehension outcomes and other studies reporting no significant differences noted between groups.
Studies with significant improvements in language comprehension for the intervention group compared to control included home visitation models (Cronan et al.,
1996; Olds, Kitzman, et al., 2004), a clinic-based model (High et al., 2000), a mixed- approach of both home visitation and center-based models (Love et al., 2005), and a community-based group model (Vally et al., 2015). Of those interventions that had significant improvements in language comprehension for the intervention group compared to control, three of the interventions were specifically focused on language development (Cronan et al., 1996; High et al., 2000; Vally et al., 2015) and two of the interventions were broadly focused on child development (Love et al., 2005; Olds,
Kitzman, et al., 2004). The interventions that had significant outcomes in language comprehension outcomes were delivered over one to three years. Most of the reported effect sizes for language comprehension outcomes were small; however, High et al.
(2000) noted large intervention effects specifically for receptive vocabulary outcomes (p
= .004; d = 4.10).
Language production is the use of language to communicate with others. Eight studies reported language production outcomes. The most common language production outcome reported was expressed vocabulary, which was measured in six studies using four different measurement instruments (Goldfeld et al., 2011, 2012; High et al., 2000;
Landry et al., 2014; McGillion et al., 2017; Mendelsohn et al., 2005; Suskind et al.,
2016). Other measures of language production were social communication, vocalizations, utterances, word types, and conversational turns. The instruments used to measure
107 language production outcomes were the MacArthur-Bates CDI, Communication and
Symbolic Behavior Scales Developmental Profile Infant-Toddler Checklist, Clinical
Evaluation of Language Fundamentals Preschool - Second Edition, Expressive One-
Word Picture Vocabulary Test, Preschool Language Scale, the Goldin-Meadow, Levine,
Hedges, Huttenlocher, Raudenbush, and Small Coding System, and Language
Environment Analysis (LENA) Digital Language Processor. Respectively, two of these instruments are parent report, four are direct assessments, and one is a digital language processor to observe language production in the home. Four of the eight interventions that reported on language production outcomes found significant differences in the intervention group compared to the control group; three of the interventions were delivered via home visitation (Cronan et al., 1996; McGillion et al., 2017; Suskind et al.,
2016) and one intervention was delivered via a child development specialist in the clinic
(Mendelsohn et al., 2005). Suskind et al (2016) found significant differences in language production outcomes with medium effect sizes during the intervention and one week after the intervention; however, when measured again 4 months after the intervention the gains in language production for the intervention compared to control group were not sustained. Mendelsohn et al. (2015) noted significant increases in expressive language with a large effect size only for children of higher educated mothers (p = .008; d = 1.05).
McGillion et al. (2017) found significantly higher intervention effects on expressive vocabulary at 18 months of age for children that were in the low SES population; however, these gains were not sustained when re-assessed at 24 months of age.
General child language outcomes. Composite scores of overall language outcomes, which include articulation, receptive/expressive language, and conceptual
108 knowledge were reported by four studies (Goldfeld et al., 2012; Lunkenheimer et al.,
2008; Olds et al., 2002; Vally et al., 2015). These general language outcomes were measured by different instruments in each study. Three of the instruments were direct assessments: The Clinical Evaluation of Language Fundamentals – Preschool Australian
Second Edition, Fluharty-2 Preschool Speech and Language, and Preschool Language
Scale-3. One instrument was a parent report: The MacArthur-Bates CDI. Only one study,
Vally et al (2015), reported large significant intervention effects on the composite child language score (p < .001; d = .99) measured by the CDI global language scores, which is a combination of language comprehension and expression subscales.
Conceptual knowledge. Two studies measured symbolic and conceptual development as an outcome. Cronan et al. (1996) used the Bracken Basic Concept Scale to measure conceptual knowledge and Goldfeld et al. (2011) used the Communication and Symbolic Behavior Scale Infant-Toddler Checklist to measure symbolic understanding of words and use of objects. Neither study found significant differences in conceptual knowledge.
Cognitive development. Eight instruments were used to measure various aspects of child cognitive development (Table 4). Five studies measured cognitive outcomes immediate postintervention (Landry et al., 2014; Love et al., 2005; Lunkenheimer et al.,
2008; Mendelsohn et al., 2005; Olds et al., 2002) and three studies measured cognitive outcomes two to seven years postintervention (Olds et al., 2014; Olds, Kitzman, et al.,
2004; Olds, Robinson, et al., 2004). Most of the instruments used to assess cognitive development were examiner administered child assessments where the examiner observed the child complete a specific task or observations were made of the parent and
109 child in play situations. Three studies reported the Mental Development Index (MDI), measured by the Bayley Scales of Child Development, which is a composite measure of early cognitive and language development (Lowe, Erickson, Schrader, & Duncan, 2012).
Love et al. (2005) reported significant intervention effects on MDI for the Early Head
Start intervention (p = .01; d = .12). Mendelson et al. (2005) found that the change in the
MDI was significant with a large effect size (p = .01; d = .94) for children of higher educated mothers (7-11 grade education), while no significant intervention effects in the
MDI were noted for children of mothers with a 6th grade or lower education. Two studies of the Nurse Family Partnership program reported the Kaufman Mental Processing
Composite, which is a composite score of child intelligence (Aylward & Stancin, 2008;
Olds et al., 2014; Olds, Kitzman, et al., 2004). The four year follow-up of the Nurse
Family Partnership study in Memphis found significant differences in the Mental
Processing Component with a small effect size (p = .03; d = .18) (Olds, Kitzman, et al.,
2004). The four-year follow-up of the Nurse Family Partnership study in Denver did not have significant differences in the Mental Processing Component (Olds et al., 2014).
More specific assessments of cognitive performance, such as reading and arithmetic skills, were reported (Landry et al., 2014; Olds et al., 2014; Olds, Kitzman, et al., 2004), though no significant intervention effects were noted. Two studies included measurements of executive functioning (Lunkenheimer et al., 2008; Olds, Robinson, et al., 2004). Neither study found significant differences between the intervention and control groups in measures of executive functioning; however, Lunkenheimer et al.
(2008) reported a small indirect effect of the Family Check-Up intervention on child inhibitory control at age 4 through parental positive behavior support at age 3 (ß = .02).
110
Social development. Nine instruments were used to measure social outcomes
(Table 5). These outcomes were categorized into three domains: social interaction and competence, behavioral problems and regulation, and temperament (Table 3). Four of the instruments included parent or teacher reports of the child’s social or behavioral performance and five instruments were based on coding from direct assessments of the child performing specific behavior tasks. Reported and calculated effect sizes on all social outcomes were small (ranging from -.18 to .3).
Social interaction and competence. Four studies included social interaction and competence measures: three broad developmental focused interventions (Landry et al.,
2014; Love et al., 2005; Olds, Kitzman, et al., 2004) and one exclusively-focused on early language development (Goldfeld et al., 2011). Two of the broad developmental studies found significant differences in measures of social competence between the intervention groups compared to control; one was an intervention that focused on responsive teaching for childcare providers (Landry et al., 2014) and the other was Early
Head Start (Love et al., 2005). Landry et al. (2014) reported that age was associated with children's social competence levels. Love et al. (2005) found significant effects (d = 0.2) on an observational measure of parent-child interaction. The clinic-based early language development intervention and the nurse home visitation intervention, which measured outcomes 4 years after program participation, did not find significant differences in social competence (Goldfeld et al., 2011; Olds, Kitzman, et al., 2004).
Behavior problems and regulation. Six studies included measures of behavior problems and regulation using five instruments; all of these studies were broadly focused interventions (Landry et al., 2014; Love et al., 2005; Olds et al., 2014, 2002; Olds,
111
Kitzman, et al., 2004; Olds, Robinson, et al., 2004). Two of the studies found significant differences on behavioral outcomes. Landry et al. (2014) reported significant decreases in anger and aggression for children of teachers allocated to the responsive teaching intervention compared to control; however, they did not find significant decreases in the intervention that included a social-emotional component. They also did not find differences in anxiety and withdrawal between both intervention and control groups.
Love et al. (2005) found small, but significant effects (d = 0.11) in aggressive behavior.
The Nurse Family Partnership interventions did not report significant effects for measures of behavior and regulation problems (Olds et al., 2014, 2002; Olds, Kitzman, et al., 2004; Olds, Robinson, et al., 2004).
Temperament. One study (Olds et al., 2002) used the Bates Parent Perceptions of
Difficult Temperament to measure irritability of the child as perceived by the parent. The
Nurse Family Partnership intervention had no intervention effect on the parents’ perception of the child’s irritability.
Discussion
Addressing disparities in early childhood has the potential to build a solid foundation for children to reach their full potential in life (Robinson et al., 2017). There is currently a national spotlight in both the popular news media and federal policy/programming on the need to close the “language gap” (Rosenberg, 2013; Strauss,
2015; Talbot, 2015; U.S. Department of Health & Human Services Office of the
Administration for Children & Families, 2017; Walker, 2014). Numerous programs and initiatives have been funded through private philanthropy and federal, state, and local dollars to address this disparity (Clinton Foundation, n.d.; “Provid. Talks,” n.d.; Talk
112 with me Baby, n.d.; University of Chicago School of Medicine, n.d.; Warren, 2015). Yet, this review found that there are few studies that have employed rigorous designs to evaluate the effectiveness of these interventions on language or cognitive outcomes. For this review, numerous studies were excluded because they did not use rigorous study designs (n = 86), they only reported outcomes for the parent or did not report child cognitive or social outcomes (n = 7), or they were interventions conducted after the child entered preschool (n = 38). However, we found 15 rigorous studies that warranted this systematic review.
The interventions included in this review varied greatly in scope, modality, duration, intensity, and components. Fifty percent of the interventions were focused primarily on early language development, while the other half had a broader developmental scope. Four different intervention modalities were described including home visitation, clinic-based, child care center-based, and group-based. Both early language development-focused and broad developmental interventions showed weak to moderate evidence for positive impacts on development. The lowest intensity intervention did not find significant gains in language, cognitive, or social outcomes
(Goldfeld et al., 2011, 2012). The highest duration/intensity interventions had a broad developmental scope, recruited families prenatally, and included case management and social service referrals to respond to identified needs (Love et al., 2005; Olds et al., 2014,
2002; Olds, Kitzman, et al., 2004; Olds, Robinson, et al., 2004). The highest duration/intensity interventions did demonstrate small effects on language, cognitive, and social outcomes. While the effect sizes (both reported and calculated) across the studies were modest, it is clear from this review that no one intervention scope or modality is
113 consistently associated with improved language, cognitive, or social development.
