REPORT

Children of the new century

Mental health findings from the Millennium Cohort Study

Leslie Morrison Gutman, Heather Joshi, Michael Parsonage and Ingrid Schoon 21 10 13 7 3 29 28 25 differences persistence Socio-demographic and trends Recent Incidence 6 problems methods health and mental sources Introduction of Data Prevalence We are very grateful to members of our Expert Reference Group for helpful advice and support: advice helpful for Group our Expert Reference of members to very grateful are We Dr Naomi Eisenstadt Maughan Barbara Professor Layard Richard Lord Professor Acknowledgements Appendix tables References Appendix 4 5 6 Executive summary Executive 1 2 3 Contents

Centre for Mental Health REPORT Children of the new century 2 Centre for Mental Health Executive Summary

This report sets out early findings from a three- among 11-year-olds as rated by teachers was year project on children’s mental health being somewhat lower at 8%, with an even wider gap undertaken by the Institute of Education at UCL between boys and girls than suggested by the and Centre for Mental Health, with funding from parents’ scores. the Economic and Social Research Council. As many as 28% of all 11-year-old children REPORT The findings relate mainly to the mental health were assessed by parents as having severe of children around the age of 11 as recorded difficulties on at least one of the four individual in the Millennium Cohort Study (MCS), a multi- SDQ subscales. Children of the new century purpose longitudinal study which is following a large sample of children born in the UK at Social and demographic differences the start of the 21st century. The data were collected mainly in 2012. The prevalence of severe mental health problems in children is strongly related to Children’s mental health is assessed in the parental education, parental occupation and MCS using the Strengths and Difficulties family income. For example, 17% of 11-year- Questionnaire (SDQ), a widely used screening olds from families in the bottom fifth of the instrument in which parents and teachers report income distribution were identified as having on a child’s mental health in the previous six severe mental health problems in 2012, months. compared with only 4% among those from The SDQ distinguishes between four broad families in the top fifth. This income-related groups or “subscales” of mental health gradient in prevalence appears to have become difficulties (conduct problems, hyperactivity/ steeper in recent years and to be much steeper inattention, emotional problems and peer among children than it is among adults. problems). A combined score across all four Looking at 11-year-old-children living in the subscales which is above a defined cut-off is four countries of the UK, the main feature is that indicative of a clinically diagnosable disorder. Scotland showed a somewhat lower prevalence Children with scores at a level where a disorder of severe mental health problems in 2012 than is likely to be present are described in this England, Wales or Northern Ireland. report as having “severe” mental health Comparisons of ethnic groups suggest that problems. severe problems were most common in 11-year- old children of Mixed ethnic background Prevalence (particularly girls), followed by those classified as White (particularly boys). Prevalence was Based on the SDQ scores reported by parents, lowest among children of Indian origin. the proportion of 11-year-old children in the UK with severe mental health problems in 2012 When we look at the type of family, the was just over 10%. prevalence of severe mental health problems among 11-year-old children living with both There were markedly more boys reported as their natural parents was well under half the having severe problems than girls (13% vs level found among children in other family types 8%). There were also some differences by (single-parent families, step-families etc.). gender in the nature of problems: among boys ‘externalising’ problems (conduct problems and hyperactivity/inattention) were relatively Recent trends more common than ‘internalising’ problems Recent changes have been assessed by (emotional and peer problems), while among comparing data on the mental health of children girls the opposite was the case. as recorded in the 2012 sweep of the MCS with The level of severe mental health problems similar data on the mental health of 10- and 11-

3 assessing the need for mental health services. services. health mental the need for assessing explain to report this of the purpose beyond is It might they or how problems these of the causes questions important These be prevented. evidence, the of further require exploration do next. to intend we something face adolescence, particularly in the context of of context in the particularly adolescence, face Looking ahead Looking health mental of the extent charting are We they as the new century of in children problems had severe problems at three or four surveys surveys or four three at problems severe had 3.6% of that estimated is 11, it age at including category. this into fall children 11-year-old all common as times three nearly is Persistence girls. among as boys among at some point during childhood is roughly twice twice roughly is childhood during some point at at problems these of the prevalence high as as 1.5% of only On the other hand, age. single any ages. four all at problems severe had children who children those as cases persistent Defining (persistence). were in the MCS children all a fifth of over Just problems health mental severe with assessed This surveyed. ages the four of one or more at problems severe of the incidence that implies children in the sample at four different ages ages different four at in the sample children show to be analysed can This (3, 5, 7 and 11). mental severe recorded ever children many how 3 and 11 of the ages between problems health problems severe with many how and (incidence) repeatedly problems these 11 had age at Incidence and persistence Incidence so far has Study Cohort The Millennium of health on the mental information collected prevalence of severe mental health problems problems health mental severe of prevalence ratings teachers’ In contrast, group. age in this in the going changes larger somewhat suggest by noted The improvements direction. opposite to measure equal in broadly applied teachers and girls. both boys concerned and whether the rating is made by by made is and whether the rating concerned or teachers. parents if appears, there parents, reported by As increase been a small have to anything, in the overall 1999 and 2012 between Overall, the analysis suggests that the mental the mental that suggests the analysis Overall, improved group age in this children of health The picture 1999 and 2012. between slightly vary and findings mixed somewhat however is problem the type of by and girls, boys between year-old children in the national Child and and Child in the national children year-old in out carried Surveys Health Mental Adolescent 2004. 1999 and

Centre for Mental Health REPORT Children of the new century 4 Centre for Mental Health Children of the new century: Mental health findings from the Millennium Cohort Study

Key REPORT

points Children of the new century from the report 10% of 11-year-old children experienced a mental health problem in 2012

Children from low-income families are four times more likely Top to experience mental health problems 20% than children from higher-income families

Bottom 20% Mental health problems are twice as common 2500 X2 in boys as girls 2000

1500

1000 20%

500

0 1999 2012 Overall, the mental health of 11- year-old children was broadly the same in 2012 as in 1999 Over 20% of children experience a mental health problem at some point between ages 3 and 11

5 2004 BCAMHS and the MCS. 2004 BCAMHS incidence of on patterns information Finally, problems health in mental and recurrence on data drawing children, the same within the MCS of surveys in successive collected 3, 5, 7 and 11. ages child at The prevalence of mental health problems problems health mental of The prevalence in the UK, children 11-year-old among which to the extent of analysis including or more two be displaying may children time. the same at problems in the differences Socio-demographic problems health mental of prevalence the extent examining 11-year-olds, among a to according in prevalence variation of residence of country as such factors of range type, family the UK, ethnicity, within and occupation parental education, parental income. family health mental of in the prevalence Trends 10- and 11-year-old among problems on 1999 and 2012, based between children in the 1999 and findings of a comparison • readership a general with written The report is quantitative possible, in mind and, wherever simple relatively by illustrated is information tabular detailed than rather diagrams specialist a more with those For presentations. supporting information, detailed interest, and statistical numbers sample including of series in a provided is tests, significance the report. the end of at tables appendix on four main topics: topics: main on four • • • . 1 1. Introduction Following a short review of data sources and sources data of a short review Following information and discuss present we methods, Gutman. Morrison Leslie to ES/L0008211/1 grant 1 Research report is on the mental health of children around around children of health on the mental report is the at experiences reflects 11, which of the age time an important and at adolescence of onset the of because children most for transition of education. secondary to primary from move the MCS is that it is based on a substantially on a substantially based is it that is the MCS the BCAMHS, than children of sample larger sub-group detailed for scope more allowing ethnicity. by example for analysis, this of focus exclusive, though not The main, are using information which is increasingly increasingly is which information using are recent the most In contrast, date. of out out carried was in the MCS children survey of a more of data in 2012 and so provides mainly children to confined albeit nature, contemporary of advantage Another year. born in a specific relevance. For example, it is now more than ten ten than more now is it example, For relevance. and Child British official the last since years Survey (BCAMHS) was Health Mental Adolescent on policy decisions that implying in the field, source on this rely which planning or service This was a necessary first step in the analysis the analysis in step first a necessary was This generated also has the work but trajectories, of mental of levels on overall information much the 21st of the start born at in children health and interest wider of be may which century longitudinal study which is following a large a large following is which study longitudinal in 2000/01. in the UK born children of sample the extraction entailed on the project work Early on the mental data of amounts substantial of successive from in the MCS children of health 3, 5, 7 and 11. ages child at out carried surveys Council analyse to is project the joint aim of The primary during problems health mental of trajectories the Millennium from data using childhood a multi-purpose (MCS), Study Cohort three-year project on children’s mental health health mental on children’s project three-year of the Institute by jointly undertaken being College University part of now Education, with Health, Mental for and Centre London, Research Social and the Economic from funding This report sets out some early findings from a from findings early some out sets report This

Centre for Mental Health REPORT Children of the new century 6 Centre for Mental Health Chapter 2: Data sources and methods

Data Note also that parents and teachers were asked to rate problems over the last six months, which The Millennium Cohort Study (MCS) is a would have included time before the 11th longitudinal study following a sample of birthday for most of those who had already children born in all four countries of the United reached that birthday.

Kingdom between September 2000 and January REPORT 2002. Unlike previous national birth cohort Cases studied