However, evidence of significant findings suggests a promising direction on improving developmental outcomes.
There were multiple instruments that were used to assess language, cognitive, and social outcomes with only a few instruments being used across more than one study. The method of data collection also varied, with some using direct assessment and others using parent or teacher report. The heterogeneity of the interventions and assessment instruments, in addition to the variability of observed differences in outcomes between studies, suggests that further research is needed to determine the best instruments and data collection strategies to use.
Enriching the environments where children live, grow, learn, and play through comprehensive and integrated approaches is critical for healthy child development
(Buckingham, Beaman, & Wheldall, 2014; Cain & Van Horne, 2017; Neuman & Celano,
2001). Most of the studies included in this review focused on the parent and the home environment. Only one study focused on enhancing the responsive teaching of the childcare provider to foster cognitive and social development (Landry et al., 2014). While this is important, as children who are in center-based care spend a significant amount of time in that environment, not all children are exposed to these settings. Approximately
18% of children less than 5 years of age living in households at or below the federal poverty level are in center-based care as their primary child care arrangement (Federal
Interagency Forum on Child and Family Statistics, 2017). Therefore, the focus on the home environment may increase program reach. Numerous observational studies and reviews on early language development, including studies on low SES populations,
114 discuss the importance of enhancing not only the quantity, but also the quality of interaction between the parent or caretaker and the child (Cartmill et al., 2013; Rowe,
Pan, & Ayoub, 2005). Enriched language environments are described in terms of their nurturance and responsiveness toward the child, in addition to the lexical diversity of the language directed toward the child (Britto & Brooks-Gunn, 2001; Siddiqi, Irwin, &
Hertzman, 2007; Smith, Landry, & Swank, 2006). The broad developmental scope interventions in this review provided more emphasis on positive, responsive parenting/teaching and socioemotional development compared to the early language development focused interventions, which focused more on shared book reading and increasing child-directed speech.
It has been proposed that enhancing early language ability will support social development (Petersen et al., 2013). This review found promising evidence that interventions targeted to low SES populations can make small improvements in social development, however this was found in only a couple of the studies (Landry et al., 2014;
Love et al., 2005). This is similar to the finding in a review of center-based parenting programs which reported that few effects were noted on the preschool child’s socioemotional development (Brooks-Gunn & Markman, 2005). Few studies assessed social outcomes, especially those that exclusively focused on early language development.
Limitations
There are several limitations of this systematic review. The inclusion criteria excluded studies that may have methodological limitations and unpublished studies and
115 so there may be potential for publication bias, though most of the studies included reported null findings.
A limitation identified through the review are that most of the studies measured outcomes proximally post-intervention. One study that assessed effects both during and post-intervention noted that gains found during the intervention were not sustained post- intervention (Suskind et al., 2016). Four studies measured outcomes 6 months – 1-year post-intervention (Goldfeld et al., 2011, 2012; Lunkenheimer et al., 2008; McGillion et al., 2017) and three studies assessed intervention effects four to seven years post- intervention (Olds et al., 2014; Olds, Kitzman, et al., 2004; Olds, Robinson, et al., 2004).
These differences indicate the need to examine dosing and potential booster interventions to sustain intervention effects over time. Additionally, due to the state of the literature, this review assessed relatively few studies and those included were heterogeneous for both the interventions and the outcome measurements which limited the ability to assess for efficacy across study type and outcomes measured.
Implications and Recommendations
With the increasing movement in early childhood policy to bridge the “language gap,” researchers, program developers, practitioners, and policymakers need to understand what is efficacious in fostering the development of children living in low SES environments. This review points to the need for the evidence base on early language interventions targeted for low SES populations to be strengthened by utilizing more rigorous research designs and consistency of outcome measurement. Researchers should employ rigorous designs with equivalent comparison groups and standardized and validated outcome measures to confidently understand the effects of the interventions.
116
Long-term studies with frequent follow-ups should be conducted to assess the dosing and sustainability of the gains observed post-intervention. Longitudinal studies can inform program developers and practitioners whether booster doses should be implemented in order to sustain program effects. Furthermore, building on this research, comparative effectiveness and cost-benefit analyses should be conducted to assist program developers, practitioners, and policy makers with increased understanding of needed funding to scale interventions that are shown to be efficacious.
Conclusion
The purpose of this systematic review was to evaluate the efficacy of early language interventions for low SES populations on language, cognitive, and social outcomes. The studies showed modest intervention effects on language, cognitive, and social outcomes. The interventions identified in this review were multifaceted. The exclusively early language development focused interventions included provision of a developmentally appropriate children’s books, instruction, modeling or skills practice, and provided feedback to the parent or teacher. The interventions more broadly focused on child development provided education on developmental milestones and assessed the needs of the family with integrated case management, resource referrals, and navigation.
Several interventions show promising results; however, the heterogeneity of interventions and outcome measurements limit the comparability of the study results. This review points to a need for rigorous study designs and replication of early language interventions for low SES populations to assess the efficacy on language, cognitive, and social outcomes. Determining interventions that improve child language, cognitive, and social outcomes in early childhood has the potential to improve population health.
117
Acknowledgements
Work on this manuscript was supported by a doctoral scholarship from the Robert Wood
Johnson Foundation Future of Nursing Scholars Program and the University of Texas
Health Science Center Cizik School of Nursing received by Cary M. Cain.
118
References
Aylward, G. P., & Stancin, T. (2008). Screening and Assessment Tools. In Mark L.
Wolraich Dennis D. Drotar & Ellen C. Perrin (Eds.), Developmental-Behavioral
Pediatrics Evidence and Practice (pp. 123–201). Philadelphia, PA: Mosby.
Ayoub, C., Vallotton, C. D., & Mastergeorge, A. M. (2011). Developmental pathways to
integrated social skills: The roles of parenting and early intervention. Child
Development, 82(2), 583–600.
Bradley, R. H., & Corwyn, R. F. (2002). Socioeconomic status and child development.
Annual Review of Psychology, 53(1), 371–399.
Britto, P. R., & Brooks-Gunn, J. (2001). Beyond shared book reading: dimensions of
home literacy and low-income African American preschoolers’ skills. New
Directions for Child and Adolescent Development, 92(92), 73-89.
Brooks-Gunn, J., & Markman, L. B. (2005). The contribution of parenting to ethnic and
racial gaps in school readiness. Future of Children, 15(1), 139-68.
Brownell, M. D., Ekuma, O., Nickel, N. C., Chartier, M., Koseva, I., & Santos, R. G.
(2016). A population-based analysis of factors that predict early language and
cognitive development. Early Childhood Research Quarterly, 35(2), 6-18.
Buckingham, J., Beaman, R., & Wheldall, K. (2014). Why poor children are more likely
to become poor readers: the early years. Educational Review, 66(4), 428–446.
Cabrera, N. J., Karberg, E., Malin, J. L., & Aldoney, D. (2017). The magic of play: Low-
income mothers’ and fathers’ playfulness and children’s emotion regulation and
vocabulary skills. Infant Mental Health Journal, 38(6), 757–771.
Cain, C. M., & Van Horne, B. S. (2017). Current challenges in addressing the realities of
119
poverty and inequality in families with young children: A call for both policy and
intervention. APSAC Advisor, 29(2), 3–13.
Camilli, G., Vargas, S., Ryan, S., & Barnett, W. S. (2010). Meta-analysis of the effects of
early education interventions on cognitive and social development. Teachers
College Record, 112(3), 579–620.
Cartmill, E. A., III, B. F. A., Gleitman, L. R., Goldin-Meadow, S., Medina, T. N., &
Trueswell, J. C. (n.d.). Quality of early parent input predicts child vocabulary 3
years later. Proceedings of the National Academy of Sciences, 110(28), 11278–
11283.
Chapin, L. A., & Altenhofen, S. (2010). Neurocognitive perspectives in language
outcomes of early head start: Language and cognitive stimulation and maternal
depression. Infant Mental Health Journal, 31(5), 486–498.
Clinton Foundation. (n.d.). Too Small To Fail. Retrieved February 22, 2018, from
http://toosmall.org/
Cronan, T. A., Cruz, S. G., Arriaga, R. I., & Sarkin, A. J. (1996). The effects of a
community-based literacy program on young children’s language and conceptual
development. American Journal of Community Psychology, 24(2), 251-272.
Doyle, O., Harmon, C. P., Heckman, J. J., & Tremblay, R. E. (2009). Investing in early
human development: timing and economic efficiency. Economics and Human
Biology, 7(1), 1–6.
Dunn, L. M., & Dunn, L. M. (1997). Picture Vocabulary Test—Third Edition. Circle
Pines, MN: American Guidance Service.
Durham, R. E., Farkas, G., Hammer, C. S., Bruce Tomblin, J., & Catts, H. W. (2007).
120
Kindergarten oral language skill: A key variable in the intergenerational
transmission of socioeconomic status. Research in Social Stratification and
Mobility, 25(4), 294–305.
Federal Interagency Forum on Child and Family Statistics. (2017). America’s Children:
Key National Indicators of Well-Being, 2017. Retrieved February 22, 2018, from
https://www.childstats.gov/americaschildren/family3.asp
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.
Fernald, A., Marchman, V. A., & Weisleder, A. (2013). SES differences in language
processing skill and vocabulary are evident at 18 months. Developmental Science,
16, 234–248.
Gilmore, J. H., Lin, W., Prastawa, M. W., Looney, C. B., Vetsa, Y. S. K., Knickmeyer,
R. C., … Gerig, G. (2007). Regional gray matter growth, sexual dimorphism, and
cerebral asymmetry in the neonatal brain. The Journal of Neuroscience, 27(6),
1255–1260.
Gleason, J. B., & Ratner, N. B. (2009). The Development of Language. Boston, MA:
Pearson.
Goldfeld, S., Napiza, N., Quach, J., Reilly, S., Ukoumunne, O. C., & Wake, M.
(2011). Outcomes of a universal shared reading intervention by 2 years of age:
The Let’s Read trial. Pediatrics, 127(3), 445–453.
Goldfeld, S., Quach, J., Nicholls, R., Reilly, S., Ukoumunne, O. C., & Wake, M. (2012).
121
Four-year-old outcomes of a universal infant-toddler shared reading intervention:
The Let’s Read trial. Archives of Pediatrics and Adolescent Medicine, 166(11),
1045–1052.
Hart, B., & Risley, T. R. (1995). Meaningful Differences in the Everyday Experience of
Young American Children. Baltimore, MD: Paul H. Brookes Publishing Co.
Hart, B., & Risley, T. R. (2003). The early catastrophe: The 30-million-word gap by age
three. American Educator, 4–9.
Heckman, J. J. (2006). Skill formation and the economics of investing in disadvantaged
children. Science, 312(5782), 1900-1902.
Higgins, J. P. T., Altman, D. G., Gøtzsche, P. C., Jüni, P., Moher, D., Oxman, A. D., …
Cochrane Statistical Methods Group. (2011). The Cochrane Collaboration’s tool
for assessing risk of bias in randomised trials. British Medical Journal, 343,
d5928.
Higgins, J. P. T., & Green, S. (2011). Cochrane Handbook for Systematic Reviews of
Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration.
High, P., LaGasse, L., Becker, S., Ahlgren, I., & Gardner, A. (2000). Literacy promotion
in primary care pediatrics: Can we make a difference? Pediatrics, 105(4 Pt 2),
927–934.
Hoff, E. (2003). The specificity of environmental influence: Socioeconomic status affects
early vocabulary development via maternal speech. Child Development, 74(5),
1368–1378.
Hurt, H., & Betancourt, L. M. (2016). Effect of socioeconomic status disparity on child
122
language and neural outcome: how early is early? Pediatric Research, 79(1–2),
148–158.
Hurt, H., & Betancourt, L. M. (2017). Turning 1 year of age in a low socioeconomic
environment: A portrait of disadvantage. Journal of Developmental & Behavioral
Pediatrics, 38(7), 493-500.
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.
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.
Landry, S. H., Zucker, T. A., Taylor, H. B., Swank, P. R., Williams, J. M., Assel, M., …
Klein, A. (2014). Enhancing early child care quality and learning for toddlers at
risk: The responsive early childhood program. Developmental Psychology, 50(2),
526–541.
Law, J., & Roy, P. (2008). Parental report of infant language skills: A review of the
development and application of the Communicative Development Inventories.
Child and Adolescent Mental Health, 13(4), 198–206.
Lonigan, C. J., Wagner, R. K., Torgesen, J. K., & Rashotte, C. A. (2002). The Preschool
Comprehensive Test of Phonological & Print Processing. Retrieved November
12, 2018, from https://psy.fsu.edu/~lonigan/01234Pre-CTOPPP.htm
Love, J. M., Kisker, E. E., Ross, C., Raikes, H., Constantine, J., Boller, K., … Banks
123
Tarullo, L. (2005). The effectiveness of Early Head Start for 3-year-old children
and their parents: Lessons for policy and programs. Developmental Psychology,
41(6), 885–901.
Lowe, J. R., Erickson, S. J., Schrader, R., & Duncan, A. F. (2012). Comparison of the
Bayley II mental developmental index and the Bayley III cognitive scale: Are we
measuring the same thing? Acta Paediatrica, 101(2), e55-58.
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.
MacLeod, J., & Nelson, G. (2000). Programs for the promotion of family wellness and
the prevention of child maltreatment: a meta-analytic review. Child Abuse &
Neglect, 24(9), 1127–1149.
Marulis, L. M., & Neuman, S. B. (2013). How Vocabulary Interventions Affect Young
Children at Risk: A Meta-Analytic Review. Journal of Research on Educational
Effectiveness, 6(3), 223–262.
Matte-Gagné, C., & Bernier, A. (2011). Prospective relations between maternal
autonomy suport and child executive functioning: Investigating the mediating role
of child language ability. Journal of Experimental Child Psychology, 110(4),
611–625.
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
language development in infants from diverse socioeconomic status backgrounds.
124
Journal of Child Psychology and Psychiatry and Allied Disciplines, 58(10),
1122–1131.
Mendelsohn, A. L., Dreyer, B. P., Flynn, V., Tomopoulos, S., Rovira, I., Tineo, W., …
Nixon, A. F. (2005). Use of videotaped interactions during pediatric well-child
care to promote child development: A randomized, controlled trial. Journal of
Developmental and Behavioral Pediatrics, 26(1), 34–41.
Moher, D., Liberati, A., Tetzlaff, J., Altman, D., & The PRISMA Group. (2009).
Preferred reporting items for systematic reviews and meta-analyses: The PRISMA
statement. PLoS Med, 6(8), e1000097.
Needlman, R., & Silverstein, M. (2004). Pediatric interventions to support reading aloud:
How good is the evidence? Journal of Developmental and Behavioral Pediatrics,
25(5), 352-63.
Nelson, C. (2000). Neural plasticity and human development: The role of early
experience in sculpting memory systems. Developmental Science, 3(2), 115–136.
Neuman, S. B., & Celano, D. (2001). Access to Print in Low-Income and Middle-Income
Communities: An Ecological Study of Four Neighborhoods. Reading Research
Quarterly, 36(1), 8–26.
Nielsen, R. (2003). Sutherland Phonological Awareness Test - revised (SPAT-R) - ACER
Shop Online. Retrieved November 12, 2018, from https://shop.acer.edu.au/acer-
shop/group/SP
Nowakowski, R. S. (2006). Stable neuron numbers from cradle to grave. Proceedings of
the National Academy of Sciences, 103(33), 12219–12220.
Olds, D. L., Holmberg, J. R., Donelan-McCall, N., Luckey, D. W., Knudtson, M. D., &
125
Robinson, J. (2014). Effects of home visits by paraprofessionals and by nurses on
children follow-up of a randomized trial at ages 6 and 9 years. JAMA Pediatrics,
168(2), 114-121.
Olds, D. L., Kitzman, H., Cole, R., Robinson, J., Sidora, K., Luckey, D. W., …
Holmberg, J. (2004). Effects of nurse home-visiting on maternal life course and
child development: Age 6 follow-up results of a randomized trial. Pediatrics,
114(6), 1550–1559.
Olds, D. L., Robinson, J., O’Brien, R., Luckey, D. W., Pettitt, L. M., Henderson, C. R.,
… Talmi, A. (2002). Home visiting by paraprofessionals and by nurses: A
randomized, controlled trial. Pediatrics, 110(3), 486–496.
Olds, D. L., Robinson, J., Pettitt, L., Luckey, D. W., Holmberg, J., Ng, R. K., …
Henderson, C. R. (2004). Effects of home visits by paraprofessionals and by
nurses: age 4 follow-up results of a randomized trial. Pediatrics, 114(6), 1560-
1568.
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.
ProQuest. (2016). RefWorks. Retrieved from https://refworks.proquest.com
Providence Talks. (n.d.). Retrieved February 22, 2018, from
http://www.providencetalks.org/
Raikes, H., Luze, G., Brooks-Gunn, J., Abigail Raikes, H., Alexander Pan, B., Tamis-
LeMonda, C. S., … Rodriguez, E. T. (2006). Mother-child bookreading in low-
126
income families: Correlates and outcomes during the first three years of life.
Child Development, 77(4), 924–953.
Ramey, C. T., & Ramey, S. L. (1998). Early intervention and early experience. American
Psychologist, 53(2), 109–120.
Roben, C. K. P., Cole, P. M., & Armstrong, L. M. (2012). Longitudinal relations among
language skills, anger expression, and regulatory strategies in early childhood.
Child Development, 84(3), 891–905.
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.
Robinson, L. R., Bitsko, R. H., Thompson, R. A., Dworkin, P. H., McCabe, M. A.,
Peacock, G., & Thorpe, P. G. (2017). CDC Grand Rounds: Addressing health
disparities in early childhood. MMWR. Morbidity and Mortality Weekly Report,
66(29), 769–772.
Rosenberg, T. (2013, April 10). The Power of Talking to Your Baby. New York Times.
Retrieved February 22, 2018, from
https://opinionator.blogs.nytimes.com/2013/04/10/the-power-of-talking-to-your-
baby/
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.
Rowe, M. L. (2018). Understanding socioeconomic differences in parents’ speech to
children. Child Development, 12(2), 122–127.
127
Rowe, M. L., Pan, B. A., & Ayoub, C. (2005). Predictors of variation in maternal talk to
children: A longitudinal study of low-income families. Parenting, 5(3), 259–283.
Senechal, M., & Young, L. (2008). The effect of family literacy interventions on
children’s acquisition of reading from kindergarten to grade 3: A meta-analytic
review. Review of Educational Research, 78(4), 880–907.
Shonkoff, J. P., & Phillips, D. A. (2000). From Neurons to Neighborhoods: The Science
of Early Childhood Development. (J. P. Shonkoff & D. A. Phillips, Eds.),
National Academy of Sciences (Vol. 23). Washington, DC: National Academy Press.
Siddiqi, A., Irwin, L. G., & Hertzman, C. (2007). Total environment assessment model
for early child development: Evidence report. Vancouver, BC. Retrieved from
https://pdfs.semanticscholar.org/722d/
Smith, K. E., Landry, S. H., & Swank, P. R. (2006). The role of early maternal
responsiveness in supporting school-aged cognitive development for children who
vary in birth status. Pediatrics, 117(5), 1608-1617.
Sperry, D. E., Sperry, L. L., & Miller, P. J. (2018). Reexamining the verbal environments
of children from different socioeconomic backgrounds. Child Development, 0(0),
1–16.
Spieker, S. J., Nelson, D. C., Petras, A., Jolley, S. N., & Barnard, K. E. (2003). Joint
influence of child care and infant attachment security for cognitive and language
outcomes of low-income toddlers. Infant Behavior and Development, 26(3), 326–
344.
Strauss, V. (2015). Why parents should talk a lot to their young kids — and choose their
128
words carefully. Retrieved February 22, 2018, from
https://www.washingtonpost.com/news/answer-sheet/wp/2015/09/30/why-
parents-should-talk-a-lot-to-their-young-kids-and-choose-their-words-
carefully/?utm_term=.d2f8e7f1c3b6
Suskind, D. L., Leffel, K. R., Graf, E., Hernandez, M. W., Gunderson, E. A., Sapolich,
S.G., … Levine, S. C. (2016). A parent-directed language intervention for
children of low socioeconomic status: A randomized controlled pilot study.
Journal of Child Language, 43(2), 366–406.
Talbot, M. (2015). Annals of Education: The Talking Cure. Retrieved February 22, 2018,
from https://www.newyorker.com/magazine/2015/01/12/talking-cure
Talk with me Baby. (n.d.). Talk With Me Baby. Retrieved February 22, 2018, from
http://www.talkwithmebaby.org/
Te Kaat-van den Os, D. J. A., Jongmans, M. J., Volman, M. J. M., & Lauteslager,
P. E. M. (2017). Parent-Implemented Language Interventions for Children with a
Developmental Delay: A Systematic Review. Journal of Policy and Practice in
Intellectual Disabilities, 14(2), 129–137.