studies, which attempted to recruit all children Children of the new century born in a particular week, this survey drew on a In our analyses we examined one child per sample using a complex clustered and stratified family present at age 11, excluding 182 children design, with deliberate over-sampling of in MCS5 who were the second or third in sets children in areas of high child poverty, minority of twins or triplets. (There were not enough ethnic populations and the three smaller of these to analyse separately.) Our sample countries of the UK (Plewis, 2007). was further reduced because about 4% of parents in MCS5 were not able to provide There have so far been five rounds of interviews usable information on their child’s mental completed with MCS families. The first survey, health, leaving a total of 12,798 cases for MCS1, was in the field in 2001 and 2002, at analysis. Almost all of the parental informants about nine months after birth of the sampled were mothers. They used computer-assisted children; MCS2 (child age 3) was mainly self-completion to answer the questions on undertaken during 2004, MCS3 (child age 5) children’s mental health. mainly during 2006 and MCS4 (child age 7) mainly during 2008, while the most recent We also drew on responses from teachers in completed survey, MCS5, at child age 11, England and Wales who answered a postal collected data mainly in 2012. A sixth survey is paper questionnaire including questions on in the field in 2015, and another one planned children’s mental health in the 2012 survey. for 2018. (The questionnaire was not sent out in Scotland or Northern Ireland). Teachers provided usable The number of families who have been information on mental health for 7,085 children. interviewed at least once is 19,244, which includes some 692 families in England who Weighting procedures were used to correct for were not interviewed until MCS2. If these cases the over-sampling of certain groups of children are included, the initial response rate to the in the MCS and for sample attrition, aiming to survey out of all those eligible was 71%. As is ensure that the socio-demographic profile of the typical of longitudinal studies, there has been whole cohort is nationally representative. some loss to follow-up, or ‘attrition’. Including some former drop-outs who had re-joined, Measurement of mental health the number of families interviewed at MCS5 in 2012 was 13,287, corresponding to 69% of As in the earlier surveys of children aged 3, 5 those ever interviewed (Mostafa, 2014). Finally, and 7 years, children’s mental health around the number of families interviewed at all five age 11 was assessed using the Strengths and surveys undertaken to date was 10,448. Difficulties Questionnaire (SDQ) (Goodman, 1997 and 2001). This is a widely tested and The fifth survey is referred to as the ‘age validated screening questionnaire in which 11’ survey and the children it surveys as parents and/or teachers report on a child’s (for convenience) ‘11-year-olds’, although mental health in the previous six months. interviewing took place at ages around, rather than necessarily after, the 11th birthday. One The SDQ includes 25 items divided into five third were still 10, though approaching 11, and groups or subscales relating to: most of the rest were closer to their 11th than • conduct problems (e.g. often has temper 12th birthday. 7 it is also “substantially higher than the higher than “substantially also is it child of reported other measures for average al., 2009). et (Collishaw psychopathology” divergence to contribute may factors Various particular reports, but and teacher parent of differences, contextual to attaches importance behave may children that partly in the sense bands are described as having ‘no problems’, ‘no problems’, having as described are bands problems’ and ‘severe problems’ ‘moderate respectively. and teachers parents by Assessments Study Cohort the Millennium above, noted As as teachers SDQ reports from collected has of both sets from and findings parents, as well points various at presented are assessments apparently with report, sometimes in this discrepant results. common is informants multiple of The use health, mental child of in studies practice is or measure instrument specific whatever observed that been widely has it and used, the presence about often disagree informants al., et Reyes Los (De symptoms of or absence While this. to no exception SDQ is The 2015). between agreement that been found has it “moderate”, is measure this using informants relevant to the analysis in Chapter 6 of the of 6 in Chapter analysis the to relevant health mental of recurrence and persistence some where childhood, during problems reflect simply may discontinuities apparent point a cut-off above just rated being children on another.) below and just on one occasion ‘normal’ numbers, actual with illustrate To difficulties total a parent-rated to corresponds 0-13, ‘borderline’ in the range is which score 14-16 and in the range a score to corresponds 17-40. An in the range a score to ‘abnormal’ but – difficulties total for score ‘abnormal’ subscales the individual for necessarily not a clinically of be indicative to understood – is and (Goodman disorder mental diagnosable a for very rare in practice is 2009). It Goodman, but 20, exceed to score difficulties total child’s where zero, be to the total for common quite reported. are no difficulties absolutely modify the but system grouping this retain We in the three children that such terminology are to an extent arbitrary, a point particularly particularly a point arbitrary, extent an to are pro-social (e.g. considerate of other of considerate (e.g. pro-social other with readily feelings…shares people’s children). younger to children…kind clingy in new situations); in new situations); clingy to tends solitary, rather (e.g. peer problems on on or bullied…gets alone…picked play and other children); than adults with better overactive…constantly fidgeting or fidgeting overactive…constantly concentration distracted, squirming…easily wanders); worries… many (e.g. problems emotional or downhearted…nervous often unhappy, tantrums…often fights with other children… with other fights tantrums…often elsewhere); or home, school from steals restless, (e.g. hyperactivity/inattention cut-off points such that 80% of children scored scored children 80% of that such points cut-off and 10% ‘abnormal’. 10% ‘borderline’ ‘normal’, on based different somewhat are points Cut-off (The cut-offs child. the of the reporter and age reflecting a continuum, it is often convenient convenient often is continuum, it a reflecting The original the scores. or group categorise to were SDQ scores the for presented groupings These ‘abnormal’. and ‘borderline’ ‘normal’, in a population- established were bandings choose to attempting survey, national based subscales in whole-population samples, samples, in whole-population subscales may subscales separate the four whereas in high-risk discrimination better provide al., 2010). et (Goodman samples as be treated can SDQ scores Although and hyperactivity/inattention subscales, and subscales, and hyperactivity/inattention the emotional of the sum is ‘internalising’ some is There subscales. and peer problem scales combined two these using that evidence separate the four using to be preferable may and peer problem subscales; the pro-social the pro-social subscales; and peer problem calculation. in this used not is subscale between drawn sometimes is A distinction scores, and ‘internalising’ ‘externalising’ conduct the of the sum is ‘externalising’ where true and 2 = very true. The five items in each each in items The five very true. true and 2 = giving then totalled, are or subscale group 10. A 0 to from ranging potentially scores is score problems) total (or difficulties total the of the scores of the sum as calculated emotional hyperactivity/inattention, conduct, • scoring on a three-point marked is item Each true, 1 = somewhat 0 = not in which system • • • •

Centre for Mental Health REPORT Children of the new century 8 Centre for Mental Health differently in different contexts (e.g. home v. school) and partly because some sorts of difficulties may be more observable in some contexts than others. To this extent, each reporter is providing useful and complementary information and no single informant’s reports are sufficient for a comprehensive assessment.

Statistical significance REPORT

Although the MCS is a large-scale survey, there Children of the new century always remains the possibility, particularly in the detailed analysis of sub-groups or trends over time, that apparent variations in the prevalence of severe mental health problems may simply be the result of chance rather than a genuine underlying difference or change. On the basis of statistical significance tests, this report particularly highlights findings which have a likelihood of 95% or more that they represent genuine differences.

9 e t None Severe Modera s Girl problems based on the SDQ scores for total total for SDQ scores on the based problems Figure in shown is teachers by given difficulties relate data these that be noted should 3.2. It therefore not and are Wales and England to 3.1, in Figure those with comparable directly the UK. of all to relate which There are, however, differences when boys and boys when differences however, are, There the girls, Among separately. analysed are girls problems, internalising severe of prevalence higher than is problems, emotional particularly problems, externalising severe of the prevalence broadly is the opposite boys among whereas between variation overall less with true, albeit subscales. the four the prevalence subscales, the four of three For higher is problems health mental severe of being the one exception girls, than boys among the prevalence where problems, emotional in both the same much is problems severe of as boys many as twice than More genders. severe having as parents by identified are girls problems. hyperactivity/inattention ratings on teacher based Prevalence health mental of prevalence The overall Boys n ildre ch All 0 20 40 60 80 100 Figure 3.1: Percentages of 11-year-old children with mental health problems: parent ratings, ratings, parent problems: health mental with children 11-year-old of 3.1: Percentages Figure UK, 2012 % Chapter 3: Prevalence of mental health problems health of mental 3: Prevalence Chapter much the same in all four SDQ categories, SDQ categories, four in all same the much problems both conduct 11.5% for from ranging for 11.9% to and hyperactivity/inattention peer and 12.3% for problems emotional problems. Detailed results for prevalence broken down down broken prevalence for results Detailed of subscales the four using problem, type of by This A.1. Table SDQ, in Appendix given are the children, 11-year-old all among that, shows are problems severe with the proportions disorder. The prevalence of both moderate and both moderate of The prevalence disorder. higher among markedly is problems severe boys 12.7% of example, For girls. than boys health mental severe having as classified are girls. 7.7% of only with compared problems This indicates that, among all children, 82.8% children, all among that, indicates This health no mental having as classified are problems moderate 6.9% have while problems in indicative problems, severe and 10.3% have mental diagnosable a clinically of case the latter Prevalence based on parent ratings on parent based Prevalence health mental of prevalence The overall in the UK, children 11-year-old among problems difficulties total for SDQ scores the on based 3.1 below. in Figure shown is parents, by given

Centre for Mental Health REPORT Children of the new century 10 Centre for Mental Health Figure 3.2: Percentages of 11-year-old children with mental health problems: teacher ratings, England and Wales, 2012 100 Severe

80 Moderate

None REPORT % 60 Children of the new century

40

20

0 All children Boys Girls

On the teacher ratings, 84.2% of all children Combinations of mental health in England and Wales have no mental health problems problems, 7.8% have moderate problems and 8.0% have severe problems. The proportion This section provides information on the extent with severe problems is thus somewhat lower to which some children exhibit two or more than the corresponding estimate for the UK severe mental health problems at the same based on parent ratings; on the other hand, the time. Findings based on parent ratings are proportion with moderate problems is slightly given in Figure 3.3. higher. More strikingly, the teachers identify This shows that 28.1% of all children have well over twice as many severe problems severe difficulties in at least one of the SDQ among boys as girls, a noticeably wider gender subscales. The total breaks down as follows: difference than in the parents’ scores. 15.7% of children are classified as having one A detailed breakdown of the teacher-rated severe problem, 7.2% have two such problems, scores by type of problem is given in Appendix 3.6% have three and 1.6% have severe ratings Table A.2. In line with the parent data, this in all four subscales. shows that among girls the prevalence of There are also gender differences in the extent severe internalising problems is higher than the to which types of problem are combined. For prevalence of severe externalising problems, 11-year-old boys, 31.4% are classified with whereas among boys the opposite is the case. one or more severe problems, including 16.4% But there are also differences. For example, with just one problem, 8.5% with two, 4.3% the teacher ratings show that the prevalence with three and 2.2% with severe ratings in all of severe conduct problems is three times as four subscales. All of these rates are lower for high in boys as in girls, while the prevalence of girls, with 24.6% having at least one problem, severe hyperactivity/inattention problems is including 15.0% with one problem only, 5.8% nearly five times as high. These are much wider with two problems, 2.5% with three and 1.0% gender differences than in the parent reports. with four.

11 s m o problems One proble No problems Tw Three problems Four problem s Girl ildren ch All s y Bo Figure 3.3: Percentages of 11-year-old children with different numbers of severe mental mental severe of numbers different with children 11-year-old of 3.3: Percentages Figure UK, 2012 ratings, parent problems: health conditions). The internalising grouping is is grouping The internalising conditions). vice and boys than girls among common more Among grouping. the externalising for versa in one problem least at 8.9% have children, all groups. and externalising both internalising children classified as having severe problems problems severe having as classified children common the most subscales, separate in two + peer problems emotional is grouping conditions), internalising the two (i.e. problems + hyperactivity/ problems conduct by followed externalising the two (i.e. inattention Detailed information on all possible possible all on information Detailed in Appendix given is problems of combinations for that example, for shows, This A.3. Table

Centre for Mental Health REPORT Children of the new century 12 Centre for Mental Health Chapter 4: Socio-demographic differences

Introduction Figure 4.1: Percentages of 11-year-old children This chapter explores differences in the with severe mental health problems (total proportions of 11-year-old children with severe difficulties) by country of residence: parent mental health problems according to a range of ratings, UK, 2012 socio-demographic factors such as ethnicity and All children REPORT family income. All the findings are based on 15

SDQ scores provided by parents. It is important Children of the new century to note that the associations shown below do 12 not necessarily imply causation, as many other % 9 factors may also be at play and further analysis is needed to disentangle their separate effects. 6

Country of residence within the UK 3 0 The relative proportions of 11-year-old children England Wales Scotland N. Ireland in the four constituent countries of the UK who have severe mental health problems, Boys as measured by their SDQ scores for total 15 difficulties, are given in Figure 4.1. 12 The main feature of Figure 4.1 is the lower prevalence of severe mental health problems in % 9 Scotland compared with the rest of the UK; thus the relative numbers in Scotland with severe 6 problems are below those elsewhere by about a quarter among both boys and girls. It should, 3 however, be noted that this difference falls just 0 short of statistical significance. In the UK as a England Wales Scotland N. Ireland whole the prevalence of severe problems among Girls boys is highest in Wales, while among girls it 15 is highest in England. Prevalence in Northern Ireland is virtually identical to the UK average, 12 both for children taken together and for boys and girls separately. % 9

Detailed information on the proportions with 6 severe problems in each of the four SDQ subscales is given by country in Appendix Table 3 A.4. Among other things this suggests that the 0 lower prevalence of severe problems in Scotland England Wales Scotland N. Ireland is somewhat more pronounced for internalising problems than it is for externalising problems. Within the UK as a whole, the prevalence of Ethnicity severe problems is highest in England for A breakdown by ethnicity of the prevalence of conduct problems and peer problems, in Wales severe mental health problems, based on SDQ for hyperactivity/inattention, and in Northern scores for total difficulties, is given in Figure 4.2 Ireland for emotional problems. (overleaf).

13 Other Other Other k k k Blac Blac Blac i gladesh Pakistani n Pakistani Pakistani /Bangladeshi /Ba /Bangladeshi Indian Indian Indian d d d Mixe Mixe Mixe e e e Whit Whit Whit ildren 9 6 3 0 9 6 3 0 s 9 6 3 0 15 12 15 12 15 12 ch Girl % Boys % All % Figure 4.2: Percentages of 11-year-old children with severe mental mental severe with children 11-year-old of Percentages 4.2: Figure 2012 UK, ratings, parent ethnicity: by difficulties) (total problems health

Centre for Mental Health REPORT Children of the new century 14 Centre for Mental Health This shows that, among all children, prevalence for each of the four SDQ subscales given in is highest in the Mixed group, followed by Appendix Table A. 6. those classified as White. In all the other four ethnic groups for which information is given, Figure 4.3: Percentages of 11-year-old children the prevalence of severe problems is below the with severe mental health problems (total national average and is particularly low among difficulties) by partnership status of parents: children of Indian origin. These differences are parent ratings, UK not, however, at a level which reaches statistical

All children REPORT significance. 20 Patterns for boys and girls within ethnic groups Children of the new century are broadly similar but with some exceptions. 15 In particular, the prevalence of severe problems among boys in the Mixed group is below rather % 10 than above the national average, meaning that the high overall prevalence of severe problems 5 in this group is fully explained by an extremely high level of problems among girls. Severe 0 Both natural Lone natural mental health problems are most common Step and other parents parent family types among White boys, closely followed by those classified as Black. Among both boys and girls, Boys prevalence is lowest in Indians, with the rate for 25 boys being particularly low. 20 Detailed information on severe problems in each of the four SDQ subscales by ethnicity is 15 % given in Appendix Table A.5. Features of note 10 include: very low levels of conduct problems - but not hyperactivity/inattention - among Indian 5 children; high levels of emotional problems 0 among children in the Mixed group, particularly Both natural Lone natural Step and other girls; exceptionally low levels of conduct parents parent family types problems among girls in the Black and Other Girls groups; and very high levels of peer problems among boys in the small and heterogeneous 15 Other group. 12 The pattern of severe problems across ethnic 9 groups just described contrasts with earlier % waves of the survey, where Indians and Mixed 6 had similar total difficulties scores to Whites, while Pakistanis and Bangladeshis, particularly 3 the former, had significantly higher levels of 0 problems (George et al., 2006). The earlier Both natural Lone natural Step and other parents parent family types surveys show a contrast between a high level of problems among Black Caribbeans and a particularly low one for Black Africans. The prevalence of severe mental health problems among children living with both their Partnership status of parents natural parents is about a half to a third of the level found among children in other family types The prevalence of severe mental health (single-parent families, step-families etc.). The problems on the total difficulties measure differential is broadly the same among boys as according to parents’ partnership status is given among girls and is also observed across all four in Figure 4.3, with matching information 15 Higher degree a Degree /Diplom l AS/ A-leve Higher degree GCSE A-C a . Low qual Degree /Diplom ations l No ademic ac AS/ qualific 5 0 A-leve 25 20 15 10 s problems in their children. For example, using using example, For children. in their problems of children the difficulties, total for SDQ scores are qualifications low very no or with parents severe exhibit to likely as times three about children those as problems health mental up to education a university have parents whose likely as four times and about level first-degree higher-degree have parents whose those as qualifications. be to appears association The inverse among than girls among stronger somewhat SDQ by analysis the detailed while boys, suggests table in the appendix given subscale pronounced more is the relationship that for is it than problems externalising for problems. internalising higher the prevalence of severe mental health health mental severe of prevalence higher the Girl % GCSE A-C Higher degree a Low qual. Degree /Diplom l ations AS/ A-leve

No ademic ac qualific 5 0 ildren 25 20 15 10 GCSE A-C ch All % Low qual. ations No ademic ac qualific 5 0 association between parental education and education parental between association the the lower i.e. problems, health mental child the qualifications, educational parents’ of level qualification held by the parents (or parent if if parent (or parents the by held qualification A.7 provides Table Appendix one present). only the four of each to relating information detailed SDQ subscales. inverse very strong a show the plots All Parental education Parental of the proportions shows 4.4 below Figure problems health mental severe with children academic highest to according analysed types of mental health problems, although to to although problems, health mental of types as externalising for degree pronounced a more problems. internalising to opposed difficulties, parent ratings) by highest parent academic qualification: UK, 2012 qualification: academic parent highest by ratings) parent difficulties, Figure 4.4: Percentages of 11-year-old children with severe mental health problems (total (total problems health mental severe with children 11-year-old of 4.4: Percentages Figure 25 20 15 10 % Boys

Centre for Mental Health REPORT Children of the new century 16 Centre for Mental Health Parental occupation parents, taking the highest ranked current occupation of a couple, where there are two Figure 4.5 below shows the proportions of parents present at child age 11. Detailed children with severe mental health problems supporting information relating to the four SDQ according to the occupational status of their subscales is given in Appendix Table A.8.

Figure 4.5: Percentages of 11-year-old children with severe mental health problems

(total difficulties) by parental occupation at child age 11: parent ratings, UK, 2012 REPORT All Children 30 Children of the new century 25

20

% 15

10

5

0 Workless Routine/ Lower Small EmployerIntermediate Professional/ Semi-Routine Supervisory Managerial Boys

30

25

20

% 15

10

5

0 Workless Routine/ Lower Small Employer Intermediate Professional/ Semi-Routine Supervisory Managerial

Girls 30

25

20

% 15

10

5

0 Workless Routine/ Lower Small EmployerIntermediate Professional/ Semi-Routine Supervisory Managerial

17 al., 2005), although income is measured in measured is al., 2005), although income quintiles. than rather week £ per of ranges of on the basis quintiles to the data Converting in that suggest the figures interpolation, proportion) over the top three bands. Targeting Targeting bands. three the top proportion) over miss would only quintile the lowest at services the cases. of two-thirds 1992) that (Rose, Paradox Rose’s illustrates This not does in a population group risk the highest risk, the of the majority for account necessarily in the prevention of a strategy for calls which a of or instead as well as a whole, as population The groups. effort on high risk of concentration difficulties in total concentration risk of pattern similar and fairly and boys, in girls similar is more are problems Conduct the subscales. for groups income two in the bottom concentrated so; hyperactivity/ less (70%) and the others 60% both have and peer problems inattention income, 40% of the bottom from the cases of 50%. have problems and emotional between on the relationship Information and family problems health mental childhood and in the 2004 Child given also is income et Survey (Green Health Mental Adolescent those in the top quintile and half as large as as large as and half quintile in the top those before, As quintile. in the bottom those among the same broadly is gradient the of the slope analysed and, girls among is it as boys among most is SDQ subscales, the four to according conduct with children among pronounced gender. of irrespective problems, are problems severe of the risks that The fact not does groups income the lower for greater other from absent totally are they that mean seen in the be can As in the population. groups in the middle children some are there figures, severe who display groups income and upper those of many how consider we If problems. group income in each are problems severe with income the proportion than in each (rather that out turns it problems), who have group found are the problems of two-thirds around range. the income of two-fifths in the bottom miss would quintiles two the bottom Targeting (in diminishing spread cases of the one third the middle quintile are twice as large as among among as large as twice are quintile the middle those from families in the top fifth. The fifth. in the top families from those in problems severe with children of proportions across the income scale. For example, 17.0% example, For scale. the income across fifth in the bottom families from children all of by identified are distribution the income of health mental severe having as their parents 4.1% among only with compared problems, with detailed information relating to the four the four to relating information detailed with A.9. Table in Appendix given SDQ subscales between links the strong In line with figures these and income, status occupational in prevalence gradient a steep show again Family income Family severe of the prevalence shows 4.6 Figure difficulties on the total problems health mental income, family to according analysed measure occupational/managerial group, whereas the whereas group, occupational/managerial and 15.2% are girls among rates corresponding SDQ subscale, by Analysed 2.9% respectively. for wider somewhat is gap the prevalence conduct particularly problems, externalising problems. internalising for is it than problems, of severe mental health problems in this this in problems health mental severe of parent-reported of the prevalence Thus group. boys among problems health mental severe with 27.2% compared is parents workless with in the are parents whose boys 5.8% among terms, the difference in prevalence between between prevalence in the difference terms, the of ends opposite at parents with children among the same much is scale occupational among larger is it but girls, among is it as boys percentage in absolute measured when boys prevalence overall the higher points, reflecting problems based on the SDQ total difficulties difficulties total SDQ on the based problems high among as times five about being score among is it as parents workless of the children professional/ in the parents of the children in relative Measured group. managerial between child mental health problems and and problems health mental child between of in the case repeated is education parental parent any of occupation ranked the highest before, 11. As age child at working who was wide, strikingly is in prevalence the gap health mental severe of the prevalence with Given the close links between education and education between links the close Given find to surprising not perhaps is it occupation, above identified association inverse the that

Centre for Mental Health REPORT Children of the new century 18 Centre for Mental Health 2004 the prevalence of severe mental health top quintile. This compares with a more than problems in children aged 11-16 was about fourfold difference among 11-year-olds in 2012, three times as high among those in the bottom suggesting that the income-related gradient in quintile of family income as among those in the the prevalence of severe mental health

Figure 4.6: Percentages of 11-year-old children with severe mental health problems (total difficulties) by quintile of family income: parent ratings, UK, 2012 REPORT All children

20 Children of the new century

16

12 % 8

4

0 Bottom Second Third Fourth Top 20% 20% 20% 20% 20% Boys 20

16

12 %

8

4

0 Bottom Second Third Fourth Top 20% 20% 20% 20% 20%

Girls 20

16

12 % 8

4

0 Bottom Second Third Fourth Top 20% 20% 20% 20% 20%

19

the four-fold difference we have found among among found have we difference the four-fold large this for The reasons children. 11-year-old not have and adults children between difference in the literature or discussed noted been widely furtherand merit research. the combined prevalence of all clinically clinically all of prevalence the combined about is problems health mental diagnosable in the bottom adults higher among 1.5 times among is it than distribution the income 20% of with contrasts This 20%. in the top those Finally, it also appears that the income-related the income-related that appears also it Finally, among steeper much is in prevalence gradient Evidence adults. among is it than children in the 2007 given is the latter to relating Survey Morbidity Psychiatric Adult national that shows al., 2009) and this et (McManus the 1970 national birth cohort with MCS using using MCS birth with cohort the 1970 national measuring to approach different a somewhat score a continuous on based health, mental problems severe for cut-off the than rather here. adopted in externalising and emotional problems problems and emotional in externalising was 2012, in MCS, 11-year-olds observed for measures health mental similar for than steeper al., 2015). et in 1980 (Goodman in 10-year-olds comparing analysis on preliminary based is This problems has become somewhat steeper in steeper somewhat become has problems years. recent gradient an income that evidence also is There

Centre for Mental Health REPORT Children of the new century 20 Centre for Mental Health Chapter 5: Recent trends

Introduction data in the MCS in order to have an equal representation of 10- and 11-year-olds. Socio- This chapter examines recent changes over demographic characteristics of the three time in children’s mental health using parent samples are given in Appendix Table A.10. The and teacher reports from three nationally profiles are fairly similar, though they also representative data sets: the British Child reflect a general trend for fewer two-parent REPORT and Adolescent Mental Health Survey of 1999 families to be married rather than cohabiting