Thalheimer, W., & Cook, S. (2002). How to calculate effect sizes from published
research: A simplified methodology. Work-Learning Research. Retrieved October
16, 2018, from http://work-learning.com/effect_sizes.htm.
U.S. Department of Health & Human Services Office of the Administration for Children
& Families. (2017). Bridging The Word Gap. Retrieved February 22, 2018, from
https://www.acf.hhs.gov/ecd/child-health-development/bridging-the-word-gap
129
University of Chicago School of Medicine. (n.d.). Thirty Million Words Initiative.
Retrieved February 22, 2018, from http://thirtymillionwords.org/
Vallotton, C. D., Harewood, T., Ayoub, C. A., Pan, B., Mastergeorge, A. M., & Brophy-
Herb, H. (2012). Buffering boys and boosting girls: The protective and promotive
effects of Early Head Start for children’s expressive language in the context of
parenting stress. Early Childhood Research Quarterly, 27(4), 695–707.
Vally, Z., Murray, L., Tomlinson, M., & Cooper, P. J. (2015). The impact of dialogic
book-sharing training on infant language and attention: a randomized controlled
trial in a deprived South African community. Journal of Child Psychology and
Psychiatry and Allied Disciplines, 56(8), 865–873.
Walker, R. (2014). How to bridge a 30-million-word gap. Retrieved February 22, 2018,
from https://www.csmonitor.com/The-Culture/Verbal-Energy/2014/1106/How-to-
bridge-a-30-million-word-gap
Warren, S. F. (2015). Right from birth: Eliminating the talk gap in young children.
Boulder, CO. Retrieved from http://www.lena.org/wp-
content/uploads/2016/10/RightFromBirth_Warren_5.12.2015_v.3.pdf
Weisleder, A., & Fernald, A. (2013). Talking to children matters: Early language
experience strengthens processing and builds vocabulary. Psychological Science,
24(11), 2143–2152.
Wiig, E. H., Secord, W. A., & Semel, E. (2004). Clinical evaluation of language
fundamentals— Preschool, second edition (CELF Preschool-2). Toronto, Canada:
The Psychological Corporation/A Harcourt Assessment Company.
Zimmerman, I. L., Steiner, V. G., & Pond, R. E. (2002). Preschool Language Scale,
130
Fourth Edition. San Antonio, TX: Psychological Corporation.
131
Table 1
Systematic Review Inclusion and Exclusion Criteria Included Excluded Population: Infants and children between 0-3 years Population: Study child with disability (developmental (human) disabilities such as autism, child with hearing loss, or Low socioeconomic status population premature birth) Age range older than 3 years of age; age range not included in manuscript Design: Randomized controlled trial or cluster Design: Not original research; non-randomized design randomized controlled trial Intervention: Intervention component Intervention: Intervention for a language other than the child's primary language (i.e. second language interventions) Outcome: At least one language, cognitive, or social Publication: Non-English language outcome measured and reported on the study child Published prior to January 1995 or after October 2018 Grey literature, dissertation, thesis, conference proceedings, or if full-text not available Publication: Published in peer reviewed literature
131
132
Table 2
Early Language Development Intervention Study Characteristics Authors, Country Population Child Program Delivered Delivered Intervention Control Intervention Sample size Child Child Race/ Date Age Name by to Condition Condition Duration by condition: Gender Ethnicity Range Baseline (Follow-Up) Cronan, United Low SES: 1 – 3 Project Trained Parent Home Visitation No One academic High Total Total Cruz, and States Qualified years PRIMER Tutor High intensity: 18 instruction year intensity: 83 Female: 104 Hispanic/ Arriaga, for Head (University 30-minute sessions High (NR) Male: 121 Latino: 94 1996 Start psychology (once a week) intensity: 18 Low intensity: White: 52 students) Instruction visits 73 (NR) High Black: 30 Modeling Low intensity: C: 69 (NR) intensity/ Low Other: 10 Return 3 visits intensity/C Demonstration Female: High intensity/ Books 41/30/33 Low intensity/C Instructional Male: Hispanic/ Materials 42/43/36 Latino: 52/43/38 Low intensity: 3 30- White: 18/15/19 minute instructional Black: 9/12/9 sessions (similar Other: 4/3/3 content) Goldfeld, Australia Low SES: 1 – 18 Let’s Nurse Parent Clinic-based Usual care Approx. 1.5 I: 365 (324) Total NR Napiza, Bottom months Read During 3 well child (tip sheet) years C: 265 (228) Female: 302 Quach, tertile on visits (4-8, 12, and 3 well-child Male: 328 Reilly, SES Indexes 18 months), a nurse visits Ukoumu of Areas spent approximately I/C nne, and Index of 5 minutes with Female: Wake, Disadvantag families to model 142/140 2011 e and discuss Male: intervention 203/125 messages, take home packs with a free age appropriate book and guidance messages. (Follow- Australia Low SES: 1 month Let’s Nurse Parent Clinic-based Same as Approx. 3.5 I: 365 (328) Total NR up from Bottom – 3 years Read Same as Goldfeld et Goldfeld et years C: 265 (235) Female: 302 Goldfeld tertile on al., 2011 al., 2011 4 well-child Male: 328 et al., SES Indexes visits 2011) of Areas I/C Goldfeld, Index of Female: Quach, Disadvantag 142/140 Nicholls, e Male: Reilly, 203/125 Ukoumu
nne, and 132
133
Wake, 2012 High, United Low SES: 5 – 11 NR Pediatrician Parent Clinic-based Routine Approx.one I: 106 (75) Total Total LaGasse, States Targeted months Pediatricians gave a pediatric year C: 99 (75) Female: 69 Hispanic/Latino: Becker, health children’s book, a care with M=3.38 well- Male: 84 72 Ahlgren, centers that handout explaining anticipator child visits White: 28 and serve low the benefits of y guidance I/C Other: 53 Gardner, income reading to children, Female: 35/34 2000 populations and literacy Male: 42/42 I/C promoting Hispanic/Latino: anticipatory 41/31 guidance. White: 16/12 Other: 20/33 Landry, United Low SES: 2-3 years Responsi Coach Child Care Center-based Usual One academic Total: 542 Total Total Zucker, States Childcare ve Early Staff Teachers were Instruction year (392) Female: 276 Hispanic/ Taylor, Centers Childhoo coached by trained RECC: 900 Male: 266 Latino: 73 Swank, where at d staff activities RECC: NR White: 37 Williams, least 50% Curriculu RECC: Responsive RECC Plus: RECC Plus: Black: 423 Assel, . . children m teaching practices: 720 activities NR Other: 9 . Klein, receive (RECC) developmentally C: NR 2014 federal appropriate assistance cognitive readiness (Federal activities, childcare interactive book subsidy or reading, and early free/reduced phonological price lunch) awareness RECC Plus: Addition of social emotional training and social emotional curriculum Love, United Low SES: 0 -12 Early Program Parent Home Visitation Did not Average I: 1513 (936) Total Total Kisker, States Qualified months Head staff Center-Based participate enrollment C: 1488 (833) Female: 1,469 Hispanic/Latino: Ross, for Early Start (teachers Mixed-Approach in Early 20-23 months Male: 1,532 708 Constanti Head Start and/or home (Home Visitation Head Start Prenatal - 3 White: 1,116 ne, visitors) & Center-Based) Program years of age I/C Black: 1,039 Boller, Female: Other: 138 Chazan- 4 Center-based 731/738 Cohen, . . programs with Male: I/C ., Vogel, parenting education 782/750 Hispanic/Latino: 2005 and at least 2 home 360/348 visits each year to White: 564/552 each family Black: 518/521 7 Home-based Other: 71/67 programs with
weekly home visits 133
and at least 2
134
parent-child social activities per month for each family 6 Mixed-approach programs with home-based and center-based services Content: Parenting education, case management, health care provision and referrals, and family support Lunkenh United Low SES 2 -3 years Family Parent Parent Home Visitation Participati 6 home visits Total: 731 Total Total eimer, States and risk Check consultant Meeting with a on in (3 at age 2 (619) Female: 358 White: 366 Shaw, factor: Up parenting assessment and 3 at age I: NR Male: 373 Black: 204 Gardner, Qualify for consultant. 3); 2 years C: NR Other: 161 Dishion, WIC (low Assessment, Hispanic 13.4% Connell, education interview session, and achievement and feedback Wilson, and low session. Interview 2008 family session focuses on income); family issues; Child feedback session behavior explores parent problems, or willingness to Family change parenting problems practices, support existing parenting strengths, and identify services appropriate for the family's needs. McGillio United High and Range NR Study staff Parent Home Visitation Dental 1 month I: 72 (69) Total NR n, Pine, Kingdom low SES: NR Video on health C: 70 (59) Female: 73 Herbert, Indices of (Mean contingent talk interventio Male: 69 and Deprivation, age 11 Daily practice n - home- Matthew primary months) Diary visitation s, 2017 caregiver visits education, where and annual home income visitors showed video on healthy eating and
tooth-
134
135
brushing practices. Mendels United Low SES: 0-3 years Video Child Parent Clinic-based Standard Approximatel I: 77 (54) NR 100% Hispanic/ ohn, States Family Interactio Developmen Child development pediatric y 3 years C: 73 (46) Latino Dreyer, Hollingshea n Project t Specialist specialist meets care 12 visits Flynn, d Four- with parents and Tomopou Factor SES - children during los, Parental routine well-child Rovira, education visits. Tineo, . . and Content: Child . Nixon, occupation development 2005 discussion, age- appropriate learning material (book or toy), record a videotape of the parent-child interaction and review the tape with the specialist Olds, United Low SES: 0 – 24 Nurse Nurse/ Parent Home Visitation Developm Approx. 2 Nurse Home NR Total Robinson States Qualified months Family Paraprofessi Nurse Home ental years Visitor: 235 Hispanic/Latino: , O'Brien, for Partnersh onal Visitor: screening Prenatal to 24 (168) 330 Luckey, Medicaid or ip Developmental and months of age Paraprofessio White: 262 Pettitt, no health screening and referral nal Home Black: 121 Henderso insurance referral to services, services Visitor: 245 Other: 22 n, . . . status home visitation by a for their (188) Talmi, registered nurse. children at C:255 (204) Nurse Home 2002 Paraprofessional 6, 12, 15, Visitor/ Home Visitor: 21, and 24 Paraprofessional Developmental months Home Visitor/C screening and old. Hispanic/Latino: referral to services, 103/110/117 home visitation by a White: 87/86/89 paraprofessional. Black: 38/42/41 Other: 7/7/8 (Follow- United Low SES: 0 – 24 Nurse Nurse/ Parent Home Visitation Same as Approx. 2 Nurse Home NR Nurse Home up from States Qualified months Family Paraprofessi Same as Olds et al., Olds et al., years Visitor: 235 Visitor/ Olds et for Partnersh onal 2002 2002 Prenatal to 24 (196) Paraprofessional al., 2002) Medicaid or ip months of age Paraprofessio /C Olds, no health nal Home Mexican Robinson insurance Visitor: 245 American:100/9 , O'Brien, status (198) 1/105 Luckey, C:255 (211) Black: 35/31/31 Holmber g, Ng, . . .