(Meltzer et al., 2000), the repeat of this survey and for a decline in owner occupation. Children of the new century undertaken in 2004 (Green et al., 2005) and the survey of the MCS sample carried out in 2012. Changes in mental health at a Because the 1999 and 2004 surveys did not cover Northern Ireland, parent reports relate to population level children living in Great Britain. Furthermore, as The analyses presented so far in this report the 2012 teacher survey did not cover Scotland, have focused mainly on children with severe teacher reports relate to children living in mental health difficulties. However, in looking England and Wales. at trends over time, there may also be interest More importantly, because the 1999 and 2004 in taking a wider perspective, using SDQ scores surveys covered children of all ages between 5 averaged across the whole population of 10- and 16, the numbers of 11-year-olds in these and 11-year-olds in each of the three samples, surveys are relatively small, reducing the in order to identify whether there has been any likelihood that observed changes over time general improvement or deterioration in the will be found to be statistically significant. To mental health of children in this age group. address this, the samples of children from the Average SDQ scores for total difficulties in 1999 and 2004 surveys have been expanded 1999, 2004 and 2012 are shown in Figure 5.1 to cover 10-year-olds as well as 11-year-olds, (parent ratings) and Figure 5.2 (teacher ratings). taking advantage of the fact that about a third Detailed information relating to the four SDQ of all children in the MCS sample were aged subscales is given in Appendix Tables A.11 10 at the time they were surveyed in 2012. and A.12. In all cases, higher scores represent To minimise the extent of possible age bias, worse mental health. additional weights were applied to the SDQ

Figure 5.1: Average SDQ scores for total difficulties among 10- and 11-year-old children: parent ratings, Great Britain

10 BCAMHS 1999

BCAMHS 2004 8 MCS 2012

6 SDQ scores 4

2

0 All children Boys Girls

21 1999 2004 2012 BCAMHS MCS BCAMHS s Girl Teacher reports, on the other hand, show reports, show on the other hand, Teacher in and girls boys both among improvements exception the main health, mental of areas most problems. emotional being severe of in the prevalence Changes problems health mental children of on the group a focus to Returning changes SDQ symptoms, the worst showing health mental severe of in the prevalence shown 1999 and 2012 are between problems 5.4 and Figure ratings) 5.3 (parent in Figure SDQ on the four data Detailed ratings.) (teacher A.13 and Tables in Appendix given is subscales A.14. s y Bo ildren ch All 8 6 4 2 0 10

Figure 5.2: Average SDQ scores for total difficulties among 10- and 11-year-old and 11-year-old 10- among difficulties total for SDQ scores 5.2: Average Figure Wales and England ratings, teacher children:

SDQ SDQ scores and peer problems from 1999 to 2004 for girls, girls, 2004 for 1999 to from and peer problems 2012. In fact, by lost were improvements these in increase significant a statistically was there 2004. from hyperactivity/inattention girls’ show contrasting results between parents parents between results contrasting show reports, the only On parent and teachers. was 1999 and 2012 between change significant hyperactivity/ for in scores an improvement reports parent While boys. among inattention in hyperactivity/inattention a decrease show and 2012. For all children, and for girls but not not but girls and for children, all and 2012. For statistically also were there on their own, boys sub-period in the improvements significant 2012. 2004 to also SDQ subscales on the four data Detailed The reports by teachers paint a more optimistic optimistic a more paint teachers The reports by significant statistically showing picture, total of score in the average improvements for both boys and children, for all difficulties 1999 between separately, considered and girls comparing 1999 and 2012. There was a was There 1999 and 2012. comparing in average improvement significant statistically particularly difficulties, total for SDQ scores much 1999 and 2004, but between girls, among of this was lost by 2012. Looking first at the data for total difficulties, difficulties, total for data the at first Looking change little suggest parents the reports by

Centre for Mental Health REPORT Children of the new century 22 Centre for Mental Health

Figure 5.3: Percentages of 10- and 11-year-old children with severe mental health problems (total difficulties) in 1999, 2004 and 2012: parent ratings, Great Britain

16 BCAMHS 1999 REPORT BCAMHS 2004 12

MCS 2012 Children of the new century

% 8

4

0 All children Boys Girls

Figure 5.4: Percentages of 10- and 11-year-old children with severe mental health problems (total difficulties) in 1999, 2004 and 2012: teacher ratings, England and Wales

16

BCAMHS 1999 12 BCAMHS 2004 MCS 2012 % 8

4

0 All children Boys Girls

23 the most recent sweep of the MCS (at child age age child (at the MCS of sweep recent the most the with comparison for available 14) become surveys. earlier types. These improvements were observed in were improvements These types. more reports and were and teacher both parent girls. than boys for marked varied the more explain might factors Various and 11-year-old 10- among found picture the impact example for including children, which recession the economic of on families 2008/09. of crash the financial followed on focus be an increased may Another factor which classroom, in the management behaviour between the differences further explain may be also may reports. It and teacher parent children 10- and 11-year-old that the case in the problems health mental fewer show 11 marks 10 to home. Age at than classroom adolescents and early puberty, of the beginning in front behaviours different demonstrate might their teachers to opposed as their parents, of assess to interest be of will It and peers. clear-cut and more the earlier whether or not among health mental improving towards trend from when data been maintained has children among boys. In contrast, the assessments by by the assessments In contrast, boys. among on a wider improvements suggest teachers as well as problems conduct including front, girls to and applying hyperactivity/inattention total for scores The parents’ boys. as well as and non-significant a small suggest difficulties problems, severe of in the prevalence increase a significant indicate ratings teachers’ whereas as boys among particularly prevalence, in fall 2004. from girls as well contrast in is findings these of nature The mixed which study reported those in an earlier to in the children of health the mental compared of that with old years 7 were when they MCS in the 1999 and 2004 age the same of children found This al., 2014). et (Sellers BCAMHS health mental improving of evidence strong (1999-2008), including the period studied over scores problem in average both a decline in the relative and a fall 7-year-olds all among symptom all for problems severe with numbers problems, namely hyperactivity/inattention hyperactivity/inattention namely problems, For example, reports by parents identify only identify only parents reports by example, For between improvement consistent of one area average to relates 1999 and 2012, whether this severe of or the prevalence SDQ scores vary according to the gender of the child, the the child, of the gender to according vary and concerned problem health mental type of parents by provided are whether assessments or teachers. important. particularly is points these of The last suggest that the mental health of 10- and of health the mental that suggest improved country in this children 11-year-old 1999 between deteriorated than rather slightly somewhat however, is, The picture and 2012. perceptions that apparent also is and it mixed prevalence of emotional problems among girls girls among problems emotional of prevalence 2004 and 2012. between Discussion chapter in this out set the findings Overall, the prevalence of severe peer problems for for peer problems severe of the prevalence the assessments 2004. In contrast, from girls widespread more show again teachers by problems conduct covering improvements, among hyperactivity/inattention as well as in the a decline and also and girls both boys the only significant change in parent reports parent in change significant the only in a decline 1999 and 2012 was between relating problems severe of the prevalence particularly and inattention, hyperactivity to in an increase also was There boys. among boys on their own but not to girls. In addition, In addition, girls. to not but their own on boys in all in prevalence falls significant were there 2004 in the sub-period girls including groupings 2012. to SDQ subscales, the individual to Turning separately. separately. the that reports suggest teacher In contrast, problems health mental severe of prevalence 1999 and 2012. between significantly fell to and also a whole as children to applies This or teachers. According to parents’ reports, parents’ to According or teachers. children 11-year-old 10- and the proportion of somewhat increased problems severe with enough to not but 1999 and 2012, between all whether among significance, statistical reach taken or girls boys or among combined children Looking first at the data on total difficulties, difficulties, total on data the at first Looking depending results conflicting find again we parents by made are the ratings on whether

Centre for Mental Health REPORT Children of the new century 24 Centre for Mental Health Chapter 6: Incidence and persistence

Introduction compared with the estimates given elsewhere in this report. For example, the prevalence This chapter provides an analysis of severe of severe problems at age 11 on the total mental health problems among those children difficulties score is 10.3% in the full sample but in the MCS for whom valid data on mental only 8.9% in the longitudinal sample. For the health are available at all four of the surveys purpose of plotting the longitudinal results in REPORT undertaken when the child was aged 3, 5, graphs, we have made a crude adjustment for

7 and 11. We describe this data set as the the missing cases, distributing them pro rata Children of the new century longitudinal sample. From these records we over those with and without severe problems can see how many children ever recorded a at age 11 according to the information we particular type of severe problem between the have about them at that point. It is likely that ages of 3 and 11 (incidence) and also how many children with recurrent severe problems are still with severe problems at age 11 also had these under-represented after this adjustment. problems repeatedly before that (persistence). This is the sort of picture that can only be Incidence provided by longitudinal data. Figure 6.1 provides information on the number A drawback with our longitudinal sample is that of different ages at which children in the the picture has some pieces missing, in the form longitudinal sample were assessed as having of nearly three thousand cases with incomplete severe mental health problems based on the records. We can tell that these exclusions have total difficulties score on parent's reports. A a somewhat different profile from those who more detailed but unadjusted breakdown of the are observed on all occasions, because the longitudinal sample is given in Appendix Table longitudinal sample understates the prevalence A.15. of severe mental health problems at age 11

Figure 6.1: Percentages of children with severe mental health problems (total difficulties) at different ages: parent ratings, UK

No problems

Problems at one age

Problems at two ages

Problems at three ages

Problems at four ages

All Children

Boys Girls

25 s 11 and 2 or 3 other age 11 alone or + one other Some but not age 11 s Girl We have also chosen to present as ‘persistent’ ‘persistent’ as present to chosen also have We at problems severe who had children all cases 11. For age at including surveys or four three surveys, it three at problems severe with these the apparently far how be checked to remains the to close a score had year problem-free be the might above suggested we as threshold, case. the using problems severe of The persistence in Figure summarised is score difficulties total 6.2 below. four SDQ subscales, as shown in Appendix in Appendix shown as SDQ subscales, four higher somewhat is A.16-19. Incidence Tables problems (conduct problems externalising for for is it than and hyperactivity/inattention) problems (emotional problems internalising twice least at always is but and peer problems) of the exception With prevalence. high as as very is pattern the where problems, emotional higher among is incidence gender, by similar measures. on all girls among than boys Persistence a good A.15-19 provide Tables Appendix of on patterns information detailed of deal mental in severe and recurrence persistence Because childhood. during problems health mental on the report is this of focus the main at here look we children, 11-year-old of health among specifically and recurrence persistence who were sample in the longitudinal those age. this at problems severe with assessed These features broadly apply to each of the of each to apply broadly features These s y Bo n ildre ch All 5 0 30 25 20 15 10 parent ratings, UK ratings, parent Figure 6.2: Incidence of severe mental health problems (total difficulties) by age 11: age by difficulties) (total problems health mental severe of 6.2: Incidence Figure % 11, and over three times as many had severe severe had many as times three 11, and over ages. four all at problems all measures. For example, 26.0% of boys boys 26.0% of example, For measures. all to estimated are sample in the longitudinal least at problems severe experienced have as Twice girls. 17.8% of with compared once, severe with assessed were girls as boys many 3 and between ages different three at problems had these problems at two different ages, 2.7% ages, different two at problems these had 1.5% had and just ages three at problems had ages. four all at problems severe mental severe of a higher incidence was There on girls than boys among problems health 7% of all children in the MCS were classified as as classified were in the MCS children all 7% of 3 only. age at problems severe having breaks incidence for 21.9% of The estimate severe had children 12.7% of follows: as down 5.1% only, one age at problems health mental it was noted above that the prevalence of severe severe of the prevalence that above noted was it than less is This 10.3%. 11 was age at problems between for incidence 21.9% of the figure half the excess for reason 3 and 11. A major ages as many as that is prevalence over incidence of health problems in at least one of the four MCS MCS the four one of least in at problems health 11. 3, 5, 7 and ages child at undertaken surveys of the incidence that is this of One implication childhood of the course during problems severe of the prevalence than higher considerably is example, For age. single any at problems these Figure 6.1 indicates that among all children children all among that 6.1 indicates Figure a fifth over just sample in the longitudinal mental severe with assessed were (21.9%)

Centre for Mental Health REPORT Children of the new century 26 Centre for Mental Health This shows that 3.6% of all 11-year-old children are classified as having persistent problems, while a further 6.7% may be described as intermittent cases, referring to those who had severe problems at age 11 only or at age 11 and one other age. A further 11.6% of the sample had a severe rating at least once but not at age 11, while 78.1% had no severe problems at any age. Persistence as we have defined it is REPORT nearly three times as common among boys as among girls (5.1% against 1.8%). The numbers Children of the new century of intermittent cases are also higher among boys than girls although the difference is less pronounced (7.6% against 5.9%). The percentages of 11-year-old children with persistent severe problems for each of the four SDQ subscales are as follows:

Boys Girls All Children Conduct 6.3 3.4 4.9 Problems Hyperactivity/ 8.1 3.0 5.6 inattention Emotional 2.1 2.0 2.1 Problems Peer Problems 4.2 2.3 3.2

These estimates, based on data adjusted for prevalence in incomplete cases, show that a substantial proportion, but not all, of those with severe problems at age 11 have these problems persistently, and that rates of persistence are noticeably higher for externalising problems than they are for internalising problems. They are also roughly twice as high among boys as they are among girls except in the case of emotional problems.