Henderso 135
n, 2004
136
(Follow- United Low SES: 0 – 24 Nurse Nurse/ Parent Home Visitation Same as Approx. 2 NR NR NR up from States Qualified months Family Paraprofessi Same as Olds et al., Olds et al., years Olds et for Partnersh onal 2002 2002 Prenatal to 24 al., 2002) Medicaid or ip months of age Olds, no health Holmber insurance g, status Donelan- McCall, Luckey, Knudston , and Robinson , 2014 Olds, United Low SES: 0 – 24 Nurse Nurse Parent Home Visitation Transporta Approx. 2 I: 228 (197) NR 92% Black Kitzman, States Unmarried, months Family Transportation for tion for years C: 515 (444) Cole, < 12 years Partnersh prenatal prenatal Prenatal - 24 Robinson education, ip appointments, appointme months of age , Sidora, and/or Nurse home visiting nts Luckey, . unemployed services, Developm . . Developmental ental Holmber screening and screening g, 2004 referral services for and child at 6, 12, and referral 24 months of age services for child at 6, 12, and 24 months of age Suskind, United Low SES: 1 – 3 Parent- Trained Parent Home Visitation Nutrition 8 weekly I: 18 (12) Total Total Leffel, States Mothers' months directed home visitor Trained coach interventio hour-long C: 19 (11) Female: 9 White: 3 Graf, eligibility Language worked with parent n - 8 visits Male: 14 Black: 20 Hernande for Interventi through 8 modules weekly z, Medicaid on to provide strategies home I/C I/C Gunderso and/or WIC to enrich the home visits from Female: 5/4 White: 2/1 n, language a research Male: 7/7 Black: 10/10 Sapolich, environment. assistant . . . Including an for 5-10 Levine, interactive minutes. 2016 component, behavior feedback component, skills practice, and goal- setting. Provided with children's book at the end of each
home visit.
136
137
Vally, South Low SES: 14 – 16 NR Trained Parent Group sessions NR 8 weeks I: 49 (45) Total NR Murray, Africa endemic months community Book sharing 8 sessions C: 42 (37) Female: 32 Tomlinso poverty, (I: 15.45 member training program Male: 59 n, and mass months Weekly 90-minute Cooper, unemploym C: 15.29 training sessions I/C 2015 ent, and months) delivered in groups Female: 16/16 rampant Male: 33/26 crime
Note. Abbreviations: I: Intervention; C: Control; RCT: Randomized Controlled Trial; NR: Not Reported
137
138
Table 3
Language Outcomes Author, Date Follow-Up Language Language Method of Intervening or Data on Data on Finding on Effect Size Results Period Outcome Outcome Analysis Control Outcome Outcome for Differences Assessment Variables for Control Intervention Instrument Group: Group: Baseline Baseline (Follow- (Follow-Up) Up) Cronan et al., Postintervention Language PRIMER ANCOVA Baseline (26.6) High intensity: High intensity: p High intensity: Significant 1996 comprehension Language Scores (29.0) = .024 .06* findings Comprehension Low intensity: Low intensity: p Low intensity: between high Book (Direct (25.9) = .877 0* intensity Assessment) intervention and Conceptual Bracken Basic (29.4) High intensity: High intensity: p High intensity: control on all knowledge Concept Scale (31.5) = .056 .05* language (Direct Low intensity: Low intensity: p Low intensity: outcomes. Low Assessment) (19.9) = .507 .02* intensity Language MacArthur- (39.4) High intensity: High intensity: High intensity: intervention did production Bates (52.5) p = .009 .07* not differ from Communicative Low intensity: Low intensity Low intensity: control. Development (38.9) p=.581 .02* Inventory (Parent Report) Goldfeld et 6 months Expressive MacArthur- Linear Parental mental (53.9) (51.1) NS (p = .36) .11* No significant al., 2011 postintervention; vocabulary Bates regression health, child intervention at 2 years of age Communicative models gender, effects noted on Development English main language Inventory spoken outcomes. (Parent Report) language in the Communication Communication home, primary (104.5) (103.6) NS (p = .87) .06* Expressive and Symbolic caregiver's (12.6) (12.7) NS (p = .58) .02* Speech Behavior Scales education Symbolic Developmental level, health (12.6) (12.5) NS (p = .77) .02* (understanding Profile Infant- care status, of words and use Toddler unemployment, of objects) Checklist and local (Parent Report) government area (Follow-up 6 months Literacy - Sutherland Linear Parental mental (6.8) (7.0) NS (p = .29) .08* No significant assessment of postintervention; Intrasyllabic Phonological regression health, child intervention outcomes) at 4 years of age Literacy - Awareness Test models gender, (5.4) (5.4) NS (p = .85) 0* effects noted on Phonemic Revised (Direct English main language Goldfeld et Literacy - Letter Assessment) spoken (9.6) (9.3) NS (p = .92) .03* outcomes. al., 2012 awareness language in the 138
Core language home, primary (97.8) (99.0) NS (p = .25) .07*
139
Receptive Clinical caregiver's (94.7) (95.1) NS (p = .56) .02* language Evaluation of education Expressive Language level, health (98.4) (99.1) NS (p = .32) .04* language Fundamentals care status, Preschool - unemployment, Second Edition, and local (Direct government Assessment) area High et al., A month after Receptive MacArthur- T-Test & (39.3) (51) p = .004 4.10* Significant 2000 child completed vocabulary Bates Hierarchical differences 3 well child Expressive Communication Linear (15.9) (22.1) NS (p = .11) 2.27* noted for visits or when vocabulary and Regression intervention the child turned Development Models group in 22 months old Inventories receptive (Parent Report) vocabulary. Regression analyses found expressive and receptive vocabulary were mediated through enjoyment of shared reading in intervention group. Landry et al., Postintervention Expressive Expressive One- Mixed Baseline 17.55 RECC: 17.69 NS NR** No significant 2014 Vocabulary Word Picture Model condition (23.80) (24.71) group Vocabulary Test Analyses RECC Plus: differences were (Direct 15.60 (22.97) found for Assessment) growth of Receptive Preschool 4.04 (8.73) RECC: 3.03 NS NR** vocabulary, Vocabulary Language Scale (7.40) receptive (Direct RECC Plus: language, or Assessment) 3.06 (6.67) early literacy Early Literacy Preschool 34.69 RECC: 35.05 NS NR** Comprehensive (40.56) (41.06) Test of RECC Plus: Phonological 33.28 (39.29) and Print Processing (Direct Assessment) Love et al., Near intervention Receptive Peabody Picture Regression Baseline (81.1) (83.3) p = .02 .13 Mixed-approach 2005 completion; at vocabulary Vocabulary Test models condition showed stronger approximately 36 Receptive (Direct Hierarchical (81.8) (83.2) NS .09 effects for months of age vocabulary (for Assessment) Linear receptive vocabulary.
the center-based Models 139
program)
140
Receptive (78.5) (82.2) p = .04 .23 vocabulary (for the mixed- approach program) Receptive (83.1) (84.6) NS .09 vocabulary (for the home-based program) Lunkenheimer Postintervention; Language skills Fluharty-2 Repeated 80.82 76.65 (87.74) NS NR** Direct effect of et al., 2008 at 4 years of age Preschool Measures (88.09) intervention on Speech and ANOVA child language Language at age 4 not Screening Test significant. (Direct Indirect effect of Assessment) intervention on child language at age 4 through parent’s positive behavior support at age 3 significant (ß = .03 - modest). McGillion et 1-year Expressive MacArthur- T-Test and SES, baseline (370.67) (348.67) p < .01 -.17 T-tests al., 2017 postintervention vocabulary Bates Linear outcomes comparing (Child 24 months Communicative regression intervention and of age) Development models control (without Inventory controlling for (Parent Report) SES): No 6 months Child Language (293.38) (296.38) p < .05 .32 intervention postintervention vocalizations Environment effect on (Child 18 months Analysis outcome of age) (LENA) Digital variables at 24 Language months. Processor Effects of the (Observational) intervention did not last 1 year after. Low SES population: significantly higher intervention effects on expressive vocabulary at 18 months. Mendelsohn Expressive Preschool T-Test (80.0) (78.5) NS (p = .58) .15* For children of
et al., 2005w language (for Language Scale higher educated 140
141
Postintervention; participants with - 3 (Direct mothers (7-11 At 21 months of maternal Assessment) grade age education <7 education), years) intervention was Expressive (76.1) (83.3) p = .008 1.05* associated with language (for increases in participants with expressive maternal language. education >7 For children of years) lower educated Receptive (79.7) (80.6) NS (p = .72) .10* mothers (6th language (for grade or less), participants with no differences maternal noted between education <7 intervention and years) control. Receptive (78.9) (82.5) NS (p = .25) .42* language (for participants with maternal education >7 years) Olds et al., Postintervention Child language Preschool General Maternal age, (99.49) (Nurse Home NS NR*** Nurse visited 2002 Language Scale Linear housing Visitor: 101.22) group less likely - 3 (Direct Model density, (Paraprofession to exhibit Assessment) gestation, al Home language delays maternal Visitor: 99.89) than children in Language delay conflict with (12) (Nurse Home Nurse Home NR*** the control her partner, Visitor: 6) Visitor: p = .05 group. No maternal (Paraprofession Paraprofessional: significant conflict with al Home NS effects on other her mother Visitor: 11) outcomes. (Follow-up 2 years Child Language Preschool General Maternal (92.01) (Nurse Home NS Nurse Home No significant assessment of postintervention Language Scale Linear psychologic Visitor: 92.65) Visitor: .04 intervention outcomes (Child 4 years of - 3 (Direct Model resources, age (Paraprofession Paraprofessional: effects noted on from Olds et age) Assessment) of gestation, al Home .08 language al., 2002) maternal age, Visitor: 93.24) outcomes for housing intervention Olds, density, groups. Robinson, et conflict with al., 2004 partner (Follow-up 4 years Receptive Peabody Picture General Maternal (90.56) (Nurse Home NS Nurse Home No significant assessment of postintervention Language Vocabulary Test Linear psychological Visitor: 93.31) Visitor: .21 intervention outcomes (Child 6 years of (Direct Model resource index, (Paraprofession Paraprofessional: effects noted for from Olds et age) Assessment) smoking status, al Home .16 receptive al. 2002) age of Visitor: 92.59) language for gestation, intervention
Olds et al., housing groups. 141
2014 density,
142
conflict with partner, neighborhood disadvantage Olds, 4 years Receptive Peabody Picture General (82.13) (84.32) p = .04 .17 Intervention Kitzman, et postintervention vocabulary Vocabulary Test Linear condition al., 2004 (Child 6 years of (Direct Model showed higher age) Assessment) receptive language outcomes. Suskind et al., Postintervention Child number of Goldin- Ordinary Child Age 384.46 507.02 (787.29 1 week: NS (p < 1 week: .72* Child 2016 1 week and words Meadow, Least (541.26 - 1 1 week; 904.18 .1) 4 months: .87* vocalizations Postintervention Levine, Hedges, Squares week; - 4 months) 4 months: NS higher during 4 months Huttenlocher, Regression 616.06 - 4 intervention for Raudenbush, Models months) the intervention Child number of and Small 210.35 255.9 (334.72 - 1 week: NS 1 week: .69* group. Number utterances Coding System (260.14 - 1 1 week; 359.79 4 months: NS 4 months: .79* of word types (Direct week; - 4 months) significantly Assessment) 274.07 - 4 increased 1 months) week after Diversity of 101.22 125.03 (168.91 1 week: p < .01 1 week: .57 intervention but child (109.88 - 1 - 1 week; 4 months: NS (p 4 months: 1.12 did not sustain vocalizations - week; 182.41 - 4 < .08) significant number of word 131.12 - 4 months) differences at 4 types months) months. Short Diversity of 1.69 (1.77 - 1.81 (2.10 1 1 week: NS 1 week: .57* term gains child 1 week; week; 2.33 - 4 4 months: NS 4 months: .56* noted, not vocalization - 2.02 - 4 months) sustained mean length of months) postintervention. utterance in words During Conversational Language 28.73 (29.46 26.24 (41.95 - DI: p < .01 .66 intervention (DI) turns Environment - DI; 28.64 - DI; 35.14 - PI) PI: NS and Analysis PI) Postintervention Child (LENA) Digital 117.39 114.36 (162.01 DI: p < .04 .43 (PI) vocalizations Language (126.28 - - DI; 160.45 - PI: NS Processor DI; 123.91 - PI) (Observational) PI) Vally et al., Postintervention Child Language MacArthur- ANCOVA Baseline (9.62) (26.04) p < .001 .99* Significant 2015 Bates Scores differences Communicative noted for the Development intervention Inventory group in (Parent Report) language Language Peabody Picture (35.01) (46.83) p < .05 .5 outcomes. Comprehension Vocabulary Test (Direct
Assessment) 142
143
Table 4
Cognitive Outcomes Author, Date Follow-Up Cognitive Cognitive Method of Intervening or Data on Data on Finding on Effect Size Results Period Outcome Outcome Analysis Control Outcome Outcome for Differences Assessment Variables for Intervention Instrument Control Group: Baseline Group: (Follow-Up) Baseline (Follow- Up) Landry et al., Postintervention Mathematical Child Math Mixed Model Baseline 19.09 RECC: 16.23 NS NR** No significant 2014 Knowledge Assessment Analyses condition (31.56) (28.08) group (Direct RECC Plus: differences Assessment) 18.48 (34.99) found for growth of mathematical knowledge Love et al., Near intervention Mental Bayley Scales Regression Baseline (89.9) (91.4) p = .01 .12 Overall (all 2005 completion; at development of Infant models condition program types approximately 36 index Development combined) months of age Mental (Direct Hierarchical (88.9) (89.8) NS .07 significant development Assessment) Linear intervention index (for the Models effects noted in center-based cognitive program) development Mental (87.9) (89.3) NS .11 outcomes. development index (for the mixed-approach program) Mental (92.8) (94.1) NS .1 development index (for the home-based program) Lunkenheimer During Inhibitory Children's Repeated 4.16 (4.36) 4.24 (4.44) NS (p = .07) NR** Structural et al., 2008 intervention (3 control Behavior measures equation years of age) and (Executive Questionnaire ANOVA and modeling found Postintervention functioning) (Parent Report) Repeated direct effect of (4 years of age) Measures intervention on Structural child inhibitory Equation control at age 4 Modeling was not significant. A small indirect
effect of 143
144
intervention on child inhibitory control at age 4 through parental positive behavior support at age 3 noted (ß = .02) Mendelsohn Postintervention; Mental Bayley Scales T-Test (73.1) (74.2) NS (p = .75) .09* Children of et al., 2005 At 21 months of development of Infant higher educated age index Development mothers (7-11 (Participants (Direct grade education: with maternal Assessment) intervention education <7 associated with years) increases in Mental (70.3) (81.8) p = .01 .94* cognitive development development. index Children of (participants lower educated with maternal mothers (6th education >7 grade or less): years) no differences noted between intervention and control groups. Olds et al., Postintervention Mental Bayley Scales General Maternal age, (89.38) (Nurse Home NS NR*** No significant 2002 Development of Infant Linear Model housing Visitor: 90.13) effects on Index Development density, (Paraprofessional cognitive (Direct gestation, Home Visitor: outcomes. Assessment) maternal 89.45) Dichotomized conflict with (13) (Nurse Home NS NR*** developmental her partner, Visitor: 11) delay maternal (Paraprofessional conflict with Home Visitor: her mother 14) (Follow-up 2 years Executive Leiter General Maternal (99.69) (Nurse Home NS Nurse Home No significant assessment of postintervention Function International Linear Model psychologic Visitor: 100.64) Visitor: .09 intervention outcomes (Child 4 years of Composite Index Performance resources, age (Paraprofessional Paraprofessional: effects on from Olds et age) Scale-Revised; of gestation, Home Visitor: .00 executive al., 2002) Tap Test; Walk- maternal age, 99.70) function for A-line Test; housing intervention Olds, Day-Night Test density, groups. Robinson, et (Direct conflict with al., 2004 Assessment) partner (Follow-up 4 years Mental Kaufman General Maternal (94.85)) (Nurse Home NS Nurse Home No significant assessment of postintervention processing Assessment Linear Model psychological Visitor: 96.58) Visitor: .16 intervention outcomes (Child 6 years of composite Battery for resource index, (Paraprofessional Paraprofessional: effects noted for
from Olds et age) Children (Direct smoking status, Home Visitor: .18 cognitive 144
al. 2002) Assessment) age of 96.81) outcomes for
145
4 & 7 years Arithmetic Peabody gestation, Age 6: Age 6: (Nurse NS Age 6: Nurse intervention Olds et al., postintervention achievement Individual housing (92.66) Home Visitor: Home Visitor: groups. 2014 (Child 6 & 9 standard score Achievement density, Age 9: 94.27) .13 years of age) Test conflict with (96.53) (Paraprofessional Paraprofessional: partner, Home Visitor: .03 neighborhood 93.04) Age 9: Nurse disadvantage Age 9: (Nurse Home Visitor: Home Visitor: .02 Paraprofessional: .05 Reading Age 6: Age 6: (Nurse NS Age 6: Nurse achievement (93.4) Home Visitor: Home Visitor: standard score Age 9: 95) .46 (94.6) (Paraprofessional Paraprofessional: Age 9: (66) Home Visitor: .17 96.16) Age 9: Nurse Age 9: (Nurse Home Visitor: Home Visitor: .27 97.2) Paraprofessional: (Paraprofessional .13 Home Visitor: 97.24) Olds, 4 years Mental Kaufman General (90.24) (92.34) p = .03 .18 Intervention Kitzman, et postintervention processing Assessment Linear Model condition al., 2004 (Child 6 years of composite Battery for showed higher age) Reading Children (Direct (93.56) (93.79) NS (p = .84) .02 cognitive achievement Assessment) development outcomes.