27 . Journal of Journal , 56(1), 49-57. . TSO. London: Mental health of health Mental The Millennium Cohort Cohort The Millennium Millennium Cohort Study: Study: Cohort Millennium The strategy of preventive of preventive strategy The The mental health of children and and of children health The mental . Oxford: Oxford Medical publications. Medical Oxford . Oxford: London: Centre for Longitudinal Longitudinal for Centre London: . London: Palgrave Macmillan. Palgrave . London: . London: Centre for Longitudinal Studies. Longitudinal for Centre . London: Studies. Studies. G. (1992). Rose, medicine A. Thapar, A., Pickles, B., R., Maughan, Sellers, and in parent- Trends (2014) and Collishaw,S. and ADHD conduct emotional, teacher-rated in prepubertal and their impact problems 1999-2008. Britain: in Great children Psychiatry and Psychology Child T. and Goodman, R. (2005) R. (2005) Goodman, and T. Britain, in Great people young and children 2004 T., H., Brugha, S., Meltzer, McManus, (2009) R. (eds.) and Jenkins, P. Bebbington, 2007 in England, morbidity psychiatry Adult and Health for Centre Information NHS London: Care. Social R. and Ford, R., Goodman, H., Gatward, Meltzer, (2000) T. Britain in Great adolescents (2014) T. Mostafa, 5 (age sweep in on response report technical 11) I. (ed.) (2007 Plewis, (4th sampling on report technical Study: edition). , , Journal Journal , 141(4), , 38, 1179-1191. . London: Centre for for Centre . London: , 48(4), 400-403. , 40(11), 1337-1345. Social and emotional skills skills emotional and Social Psychological Bulletin Psychological . Early Intervention Foundation Intervention Foundation . Early Millennium Cohort Study Second Second Study Cohort Millennium References of the Strengths and Difficulties Questionnaire. Questionnaire. and Difficulties Strengths the of and of Child Academy the American of Journal Psychiatry Adolescent H., Ford, A., Meltzer, McGinnity, H., Green, Goodman, R. (1997) The Strengths and Strengths The R. (1997) Goodman, note. research a Questionnaire: Difficulties Psychiatry and Psychology of Child Journal 38(5), 581-586. properties R. (2001) Psychometric Goodman, Strengths and Difficulties Questionnaire as a as Questionnaire and Difficulties Strengths health. mental child of measure dimensional and of Child Academy the American of Journal Psychiatry Adolescent hypothesised five subscales on the Strengths Strengths on the subscales five hypothesised from data (SDQ): Questionnaire and Difficulties and children. teachers parents, British Psychology Child of Abnormal R. (2009) and Goodman, Anna, Goodman, Review, March 2015. London: UCL Institute of of Institute UCL London: 2015. March Review, Education. and Ploubidis, D. Lamping, Anna, Goodman, internalising broader use When to G. (2010) the of instead subscales and externalising Longitudinal Studies. Longitudinal B. and H., Nasim, Joshi, Alissa, Goodman, C. (2015) Tyler, on effects the long-term and in childhood life adult 858-900. I. (2006). Schoon, A., K., Hansen, George, & H. Joshi S. Dex In: Development. Child (Eds.). Report Descriptive Survey. 50(5), 571-580. M. et Wang, T., A., Augenstein, Reyes, Los De the multi-informant of validity The al. (2015) and adolescent child assessing to approach health. mental Collishaw, S., Goodman, R., Ford, T., Rabe- T., R., Ford, S., Goodman, Collishaw, A. far How (2009) Pickles, S. and Hesketh, and family child, between associations are psychopathology and child factors community and informant-general? informant-specific Psychiatry and Psychology of Child Journal

Centre for Mental Health REPORT Children of the new century 28 Centre for Mental Health 2 χ N * 98.01 45.39 p<.001 p<.001 p<.001 1.17 p = .39 p <.001 273.14 118.18 12,798 12,761 12,782 12,790 12,751 p-value 11.5 11.5 11.9 12.5 10.3 Severe REPORT All ChildrenAll 6.8 7.3 8.9 6.9 10.6 Children of the new century Moderate 77.9 81.7 80.8 78.6 82.8 None n 6,322 6,340 6,333 6,336 6,320 8.9 7.1 7.8 12.1 10.8 Severe Girls 9.7 5.5 7.5 8.3 5.7 Moderate 81.4 87.4 80.4 80.9 86.5 None n 6,458 6,439 6,449 6,454 6,431 15.5 11.7 14.2 12.7 13.9 Severe Boys 7.2 9.4 8.0 11.4 8.1 Moderate 2 test probability for comparison of boys and girls. boys of comparison for probability 2 test χ 74.7 76.4 81.1 76.4 79.3 None Peer Peer Problems Total Problems SDQ Conduct Problems Hyper- activity/ inattention Emotional Problems Sample size unweighted. Statistical tests use corrections for sample design. sample for corrections use tests Statistical unweighted. size Sample ratio * Likelihood Appendix Table A.1. Appendix Table of 11-year-old children with mental health problems (parent-rated) Percentages by SDQ subscale: United Kingdom, 2012 29

χ 2 test probability for comparison of boys and girls. and boys of comparison for probability test 2 ratio Likelihood *

Sample size unweighted. Statistical tests use corrections for sample design. sample for corrections use tests Statistical unweighted. size Sample

p<.001

Problems 200.25 200.25

78.2 Total 5.1 90.4 3,538 11.3 10.5 7,085 8.0 7.8 84.2 3,547 4.5

P<.001

Problems 48.00

85.9 Peer 3.6 91.1 3,538 8.5 5.6 7,085 6.9 4.6 88.5 3,547 5.3

P<.47

Problems 1.34

90.8 Emotional 4.2 91.5 3,538 4.8 4.4 7,085 4.6 4.3 91.1 3,547 4.3

inattention P<.001

activity/ activity/ 388.70

78.6 Hyper- 2.1 94.6 3,538 14.8 6.6 7,085 9.1 4.4 86.5 3,547 3.3

P<.001

Problems 208.52

Conduct Conduct 10.5 5.8 83.7 94.5 3,538 3,547 3.5 2.0 7.0 4.0 89.0 7,085

p-value

χ SDQ Severe Moderate None None n n Severe Moderate Severe Moderate None 2 * N

Boys Girls All Children All

Percentages of 11-year-old children with mental health problems (teacher-rated) by SDQ subscale: England and Wales, 2012 Wales, and England subscale: SDQ by (teacher-rated) problems health mental with children 11-year-old of Percentages Appendix Table A.2. A.2. Table Appendix

Centre for Mental Health REPORT Children of the new century 30 Centre for Mental Health Appendix Table A.3. Percentages of 11-year-old children with severe problems (parent-rated) in combinations of SDQ subscales: United Kingdom, 2012

Boys Girls All Children None 68.6 75.4 71.9 REPORT Conduct 3.8 3.3 3.6

Hyperactivity 5.0 2.4 3.7 Children of the new century Emotional 3.2 5.2 4.1 Peer 4.4 4.1 4.3 Hyper/Conduct 2.5 0.8 1.7 Emotional/Peer 1.6 2.1 1.8 Emotional/Hyper 1.0 0.6 0.8 Emotional/Conduct 0.8 0.9 0.9 Hyper/Peer 1.3 0.7 1.0 Conduct/Peer 1.3 0.7 1.0 Emotional/Hyper/Conduct 1.0 0.7 0.8 Emotional/Conduct/Peer 0.8 0.9 0.9 Emotional/Hyper/Peer 1.0 0.7 0.9 Conduct/Hyper/Peer 1.5 0.5 1.0 Emotional/Hyper/Conduct/Peer 2.2 1.0 1.6 N 6,431 6,320 12,751

Note: unweighted valid numbers of observations are presented.

31 7.9 8.1 7.2 5.7 7.7 10.6 10.6 10.4 5.31 12.9 13.7 10.1 12.8 3.92 4.26 (768) (634) (413) (686) (341) (377) (704) (307) 6,431 6,320 12,751 (1,376) (6,644) (6,248) (12,941) Total Total Problems 9.0 9.1 8.9 13.0 12.2 11.5 14.9 13.7 12.8 11.1 10.6 10.0 2.78 (636) (414) (686) (342) (378) (704) (308) (770) 6,336 6,454 12,790 (1,376) (6,673) (6,259) 12.27** (12,981) 14.13*** Peer Peer Problems 9.0 9.4 8.6 12.2 11.3 13.0 5.80 11.4 11.9 12.3 3.08 12.5 11.3 13.7 (769) (636) (413) (686) (342) (377) (704) (308) 8.46* 6,449 6,333 12,782 (1,376) (6,688) (6,255) (12,972) Problems Emotional Emotional Girls Boys All ChildrenAll 11.7 9.1 7.5 8.0 4.7 7.2 12.8 11.9 15.6 17.2 13.5 16.3 2.51 (634) (413) (686) (341) (378) (704) (307) (12,952) 5.90* 6.27* (769) 6,439 6,322 12,761 (1,376) (6,654) (6,249) Inattention Hyperactivity/ Hyperactivity/ 8.6 8.8 11.6 11.3 3.66 10.2 11.4 14.2 13.3 11.9 12.5 10.3 1.94 2.98 10.6 (771) (636) (414) (342) (686) (704) (308) (378) 6,458 6,340 12,798 (1,376) (6,677) (6,261) Conduct Conduct (12,987) Problems ² difference ² difference ² difference Wales England N. Ireland Scotland χ N Unweighted Scotland England Wales N. Ireland χ Scotland N. Ireland χ N Unweighted Unweighted N Unweighted Wales England Appendix Table A.4. Table Appendix by (parent-rated) with severe problems children of 11-year-old Percentages 2012 Kingdom, subscale: United MCS5 and SDQ of residence at country for any aggregation across subgroups that may be required. The sum of these cell sizes exceeds the exceeds sizes cell these of The sum be required. may that subgroups across aggregation any for to numbers up the unweighted gross 5 attrition weights Wave the since size sample unweighted total 2014). (Mostafa, losses sample reflect Note: the numbers in brackets are the weighted sample sizes for each cell; they should form the base the base form should they cell; each for sizes sample the weighted are in brackets numbers the Note:

Centre for Mental Health REPORT Children of the new century 32 Centre for Mental Health Appendix Table A.5. Percentages of 11-year-old children with severe problems (parent-rated) by child ethnicity and SDQ subscale: United Kingdom

Conduct Hyperactivity/ Emotional Peer Total Problems Inattention Problems Problems Problems All Children

White 12.0 11.9 12.0 12.4 10.6 REPORT (13,521) (13,486) (13,481) (13,495) (13,476)

Mixed 12.7 12.9 15.9 13.9 12.3 Children of the new century (553) (551) (553) (553) (551) Indian 3.7 12.5 8.2 10.6 5.0 (287) (285) (287) (287) (285) Pakistani/ 8.2 7.3 11.4 14.0 7.8 Bangladeshi (712) (697) (706) (710) (695) Black 7.5 7.8 8.4 13.2 10.1 (519) (505) (517) (517) (502) Other 5.0 7.4 13.2 17.9 7.9 (202) (201) (202) (202) (201) χ² difference 38.45** 56.31*** 16.09 8.27 15.43 Unweighted N 12,798 12,761 12,782 12,790 12,751 Boys White 14.7 16.2 11.8 14.1 13.2 (6,916) (6,912) (6,910) (6,914) (6,906) Mixed 12.0 14.1 12.5 12.7 10.9 (297) (296) (296) (296) (296) Indian 3.6 16.9 9.5 12.8 4.5 (162) (162) (162) (162) (162) Pakistani/ 9.2 10.8 10.8 15.6 9.3 Bangladeshi (352) (344) (349) (352) (342) Black 10.9 8.4 8.8 15.8 12.9 (286) (273) (284) (284) (271) Other 8.4 10.6 14.0 20.8 9.6 (104) (104) (104) (104) (104) χ² difference 29.90** 22.09* 3.61 4.92 13.75 Unweighted N 6,458 6,439 6,449 6,454 6,431

33 6.9 6.1 6.3 7.8 5.5 (98) 13.8 (231) (354) (255) (124) 13.15 6,320 (6,571) Total Total Problems 7.7 (98) 14.8 8.73 10.0 15.4 12.5 10.6 (232) (359) (256) (126) 6,336 (6,573) Peer Peer Problems 7.9 6.5 (98) 12.4 11.9 12.1 19.9 (232) (256) (125) (357) 18.66 6,333 (6,572) Problems Emotional Emotional Girls 6.1 4.6 4.9 7.3 6.7 (98) 11.6 (354) (231) (255) (124) 6,322 (6,572) 60.71*** Inattention Hyperactive/ Hyperactive/ 3.4 1.4 7.3 9.2 3.9 (98) 13.6 (234) (360) (256) (126) 6,340 23.88* (6,574) Conduct Conduct Problems ² difference any aggregation across subgroups that may be required. The sum of these cell sizes exceeds the total the total exceeds sizes cell these of The sum be required. may that subgroups across aggregation any reflect to numbers up the unweighted gross 5 attrition weights Wave the since size sample unweighted 2014). (Mostafa, losses sample Note: the numbers in brackets are the weighted sample sizes for each cell; they should form the base for for the base form should they cell; each for sizes sample the weighted are in brackets the numbers Note: Appendix Table A.5. continued Table Appendix χ N Unweighted Other Pakistani/ Pakistani/ Bangladeshi Black Mixed Indian White

Centre for Mental Health REPORT Children of the new century 34 Centre for Mental Health Appendix Table A.6. Percentages of 11-year-old children with severe problems (parent-rated) by parents’ partnership status at MCS5 and SDQ subscale: United Kingdom, 2012

Conduct Hyperactivity/ Emotional Peer Total Problems Inattention Problems Problems Problems Girls All children

White 9.2 7.3 12.1 10.6 7.8 Both natural 7.1 8.0 9.9 9.9 6.6 REPORT (6,574) (6,572) (6,572) (6,573) (6,571) parents (9,530) (9,511) (9,520) (9,528) (9,506)

Mixed 13.6 11.6 19.9 15.4 13.8 Lone natural 17.4 15.4 15.1 15.9 15.0 Children of the new century (256) (255) (256) (256) (255) parent (4,130) (4,122) (4,124) (4,127) (4,116) Indian 3.9 6.7 6.5 7.7 5.5 Step and other 19.5 20.0 14.8 17.9 18.1 (126) (124) (125) (126) (124) family types (2,110) (2,099) (2,110) (2,109) (2,097) 2 Pakistani/ 7.3 4.9 11.9 12.5 6.3 χ difference 366.13*** 264.10*** 78.55*** 129.17*** 303.00*** Bangladeshi (360) (354) (357) (359) (354) Unweighted N 12,798 12,761 12,782 12,790 12,751 Black 3.4 6.1 7.9 10.0 6.9 Boys (234) (231) (232) (232) (231) Both natural 8.6 11.0 9.8 11.0 8.2 Other 1.4 4.6 12.4 14.8 6.1 parents (4,844) (4,833) (4,837) (4,843) (4,828) (98) (98) (98) (98) (98) Lone natural 20.5 20.0 13.9 18.0 17.9 χ² difference 23.88* 60.71*** 18.66 8.73 13.15 parent (2,157) (2,154) (2,153) (2,155) (2,149) Unweighted N 6,340 6,322 6,333 6,336 6,320 Step and other 24.0 26.7 15.7 21.1 22.5 family types (1,118) (1,107) (1,118) (1,116) (1,106) χ2difference 228.15*** 168.29*** 35.86*** 87.97*** 186.55*** Unweighted N 6,458 6,439 6,449 6,454 6,431 Girls Both natural 5.6 4.8 9.9 8.8 5.0 parents (4,686) (4,678) (4,683) (4,685) (4,678) Lone natural 14.0 10.4 16.5 13.7 11.9 parent (1,973) (1,968) (1,970) (1,972) (1,967) Step and other 14.3 12.6 13.8 14.4 13.1 family types (992) (992) (992) (992) (992) χ2difference 133.97*** 91.85*** 49.23*** 40.74*** 113.20*** Unweighted N 6,340 6,322 6,333 6,336 6,320

Note: the numbers in brackets are the weighted sample sizes for each cell; they should form the base for any aggregation across subgroups that may be required. The sum of these cell sizes exceeds the total unweighted sample size since the Wave 5 attrition weights gross up the unweighted numbers to reflect sample losses (Mostafa, 2014).

35 9.2 5.5 3.9 6.9 5.5 17.5 10.6 11.0 16.3 16.4 12.8 12.1 15.2 20.8 (846) (593) (435) (537) (821) (310) Total Total 12,301 (3,717) (1,078) (4,062) (1,614) (1,940) (2,122) (3,106) (1,544) Problems 288.08*** 166.47*** 6,196 9.7 8.1 8.4 8.9 18.0 12.8 13.7 18.7 16.0 14.7 11.5 10.7 16.7 21.7 (846) (435) (596) (537) (821) (311) 12,338 (3,721) (1,079) (4,066) (1,614) (1,943) (2,124) (3,136) (1,568) 196.81*** 109.50*** 6,218 Peer Problems Peer 8.1 7.7 7.8 7.9 16.7 12.0 12.0 12.6 16.4 11.2 11.3 12.3 13.7 17.7 (846) (435) (595) (537) (821) (310) 12,330 (3,718) (1,078) (4,064) (1,614) (1,940) (2,124) (3,133) (1,566) 81.78*** 134.63*** 6,213 Emotional Problems Emotional Boys All children All 7.3 5.4 7.6 18.7 12.8 11.7 11.5 23.0 17.4 15.0 16.2 10.5 15.8 20.5 (846) (435) (595) (537) (821) (312) 12,311 (3,719) (1,078) (4,065) (1,614) (1,942) (2,125) (3,111) (1,549) 179.47*** 108.07*** Inattention Hyperactivity/ 6,204 9.7 6.7 2.9 8.1 2.6 18.5 11.8 14.8 18.6 19.1 14.4 12.0 19.0 22.7 (847) (597) (435) (537) (821) (312) 12,346 (3,721) (1,079) (4,067) (1,614) (1,943) (2,126) (3,139) (1,570) Conduct Conduct 346.11*** 219.77*** Problems 6,222 difference difference 2 2 GCSE at grade D-G GCSE at grade A*-C and AS A-level Degree/diploma Degree/diploma Higher degree Other overseas qualifications GCSE at grade D-G GCSE at grade A*-C No academic No academic qualifications χ N Unweighted AS and AS A-level Degree/diploma Higher degree χ N Unweighted Other overseas Other overseas qualifications No academic qualifications Appendix Table A.7. Table Appendix by highest (parent-rated) with severe problems children of 11-year-old Percentages 2012 Kingdom, subscale: United and SDQ academic qualification parent

Centre for Mental Health REPORT Children of the new century 36 Centre for Mental Health

Conduct Hyperactivity/ Emotional Peer Total Problems Inattention Problems Problems Problems

Girls GCSE at grade 17.7 14.1 22.6 20.2 18.6 REPORT D-G (412) (411) (411) (411) (411)

GCSE at grade 8.8 7.8 12.8 10.7 8.2 Children of the new century A*-C (1,978) (1,977) (1,977) (1,978) (1,977) AS and 7.4 7.3 11.7 7.9 6.3 A-level (542) (541) (541) (542) (541) Degree/di- 5.1 3.7 8.5 7.1 4.0 ploma (1,941) (1,941) (1,941) (1,941) (1,940) Higher degree 3.2 3.2 7.4 6.1 2.2 (793) (793) (793) (793) (793) Other overseas 10.2 6.8 11.4 10.4 6.3 qualifications (285) (284) (285) (285) (284) No academic 14.4 9.5 15.0 15.6 11.8 qualifications (1,570) (1,562) (1,567) (1,569) (1,561) 2difference 140.27*** 83.16*** 81.44*** 106.44*** 153.27*** Unweighted N 6,124 6,107 6,117 6,120 6,105

Notes:

(i) The numbers in brackets are the weighted sample sizes for each cell; they should form the base for any aggregation across subgroups that may be required. The sum of these cell sizes exceeds the total unweighted sample size since the Wave 5 attrition weights gross up the unweighted numbers to reflect sample losses (Mostafa, 2014).

(ii) ‘Other overseas qualifications’ are included with ‘GCSE at grade D-G’ as ‘Low qualifications’ in Figure 4.4.