Note. * Effect size not reported in the article; Cohen's d effect size calculated using methodology from Thalheimer and Cook (2002) **n not reported per study condition; unable to calculate effect size ***Standard deviation not reported; unable to calculate effect size
145
146
Table 5
Social Outcomes Author, Date Follow-Up Social Outcome Social Method of Intervening Data on Data on Finding on Effect Size Results Period Outcome Analysis or Control Outcome Outcome for Differences Assessment Variables for Intervention Instrument Control Group: Baseline Group: (Follow-Up) Baseline (Follow- Up) Goldfeld et 6 months Social Skills The Linear Caregiver (10.2) (10.0) NS (p = .90) .06* No significant al., 2011 postintervention; Communication regression education differences at 2 years of age and Symbolic models noted for social Behavior Scale skills. Infant-Toddler Checklist (Parent Report) Landry et al., Postintervention Social competence Social ANCOVA Baseline 3.59 RECC Plus: 3.47 p = .0098 NR** Age associated 2014 Competence and Mixed outcome, (3.54) (3.78) with children's and Behavior Model Child Initial RECC: 3.40 social Evaluation Analysis Age (3.76) competence Anger/Aggression (Direct 2.54 RECC Plus: 2.47 p = .0121 NR** levels; Observation) (2.45) (2.45) intervention RECC: 2.72 groups showed (2.32) more change Anxiety/Withdrawal 2.55 RECC Plus: 2.38 NS NR** over time in (2.36) (2.30) social RECC: 2.42 competence; (2.05) control group had no change over time. Children in RECC condition demonstrated greater decreases in anger/aggression scores. No differences noted on anxiety/ withdrawal. Love et al., Near Aggressive Child Behavior Regression Baseline (11.3) (10.6) p = .04 .11 Overall (all 2005 intervention behavior Checklist - models condition program types completion; at Aggressive combined)
Behavior significant 146
147
approximately Subscale intervention 36 months of age (Parent Report) effects noted in lower levels of Semi structured (4.6) (4.8) p < .01 .02 aggressive Play (using behavior and Three Box higher Coding Scales engagement in from NICHD parent-child Study of Early interaction. Child Care) (Direct Assessment) Aggressive Child Behavior Hierarchical (10.8) (9.6) NS (p = .71) -.18 behavior (for the Checklist - Regression center-based Aggressive Models program) Behavior Aggressive Subscale (11.3) (10.7) NS (p = .26) -.09 behavior (for the (Parent Report) mixed-approach program) Aggressive (11.7) (11.2) NS (p = .13) -.08 behavior (for the home-based program) Parent-child Semi structured (4.7) (4.9) p = .048 .17 interaction (for the Play (using center-based Three Box program) Coding Scales Parent-child from NICHD (4.4) (4.7) p = .04 .3 interaction (for the Study of Early mixed-approach Child Care) program) (Direct Parent-child Assessment) (4.6) (4.8) p = .000 .19 interaction (for the home-based program) Olds et al., Postintervention Irritability Bates Parent General Maternal age, (2.84) (Nurse Home NS NR*** No significant 2002 Perceptions of Linear housing Visitor: 2.80) effects on Difficult Model density, (Paraprofessional temperament or Temperament gestation, Home Visitor: behavior (Parent Report) maternal 2.83) problems. Behavior Problem Child Behavior conflict with (45.26) (Nurse Home NS NR*** Checklist for her partner, Visitor:43.71) Ages 2-3 maternal (Paraprofessional (Parent Report) conflict with Home Visitor: her mother 45.49) (Follow-up 2 years Behavioral Leiter General Maternal (99.71) (Nurse Home NS Nurse Home No significant assessment of postintervention Adaptation International Linear psychologic Visitor: 99.63) Visitor: -.01 intervention
outcomes Performance Model resources, age (Paraprofessional effects on 147
148
from Olds et (Child 4 years of Scale-Revised of gestation, Home Visitor: Paraprofessional: behavioral al., 2002) age) (Direct maternal age, 100.66) .09 adaptation, Emotion Regulation Assessment) housing (99.61) (Nurse Home NS Nurse Home emotion Olds, density, Visitor: 99.54) Visitor: -.01 regulation or Robinson, et conflict with (Paraprofessional Paraprofessional: externalizing al., 2004 partner Home Visitor: .13 behavior 100.86) problems. Externalizing Child Behavior (12.20) (Nurse Home NS Nurse Home Behavior Checklist Visitor: 12.16) Visitor: -.01 (Parent Report (Paraprofessional Paraprofessional: of rule-breaking Home Visitor: -.08 and aggressive 11.65) behavior) (Follow-up 4 & 7 years Internalizing Child Behavior General Age 6: Age 6: (Nurse NS NR** No significant assessment of postintervention Behavior Checklist Linear (176) Home Visitor: intervention outcomes (Child 6 & 9 (Parent Report) Model Age 9: 169) effects noted for from Olds et years of age) (164) (Paraprofessional behavioral al. 2002) Home Visitor: outcomes for 173) intervention Olds et al., Age 9: (Nurse groups. 2014 Home Visitor: 138) (Paraprofessional Home Visitor: 153) Externalizing Maternal Age 6: Age 6: (Nurse NS NR** Behavior psychological (176) Home Visitor: resource index Age 9: 169) (165) (Paraprofessional Home Visitor: 173) Age 9: (Nurse Home Visitor: 138) (Paraprofessional Home Visitor: 155) 7 years Attention Conners' (187) (Nurse Home NS NR** postintervention dysfunction Continuous Visitor: 166) (Child 9 years of Performance (Paraprofessional age) Test and Home Visitor: Clinical 177) Confidence Index (Direct Assessment) Olds, 4 years Classroom social Hightower General (24.53) (24.93) NS (p = .71) .03 No significant Kitzman, et postintervention skills Teacher-Child Linear effects noted on
al., 2004 Rating Scale Model 148
149
(Child 6 years of (Teacher social age) Report) development. Dysregulated McArthur Story (100.26) (99.24) NS (p = .26) -.1 aggression Stem Battery Warmth/empathy (Direct (99.51) (100.86) NS (p = .13) .14 Assessment)
149
150
Records identified through PubMed, Additional records identified Embase, PsycINFO, and Cinahl through hand search of citations
Databases (n = 4) (n = 3,352)
Identification Records excluded (n = 1,884) Records after duplicates removed Date of publication (306) (n = 2,546) Language (8)
Title (1,570)
Screening Records excluded by abstract Abstracts Screened (n = 506) (n= 662)
Full-text articles excluded (n = 141) Full-text articles assessed for eligibility Age of child (38) Duplicate dataset (6)
Eligibility (n = 156) Study design (86) Study outcomes (7) Study population (4)
Studies included in synthesis (n =15)
Included
Figure 1. PRISMA Flow Diagram
151
Appendix A
University of Texas Health Science Center Houston
CPHS Approval
152
153
154
Appendix B
Baylor College of Medicine Approval Letters
155
156
157
158
159
160
161
Appendix C Data Dictionary
162
1
62
163
1
63
164
1
64
165
1 65
166
1
66
167
1
67
168
1
68
169
1
69
170
Appendix D Attachment Style Questionnaire
171
Attachment Style Questionnaire-SF (ASQ) (Feeney, Noller, & Hanrahan,1994; Karantzas et al., 2010) Please show how much you agree with each of the following items by rating them on this scale:
1=totally disagree; 2=strongly disagree; 3=disagree; 4=agree; 5=strongly agree; 6=totally agree.
Totally Strongly Disagree Agree Strongly Totally Disagree Disagree Agree Agree 3. I feel confident that other people will be there when I 1 2 3 4 5 6 need them. 4. I prefer to depend on myself rather than other 1 2 3 4 5 6 people. 5. I prefer to keep to myself. 1 2 3 4 5 6 8. Achieving things (having goals and fulfilling them) 1 2 3 4 5 6 is more important than building relationships. 9. Doing your best is more important than getting along 1 2 3 4 5 6 with others. 10. If you've got a job to do, you should do it no matter 1 2 3 4 5 6 who gets hurt (emotionally hurt). 11. It's important to me that others like me. 1 2 3 4 5 6 13. I find it hard to make a decision unless I know what 1 2 3 4 5 6 other people think. 14. My relationships with people are generally 1 2 3 4 5 6 superficial (shallow or phony) 15. Sometimes I think I am no good at all. 1 2 3 4 5 6 16. I find it hard to trust people. 1 2 3 4 5 6 17. I find it difficult to depend on others. 1 2 3 4 5 6 18. I find that others are reluctant (hesitate) to get as 1 2 3 4 5 6 close as I would like. 19. I find it relatively easy to get close to other people. 1 2 3 4 5 6 20. I find it easy to trust others. 1 2 3 4 5 6
21. I feel comfortable depending on other people. 1 2 3 4 5 6
1
71
172
22. I worry that others won't care about me as much I 1 2 3 4 5 6 care about them. 23. I worry about people getting too close. 1 2 3 4 5 6 24. I worry that I won't measure up to other people. 1 2 3 4 5 6 25. I have mixed feelings about being close to others. 1 2 3 4 5 6 27. I wonder why people would want to be involved 1 2 3 4 5 6 with me. 29. I worry a lot about my relationships. 1 2 3 4 5 6 30. I wonder how I would cope without someone to love 1 2 3 4 5 6 me. 31. I feel confident about relating to others. 1 2 3 4 5 6
32. I often feel left out or alone. 1 2 3 4 5 6 33. I often worry that I do not really fit in with other 1 2 3 4 5 6 people. 34. Other people have their own problems, so I don't 1 2 3 4 5 6 bother them with mine. 37. If something is bothering me, others are generally 1 2 3 4 5 6 aware and concerned. 38. I am confident that other people will like and respect 1 2 3 4 5 6 me. Scoring: The 2 factors are (R = reverse scored): Avoidance: 3R, 4, 5, 8, 9, 10, 14, 16, 17, 19R, 20R, 21R, 23, 25, 34, 37R. Relationship Anxiety: 11, 13, 15, 18, 22, 24, 27, 29, 30, 31R, 32, 33, 38R.
1
72
173
Appendix E Edinburgh Postnatal Depression Scale
174
175
176
Appendix F
Collaborative Institutional Training Initiative Certificates
177
1
77
178
179
1
79
180
1
80
181
1
81
182
1
82
183
183
184
184
185
CURRICULUM VITAE Cary M. Cain, MPH, BSN, RN
The University of Texas Health Science Center at Houston Cizik School of Nursing 6901 Bertner Avenue Houston, Texas 77030 [email protected]
EDUCATION & LICENSURE Registered Nurse, Texas Board of Nursing, License No: 684857
University of Texas Health Science Center at Houston, Cizik School of Nursing (anticipated May 2019), Doctor of Philosophy Major: Nursing Dissertation: Impact of adult attachment style on outcome trajectories in a parent- directed early language development intervention Advisor: Cathy Rozmus, PhD, RN
University of Texas Health Science Center at Houston School of Public Health (May 2012), Master of Public Health Major: Management, Policy & Community Health Concentration: Global Health Practicum: Neurocysticercosis prevention in Eryuan County, Yunnan, China Advisor: Beatrice Selwyn, ScD, MScHyg, BSN
University of Texas at Arlington, School of Nursing (December 2001), Bachelor of Science in Nursing Major: Nursing
EMPLOYMENT & PRACTICA 2019 (anticipated) Baylor College of Medicine at Texas Children’s Hospital, Houston, Texas Assistant Professor, Department of Pediatrics, Section of Public Health & Child Abuse Pediatrics
2016 to present Baylor College of Medicine at Texas Children’s Hospital, Houston, Texas Research Associate, Department of Pediatrics , Section of Public Health & Child Abuse Pediatrics
2015 to 2016 Texas Children’s Hospital, Houston, Texas Community Outreach Coordinator, Child Abuse Pediatrics Department
186
2012 to 2015 Children’s Memorial Hermann Hospital, Houston, Texas Pediatric Injury Prevention and Outreach Education Coordinator
2011 to 2012 Children’s Memorial Herman Hospital, Houston, Texas Registered Nurse II
Summer 2011 Dali Institute of Parasitic Diseases, Eryuan County, Yunnan, China Practicum Internship
2009 to 2011 Houston Independent School District, Houston, Texas School Nurse
2007 to 2009 Lubbock Independent School District, Lubbock, Texas School Nurse
2002 to 2005 International Mission Board, Richmond, Virginia Asia American Cultural Exchange & Service Agency, Guizhou, China Community Health Advisor
2001 to 2002 Children’s Medical Center of Dallas, Dallas, Texas Registered Nurse
AWARDS 2018 Sigma Theta Tau Zeta Pi PhD Student Award for Excellence in Nursing 2018 Sigma Theta Tau International Edith Anderson Leadership Education Grant ($1,000 Travel Award) 2016 to 2019 Robert Wood Johnson Foundation Future of Nursing Scholar ($75,000 scholarship) 2016 to 2019 McGovern Foundation Accelerated PhD Program Scholarship ($50,000 scholarship) 2012 Children’s Memorial Hermann Patient Services Award 2012 Children’s Memorial Hermann C.H.A.M.P.S. (Colleagues in Healthcare Always Making People Smile) 2011 University of Texas School of Public Health HRSA Trainee Scholarship 2001 University of Texas at Arlington Outstanding Student Clinical Award for Nursing Leadership and Management 2001 University of Texas at Arlington Outstanding Student Clinical Award for Pediatrics 2001 University of Texas at Arlington Outstanding Student Clinical Award for Gerontology
187
GRANTS 2019 Episcopal Health Foundation: Early Childhood Brain Development. $667,600. Role: Early Brain Development Domain Lead/Analyst 2015 to 2016 Texas Department of Transportation: Live Your DREAMS (Distraction Reduction Among Motivated Students). $131,212.00 (49% match). Role: Co-director 2015 State Farm: State Farm Good Neighbor Citizen Program Safety Grant. $15,000.00. 2015. Role: Co-director 2014 to 2015 Texas Department of Transportation: Live Your DREAMS (Distraction Reduction Among Motivated Students). $128,743.00 (51% match). Role: Co-director 2014 Texas Pediatric Society Foundation: Implementation of “The Period of PURPLE Crying” in the Neonatal Intensive Care Unit. $1,000.00. Role: Project Director
PUBLICATIONS Refereed Cain, C.M., Pollok, L., Rodriguez, M., Edwards, C., & Cumpsty-Fowler, C. (2019). Evaluation of the Texas Injury Prevention Leadership Collaborative essential leadership skills training. Texas Public Health Journal, 71(2), 19-22. Cain, C.M. and Van Horne, B.S. (2017). Current challenges in addressing the realities of poverty and inequality in families with young children: A call for both policy and intervention. American Professional Society on the Abuse of Children (APSAC) Advisor, 29(2), 3-13.
Book Chapter Cain, C.M., Lopez, K.K., & Levine, S. (2018). Resiliency in children and adolescents. In M.W. Kline (Ed.) Rudolph's Pediatrics (23rd ed.) (Chapter 29). New York: McGraw Hill Medical.
Non-Refereed Lopez, K.K. & Cain, C.M. (2017). A public health approach to prevent child abuse and neglect fatalities: Summary of 'Within our reach: A national strategy to eliminate child abuse and neglect fatalities.’ American Professional Society on the Abuse of Children (APSAC) Alert, 8(3), 1-4. Valadez McStay, R., Lopez, K.K., Cain, C.M., Doan, P.D., Hayes, A., Hughes, L. (2016). Community Health Needs Assessment. Houston, Texas: Texas Children’s Hospital.
Manuscripts under review Cain, C.M., Lobiando-Wood, G, Santa-Maria, D., Mandell, D.J., & Rozmus, C. (under review). Early language interventions for low socioeconomic status populations: A systematic review of the efficacy on child language, cognitive, and social development. Developmental Review. Keefe, R.J., Van Horne, B.S., Thompson, R., Budolfson, K., Cain, C.M., & Greeley,
188
C.S. (under review). Mental health diagnoses of children in foster care: A comparison study between foster and non-foster children covered by Medicaid. Pediatrics. Correa, N.P., Cain, C. M. Bertenthal, M.S., & Lopez, K.K. (under review). Screening for intimate partner violence in the healthcare setting: Survivors’ perspectives. Violence Against Women.
PRESENTATIONS & CONFERENCE SESSIONS Cain, C.M., Mandell, D.J., Padhye, N., & Rozmus, C. (submitted). Impact of adult attachment style in outcome trajectories in a parent-focused early language development intervention. American Public Health Association Annual Meeting. Philadelphia, PA. Cain, C.M., Barry, R.O., Howell, J.J., Dugan, M.K., & Lopez, K.K. (submitted). More than just words: Parent-child engagement to promote protective factors that are associated with root causes of violence. Safe States Alliance Annual Meeting. Atlanta, GA. Contreras, M.A., Williams, S.R., Edwards, C., Northway, J., Pollok, L., Rodriguez, M., Cain, C.M., & Stephens-Stidham, S. (submitted). Keep Texas Safe: A Collaborative Policy Effort. Safe States Alliance Annual Meeting. Atlanta, GA. Cain, C.M. (November 12, 2018). Association of Children’s Externalizing Behavior Disorder and Maternal Mental Health: Implications for Public Health. American Public Health Association Annual Meeting. San Diego, CA. (Poster Presentation) Cain, C.M. & Mandell, D.J. (October 5, 2018). Support for the Maternal-Child Dyad: Association of Children’s Externalizing Behavior Disorder and Maternal Mental Health. Society for Research in Child Development. Phoenix, AZ. (Poster Presentation) Cain, C.M. & Stidham-Stephens, S. (September 7, 2018). Evaluation of the Texas injury prevention leadership collaborative essential leadership skills training. Safe States Alliance Annual Meeting. Charleston, SC. (Oral Presentation) Abela, K., Acorda, D., Cain, C.M. & Gibbs, K. (August 8, 2018). Relationship of Food Insecurity to Sociodemographic and Hospitalization Characteristics in Children Living in Texas. Texas Children’s Hospital Professional Day. Houston, TX. (Oral Presentation) Cain, C.M. (July 22, 2018). Associations of Children's Externalizing Behavior Disorder and Maternal Mental Health: Implications for Nursing. Sigma Theta Tau 29th International Nursing Research Congress. Melbourne, Australia. (Oral Presentation) Cain, C.M. & Dugan, M. (April 17, 2018). Strengthening families to foster resilience through early language development. Texas Nurse Family Partnership Conference. Dallas, TX. (Oral Presentation) Correa, N.P. & Cain, C.M. (October 4, 2017). Mitigating Adverse Childhood Experiences and Fostering Resilience: Building Community Coalitions to Prevent Child Maltreatment. Texas Department of Family Protective Services 17th Annual Partners in Prevention Conference. San Antonio, TX. (Oral Presentation) Cain, C.M. & Dugan, M. (October 3, 2017). Strengthening Families to Foster Resilience
189
through Early Literacy. Texas Department of Family Protective Services 17th Annual Partners in Prevention Conference. San Antonio, TX. (Oral Presentation) Cain, C.M., Dugan, M., Lopez, K.K., & Greeley, C. (September 12, 2017). Texas Children's Hospital "upWORDS:Baby": More than just words. 2017 LENA Early Language Conference. Beaver Creek, CO. (Invited Poster Presentation) Abela, K., Acorda, D., Cain, C.M., & Gibbs, K. (June 9, 2017). Relationship of Food Insecurity to Sociodemographic and Hospitalization Characteristics in Children Living in Texas. University of Texas Health Science Center at Houston School of Nursing Annual Research Day. (Poster Presentation) Bickel, J.E. & Cain, C.M. (July 19, 2016). The Power to Prevent Child Maltreatment. 2016 Texas WIC Conference. Austin, TX. (Oral Presentation) Cain, C.M., Gonzalez, D., & Lempertz, I. (May 42, 2016). Utilization of Get Well Network to implement and evaluate an abusive head trauma prevention program. Get Well Network Interactive Patient Care Conference. Philadelphia, PN. (Oral Presentation) Cain, C.M. (April 29, 2016). Adverse Childhood Experiences: A Call to Prevent Abusive Injuries to Children. South Texas Regional Advisory Council, Regional Emergency Healthcare Systems Conference. San Antonio, TX. (Invited Oral Presentation) Cain, C.M. (April 7, 2016). Applying evidence-informed strategies to prevent child maltreatment. Texas Emergency Nurses Association Injury Prevention Conference. Seabrook, TX. (Invited Oral Presentation) Cain, C.M. & Wesson D. (January 7, 2016). Trauma data to inform child maltreatment prevention. Surgical Lecture Series for Baylor College of Medicine/Texas Children’s Hospital Academic Pediatric Emergency Department. Houston, TX. (Oral Presentation) Cain, C.M. (September 24, 2015). Selecting an evidence-informed program for the prevention of child maltreatment. Safe States Alliance Hospital Injury Prevention Special Interest Group. National Webinar. September 24, 2015. (Invited Webinar Presentation) Bickel J.E. & Cain C.M. (September 10, 2015). Abusive Head Trauma, Infant Crying, and the Stressed Out Caregiver. Pediatric CME Grand Rounds, Texas Children’s Health Plan. Houston, TX. (Invited Oral Presentation) Cain, C.M. (August 17, 2015). Abusive Head Trauma Prevention. Texas EMS, Trauma and Acute Care Foundation Injury Prevention Conference. Austin, TX. (Invited Oral Presentation) Bickel J.E. & Cain C.M. (June 23, 2015). Abusive Head Trauma, Crying, and the Stressed Our Caregiver. Texas Department of State Health Services, Health Texas Babies: A Life Course Conference. Houston, TX. (Session Presentation) Cain C.M., Rubalcava D., Frost M., Jurek M., Zhu J., Barger S., Wesson D., & Cox C. (November 10, 2014). Collaborative use of trauma registry data to inform non- accidental trauma prevention. American College of Surgeons Trauma Quality Improvement Program Annual Meeting. Chicago, IL. (Poster Presentation) Cain, C.M. (February 27 & March 12, 2013) Injury Prevention: An essential component of trauma care. Memorial Hermann Trauma Nursing Symposium. Houston, TX. (Invited Oral Presentation)
190
Cain, C.M. (February 25, 2013). Injury Prevention Program Evaluation. Texas EMS, Trauma and Acute Care Foundation Injury Prevention Conference. Austin, TX. (Invited Oral Presentation)
PROFESSSIONAL AFFILIATIONS 2018 to present Society for Research in Child Development, Member Society for Advancement of Violence and Injury Research (SAVIR), Member Safe States Alliance, Member 2016 to present Texas Injury Prevention Leadership Collaborative, Strategy Team Leader 2015 to 2016 Texas Department of Family Protective Services Prevention and Early Intervention Division, Texas Prevention Network, Evidence Based Programming Think Tank Co-Chair 2012 to present American Public Health Association, Member Maternal and Child Health and Injury Control and Emergency Health Services, Member 2012 to present Safe Kids Greater Houston Coalition, Member 2012 to 2016 Harris County Child Fatality Review Team, Hospital Representative 2001 to present Sigma Theta Tau Nursing Honor Society, Membership Chairperson (Zeta Pi Chapter)