37 4.4 6.3 5.7 5.8 7.2 7.6 11.0 10.7 21.5 13.4 12.6 27.2 (871) (454) (979) (991) Total (1926) 5,956 11,801 (3,381) (1,937) (3,753) (1,739) (1,919) (2,558) (1,344) 468.97*** 312.32*** Problems 7.9 9.6 9.8 9.3 13.2 11.8 22.7 11.6 11.7 15.4 13.0 25.3 (872) (979) (992) (455) Peer (1930) 5,975 11,835 (3,382) (1,939) (3,770) (1,740) (1,934) (2,573) (1,353) 281.72*** 148.74*** Problems 7.7 8.4 9.3 7.4 8.7 9.0 12.5 12.5 19.5 12.8 12.4 18.8 (873) (979) (992) (455) (1928) 5,972 11,829 (3,382) (1,937) (3,769) (1,740) (1,934) (2,567) (1,348) 92.95*** Problems 194.54*** Emotional Boys All children All 6.8 8.9 8.4 9.6 10.1 11.4 20.6 11.9 11.4 14.5 15.7 27.6 (871) (979) (992) (454) (1927) 5,964 11,811 (3,382) (1,937) (1,740) (3,756) (1,922) (2,564) (1,349) 264.38*** 179.01*** Inattention Hyperactivity/ 5.3 7.3 6.5 6.0 8.3 9.4 12.5 11.1 24.8 15.5 12.5 30.3 (873) (979) (992) (455) (1930) 5,979 11,843 (3,383) (1,939) (3,772) (1,741) (1,936) (2,575) (1,354) Conduct Conduct 553.42*** 363.35*** Problems difference difference 2 2 Professional/ Managerial Intermediate Employer Lower Supervisory Small Professional/ Managerial Intermediate Routine Workless χ N Unweighted Routine/ Semi- Routine/ Semi- Routine Small Employer Lower Supervisory χ N Unweighted Workless 2012 Appendix Table A.8. Table Appendix by highest (parent-rated) with severe problems children of 11-year-old Percentages United Kingdom, and SDQ subscale: (if working) at MCS5 held by parents occupation

Centre for Mental Health REPORT Children of the new century 38 Centre for Mental Health

Conduct Hyperactivity/ Emotional Peer Total Conduct Hyperactivity/ Emotional Peer Total Problems Inattention Problems Problems Problems Problems Inattention Problems Problems Problems

Girls REPORT All children Professional/ 4.5 3.8 8.1 6.4 2.9 Professional/ 5.3 6.8 7.7 7.9 4.4

Managerial (1,642) (1,642) (1,642) (1,642) (1,642) Children of the new century Managerial (3,383) (3,382) (3,382) (3,382) (3,381) Intermediate 6.3 5.9 8.0 7.5 5.3 Intermediate 7.3 8.9 8.4 9.6 6.3 (960) (959) (959) (960) (958) (1,939) (1,937) (1,937) (1,939) (1,937) Small 3.5 5.2 9.5 7.9 3.6 Small 6.5 8.4 9.3 9.8 5.7 Employer (938) (935) (936) (938) (935) Employer (1930) (1927) (1928) (1930) (1926) Lower 9.2 5.3 12.3 10.8 8.6 Lower 12.5 10.1 12.5 13.2 11.0 Supervisory (418) (417) (418) (417) (417) Supervisory (873) (871) (873) (872) (871) Routine/ 9.6 7.0 12.5 10.5 8.8 Routine/ 11.1 11.4 12.5 11.8 10.7 Semi-Routine (1,837) (1,834) (1,835) (1,836) (1,834) Semi- (3,772) (3,756) (3,769) (3,770) (3,753) Routine Workless 18.7 12.9 20.2 19.9 15.2 (1,221) (1,215) (1,218) (1,219) (1,214) Workless 24.8 20.6 19.5 22.7 21.5 (2,575) (2,564) (2,567) (2,573) (2,558) χ2difference 194.38*** 83.69*** 103.11*** 132.95*** 158.20*** χ2difference 553.42*** 264.38*** 194.54*** 281.72*** 468.97*** Unweighted N 5,864 5,847 5,857 5,860 5,845 Unweighted N 11,843 11,811 11,829 11,835 11,801 Boys Notes: Professional/ 6.0 9.6 7.4 9.3 5.8 (i) The numbers in brackets are the weighted sample sizes for each cell; they should form the base for Managerial (1,741) (1,740) (1,740) (1,740) (1,739) any aggregation across subgroups that may be required. The sum of these cell sizes exceeds the total Intermediate 8.3 11.9 8.7 11.6 7.2 unweighted sample size since the Wave 5 attrition weights gross up the unweighted numbers to reflect (979) (979) (979) (979) (979) sample losses (Mostafa, 2014). Small 9.4 11.4 9.0 11.7 7.6 (ii) This is not a standard social class variable as the workless are not classified to their previous Employer (992) (992) (992) (992) (991) occupation. Lower 15.5 14.5 12.8 15.4 13.4 Supervisory (455) (454) (455) (455) (454) Routine/ 12.5 15.7 12.4 13.0 12.6 Semi- (1,936) (1,922) (1,934) (1,934) (1,919) Routine Workless 30.3 27.6 18.8 25.3 27.2 (1,354) (1,349) (1,348) (1,353) (1,344) χ2difference 363.35*** 179.01*** 92.95*** 148.74*** 312.32*** Unweighted N 5,979 5,964 5,972 5,975 5,956

39 6.9 4.3 2.6 8.3 5.5 4.1 9.6 6.6 5.6 12.0 12.2 17.0 15.6 21.8 18.7 Total 6,320 6,431 12,751 (1,743) (1,588) (1,578) (1,513) (1,594) (1,657) (1,509) (1,421) (1,456) (3,255) (3,399) (3,097) (3,000) (2,969) (1,661) Problems 135.71*** 251.29*** 379.78*** 9.6 8.6 6.2 6.4 9.8 8.6 7.5 19.6 10.7 16.5 13.8 10.0 19.0 16.8 21.3 6,336 6,454 12,790 (1,745) (1,599) (1,581) (1,513) (1,601) (1,659) (1,511) (1,422) (1,456) (3,279) (3,404) (3,110) (3,003) (2,969) (1,678) 17.82*** 154.09*** 252.31*** Peer Problems Peer 9.7 9.6 7.2 9.4 6.9 9.5 7.1 14.2 17.1 13.6 12.8 17.0 13.9 11.2 17.0 6,333 6,449 12,782 (1,746) (1,598) (1,581) (1,513) (1,599) (1,659) (1,511) (1,421) (1,456) (3,269) (3,405) (3,109) (3,002) (2,969) (1,670) 72.95*** 77.95*** 145.35*** Problems Emotional Girls Boys All children All 9.7 6.4 5.4 2.4 9.9 7.9 6.1 10.1 19.8 13.2 10.2 11.1 17.2 15.6 23.8 6,322 6,439 12,761 (1,578) (1,513) (1,422) (1,456) (3,000) (2,969) (1,747) (1,588) (1,595) (1,657) (1,509) (3,262) (3,404) (3,097) (1,667) 96.69*** 152.83*** 234.05*** Inattention Hyperactivity/ 4.2 2.7 4.6 6.5 5.4 3.7 6.7 8.9 15.9 13.5 25.5 20.3 10.9 20.8 17.0 6,340 6,458 12,798 (1,511) (1,422) (1,456) (1,581) (1,513) (1,604) (1,660) (1,678) (1,748) (1,599) (3,282) (3,407) (3,110) (3,003) (2,969) Conduct Conduct 212.48*** 352.36*** 561.01*** Problems difference difference difference 2 2 2 Top χ Fourth χ Fourth Top Top χ Fourth Unweighted N Unweighted Second Third Unweighted N Unweighted Bottom Third Third Bottom Second Unweighted N Unweighted Second Third Bottom Note: the numbers in brackets are the weighted sample sizes for each cell; they should form the form should they cell; each for sizes sample the weighted are in brackets numbers the Note: sizes cell these of The sum be required. may that subgroups across aggregation any for base up the gross 5 attrition weights Wave the since size sample unweighted the total exceeds 2014). (Mostafa, losses sample reflect to numbers unweighted Appendix Table A.9. Table Appendix by (parent-rated) with severe problems children of 11-year-old Percentages Kingdom, 2012 United and SDQ subscale: at MCS5 income quintiles equivalised

Centre for Mental Health REPORT Children of the new century 40 Centre for Mental Health Appendix Table A.10. Socio-demographic characteristics of three study samples of children aged 10 and 11: Great Britain

BCAMHS BCAMHS MCS 1999 2004 2012 (N=1904) (N=1348) (N=11,397) Child Age REPORT 10 50.6 50.6 34.0

11 49.4 49.4 66.0 Children of the new century Child Gender Female 49.8 47.1 50.0 Male 50.2 52.9 50.0 Family Composition Married 72.0 67.5 58.2 Single 22.1 23.9 23.8 Cohabiting 5.9 8.6 18.0 Ethnicity White 91.3 86.6 87.7 Black 1.5 2.4 2.1 Indian/Pakistani/Bangladeshi 4.2 6.3 5.4 Other 3.0 4.7 4.8 Country England 85.8 85.0 86.4 Scotland 8.9 9.1 8.5 Wales 5.3 5.9 5.1 Housing Tenure Own 70.1 68.4 65.0 Rent 29.9 31.6 35.0 Numbers of children in the household 1 Child 16.3 19.5 12.2 2 Children 50.1 48.7 45.8 3 Children 23.1 20.6 26.9 4+ Children 10.5 11.2 15.1

Notes:

(i) Percentages are based on weighted data. Sample sizes represent approximate unweighted number of observations, as exact numbers vary slightly due to valid data available for each socio-demographic variable.

(ii) MCS counts more children in the household than did BCAMHS where an upper age limit of 16 is imposed. The contrast between the surveys of family size is therefore likely to be an artefact rather than a trend.

41

*Significant effects only. Cohen’s d is an effect size used to indicate the standardised difference between two means. two between difference standardised the indicate to used size effect an is d Cohen’s only. effects *Significant

MCS SDQ data were weighted so that 10- and 11-year-old children represented equivalent proportions in all analyses. all in proportions equivalent represented children 11-year-old and 10- that so weighted were data SDQ MCS

Note: the approximate sample sizes are 957 = boys, 947 = girls for 1999; 706 = boys, 642 = girls for 2004; and 5,748 = boys, 5,649 = girls for 2012. 2012. for girls = 5,649 boys, = 5,748 and 2004; for girls = 642 boys, = 706 1999; for girls = 947 boys, = 957 are sizes sample approximate the Note:

8.14-8.65 8.39(5.75) Total .08 1>2 8.03-8.39 8.21(6.13) 7.61-8.26 7.93(6.04)

Girls 6.59-7.45 7.02(5.52) 7.37-8.07 7.72(5.44) .13 1>2 7.22-7.68 7.45(5.66)

Boys 8.29-9.23 8.76(6.37) 8.68-9.44 9.06(5.97) 8.69-9.16 8.93(6.45)

Total Problems Total

Total 1.32-1.49 1.41(1.65) 1.44-1.59 1.51(1.71) 1.42-1.52 1.47(1.77)

Girls 1.13-1.36 1.24(1.45) 1.34-1.54 1.44(1.61) .12 1>2 1.28-1.41 1.35(1.65)

Boys 1.42-1.69 1.55(1.80) 1.47-1.70 1.59(1.80) 1.50-1.63 1.57(1.86)

Peer Problems Peer

Total 1.87-2.09 1.98(2.05) 1.90-2.08 1.99(2.05) 1.88-1.99 1.94(2.02)

Girls 1.82-2.07 1.96(1.99) 1.99-2.26 2.13(2.09) 1.90-2.06 1.99(2.03)

Boys 1.82-2.14 1.98(2.00) 1.73-1.98 1.86(2.01) 1.83-1.97 1.90(2.03)

Emotional Problems Emotional

Total 2.98-3.26 3.12(2.61) 3.29-3.53 3.41(2.62) .11 1>2 3.22-3.36 3.29(2.51)

Girls 2.36-2.73 2.55(2.36) 2.70-3.01 2.85(2.41) .13, .09 .13, >2 1,3 2.69-2.89 2.79(2.30)

Boys 3.45-3.85 3.65(2.71) 3.79-4.13 3.96(2.72) .12, .08 .12, 3 1>2, 3.67-3.86 3.76(2.61)

inattention

Hyperactivity/ Hyperactivity/

Total 1.48(1.66) 1.43(1.72) 1.40-1.55 1.52(1.68) 1.34-1.52 1.47-1.57

Girls 1.30(1.55) 1.25(1.53) 1.20-1.40 1.33(1.50) 1.13-1.37 1.27-1.39

Boys 1.66(1.75) 1.59(1.87) 1.54-1.77 1.69(1.82) 1.45-1.73 1.63-1.76

Conduct Problems Conduct

Mean(SD) Mean(SD) Mean(SD) (p < .05)* < (p

1999 2004 Interval 2012 Interval Difference Interval * d

SDQ BCAMHS Confidence Confidence BCAMHS BCAMHS Confidence Confidence MCS Confidence Cohort Cohen’s Cohen’s

Parent-rated means and standard deviations of SDQ scores for 10- and 11-year-olds in 1999, 2004 and 2012: Great Britain Britain Great 2012: and 2004 1999, in 11-year-olds and 10- for scores SDQ of deviations standard and means Parent-rated Appendix Table A 11. 11. A Table Appendix

Centre for Mental Health REPORT Children of the new century 42 Centre for Mental Health

*

d

.10 .09 .09

.11 .21 .14

.10

.11

.08, .08

.15, .10

.11, .09

Cohen’s Cohen’s * d .11 .12 .13 .08 .12, .08 .13, .09 Cohen’s Cohen’s REPORT

1>3 1>3 1>3

1>3 1>3 1>3

1>3

1>3

Cohort

1,2>3

1,2 >3

1, 2>3

(t-test)*

Difference Children of the new century 1>2 1>2 1>2 1>2 1>2, 3 1,3 >2 Cohort (p < .05)* Difference

Interval .40-.51 .72-.82

.98-1.16

.94-1.10

3.16-3.46 1.47-1.67 2.35-2.55

1.34-1.54 1.36-1.56 1.38-1.52

1.33-1.53

Confidence 1.15-1.29

6.91-7.57 4.23-4.73

5.66-6.08 Interval 1.63-1.76 1.27-1.39 1.47-1.57 3.67-3.86 2.69-2.89 3.22-3.36 1.83-1.97 1.90-2.06 1.88-1.99 1.50-1.63 1.28-1.41 1.42-1.52 8.69-9.16 7.22-7.68 8.03-8.39 Confidence

MCS

2012

Mean(SD)

.44(1.10) .76(1.50)

1.07(1.74)

3.31(2.82) 1.57(1.95) 2.45(2.57)

1.44(1.91) 1.46(1.92) 1.45(1.91)

1.43(1.85) 1.02(1.59) 1.22(1.74)

7.24(6.28) 4.48(4.78)

5.88(5.76) MCS 2012 Mean(SD) 1.69(1.82) 1.33(1.50) 1.52(1.68) 3.76(2.61) 2.79(2.30) 3.29(2.51) 1.90(2.03) 1.99(2.03) 1.94(2.02) 1.57(1.86) 1.35(1.65) 1.47(1.77) 8.93(6.45) 7.45(5.66) 8.21(6.13)

Interval .40-.60 .76-.96

1.02-1.35

3.14-3.66 1.65-2.07 2.50-2.85

1.32-1.68 1.35-1.73 1.39-1.65

1.35-1.70 1.02-1.33 Confidence 1.24-1.48

7.01-8.21 4.58-5.57

6.02-6.82 Interval 1.45-1.73 1.13-1.37 1.34-1.52 3.45-3.85 2.36-2.73 2.98-3.26 1.82-2.14 1.82-2.07 1.87-2.09 1.42-1.69 1.13-1.36 1.32-1.49 8.29-9.23 6.59-7.45 7.61-8.26 Confidence Confidence

2004

BCAMHS

.86(1.62)

Mean(SD)

.50(1.13)

1.18(1.90)

3.40(2.96) 1.86(2.30) 2.67(2.77)

1.50(2.06) 1.54(2.06) 1.52(2.06)

1.53(2.00) 1.17(1.76) 1.36(1.85)

7.61(6.85) 5.08(5.33)

6.42(6.31) 2004 BCAMHS BCAMHS Mean(SD) 1.59(1.87) 1.25(1.53) 1.43(1.72) 3.65(2.71) 2.55(2.36) 3.12(2.61) 1.98(2.00) 1.96(1.99) 1.98(2.05) 1.55(1.80) 1.24(1.45) 1.41(1.65) 8.76(6.37) 7.02(5.52) 7.93(6.04)

Interval .48-.68

.83-1.01

1.12-1.41

3.42-3.86 1.86-2.21 2.68-2.97

1.28-1.57 1.25-1.54 1.31-1.51

1.47-1.76 1.01-1.24 Confidence 1.29-1.44

7.47-8.43 4.74-5.53

6.21-6.85 Interval 1.54-1.77 1.20-1.40 1.40-1.55 3.79-4.13 2.70-3.01 3.29-3.53 1.73-1.98 1.99-2.26 1.90-2.08 1.47-1.70 1.34-1.54 1.44-1.59 8.68-9.44 7.37-8.07 8.14-8.65 Confidence Confidence

1999

BCAMHS

Mean(SD)

.58(1.36) .92(1.72)

1.27(1.97)

3.64(2.99) 2.04(2.35) 2.83(2.81)

1.43(1.92) 1.39(1.94) 1.41(1.93)

1.62(1.95) 1.12(1.57) 1.37(1.79)

7.95(6.57) 5.13(5.42)

6.53(6.18) 1999 BCAMHS Mean(SD) 1.66(1.75) 1.30(1.55) 1.48(1.66) 3.96(2.72) 2.85(2.41) 3.41(2.62) 1.86(2.01) 2.13(2.09) 1.99(2.05) 1.59(1.80) 1.44(1.61) 1.51(1.71) 9.06(5.97) 7.72(5.44) 8.39(5.75)

SDQ

Conduct ProblemsConduct Boys Girls Total Hyperactivity/ inattention Boys Girls Total Emotional Problems Boys Girls Total SDQ Problems Conduct Boys Girls Total Hyperactivity/ inattention Boys Girls Total Problems Emotional Boys Girls Total Problems Peer Boys Girls Total Problems Total Boys Girls Total Peer Problems Boys Girls Total Total Problems Total Boys Girls

Total Appendix Table 12. Appendix Table 2004 and 2012: England Wales means and standard deviations of SDQ scores for 10- 11-year-olds in 1999, Teacher-rated Note: the approximate sample sizes are 957 = boys, 947 = girls for 1999; 706 = boys, 642 = girls for 2004; and 5,748 = boys, 5,649 = girls for 2012. for 5,649 = girls 2004; and 5,748 = boys, for 642 = girls 1999; 706 = boys, for 947 = girls 957 = boys, are sizes sample the approximate Note: analyses. in all proportions equivalent represented children 10- and 11-year-old so that weighted were SDQ data MCS means. two between difference standardised the to indicate used size an effect is d Cohen’s only. effects *Significant

Note: are 711 sample sizes = boys, 722 = girls for 1999; 509 = boys, 453 = girls; and 3,530 = boys, 3,541 = girls forSDQ 2012. data were MCS weighted so 10- that and represented children 11-year-old equivalent proportions inanalyses. all *Significant effects only. Cohen’s *Significant is an dused size effect effectsto indicateonly. thestandardised difference between two means. 43 1>3 2<3 1>2, 3 Cohort (p < .05) Difference 8.0 8.9 7.7 MCS 11.9 13.9 10.9 12.8 13.1 10.6 14.3 11.7 15.4 11.8 11.6 12.0 2012 8.3 6.2 9.8 8.5 7.3 11.9 14.6 11.6 13.0 15.6 12.2 15.7 11.7 12.4 11.4 2004 BCAMHS BCAMHS 6.9 9.4 8.6 9.4 12.3 14.9 10.9 12.9 11.9 13.7 11.2 19.2 14.4 11.1 13.5 1999 BCAMHS Girls Total Girls Total Problems Total Boys Boys Girls Total Problems Peer Boys Boys Girls Total Problems Emotional Boys Girls Total inattention Hyperactivity/ SDQ Problems Conduct Appendix Table A.13. Table Appendix scores in SDQ with severe parent-rated 11-year-olds of 10- and Percentages Great Britain and 2012: 1999, 2004 and 5,748 = boys, 5,649 = girls for 2012. MCS SDQ data were weighted so that 10- and 11-year-old children children 10- and 11-year-old so that weighted were SDQ data 2012. MCS for = girls 5,649 = boys, and 5,748 analyses. in all proportions equivalent represented Note: the approximate sample sizes are 957 = boys, 947 = girls for 1999; 706 = boys, 642 = girls for 2004; 2004; for = girls 642 = boys, 706 1999; for 947 = girls = boys, 957 are sizes sample approximate the Note:

Centre for Mental Health REPORT Children of the new century 44 Centre for Mental Health Appendix Table A.14. Percentages of 10 and 11-year-olds with severe teacher-rated SDQ scores in 1999, 2004 and 2012: England and Wales

SDQ BCAMHS BCAMHS MCS Cohort 1999 2004 2012 Difference (p< .05) Conduct Problems REPORT Boys 13.5 12.5 10.5 1>3

Girls 5.7 3.3 3.2 1>3 Children of the new century Total 9.6 8.1 6.9 1>3 Hyperactivity Boys 19.5 17.7 15.2 1>3 Girls 6.9 5.9 3.3 1, 2>3 Total 13.2 12.1 9.4 1, 2>3 Emotional Problems Boys 4.8 5.8 4.8 Girls 4.6 6.6 4.2 2>3 Total 4.7 6.2 4.5 2>3 Peer Problems Boys 9.7 9.8 8.9 Girls 5.0 4.8 5.2 Total 7.3 7.5 7.1 Total Problems Boys 15.0 16.0 11.5 1, 2>3 Girls 5.7 6.6 4.3 2>3 Total 10.2 11.6 8.0 1, 2>3

Note: sample sizes are 711 = boys, 722 = girls for 1999; 509 = boys, 453 = girls; and 3,530 = boys, 3,541 = girls for 2012. MCS SDQ data were weighted so that 10- and 11-year-old children represented equivalent proportions in all analyses.

45 0.8 0.6 1.3 7.1 0.7 1.2 3.3 0.9 1.0 0.3 0.4 1.0 1.1 0.6 0.4 79.3 9,595 (11,344) All All children 0.5 0.3 0.6 6.5 0.6 0.6 3.1 0.7 0.8 0.2 0.2 1.0 1.0 0.6 0.2 83.1 Girls 4,798 (5,586) 1.0 1.0 2.0 7.7 0.8 1.8 3.5 1.2 1.3 0.3 0.5 1.0 1.3 0.7 0.5 75.4 Boys 4,797 (5,758) None Age 3/Age 5/Age 7/Age 11 7/Age 5/Age 3/Age Age N Age 3/Age 5/Age 7 5/Age 3/Age Age 11 5/Age 3/Age Age 11 7/Age 3/Age Age 11 7/Age 5/Age Age Age 7/Age 11 7/Age Age 7 5/Age Age 11 5/Age Age 7 3/Age Age 11 3/Age Age Age 5 Age 7 Age 11 Age 5 3/Age Age Age 3 Age counts counts are presented in brackets, showing the data on which each percentage this. below given is is observations number of unweighted based; the Note: this table is based on cases with valid data in the longitudinal sample. Weighted Weighted sample. in the longitudinal data valid with on cases based is table this Note: Appendix Table A.15. Table Appendix at parent ratings) (total difficulties, problems with severe of MCS children Percentages to 11: United Kingdom ages 3

Centre for Mental Health REPORT Children of the new century 46 Centre for Mental Health Appendix Table A.16. Percentages of MCS children with severe conduct problems (parent ratings) at ages 3 to 11: United Kingdom

Boys Girls All children None 66.7 74.1 70.3 Age 3 9.4 10.7 10.0

Age 5 2.9 1.5 2.2 REPORT Age 7 2.1 1.5 1.8 Children of the new century Age 11 3.2 2.4 2.8 Age 3/Age 5 1.9 1.7 1.8 Age 7/Age 11 1.4 .7 1.1 Age 5/Age 7 1.1 .7 .9 Age 5/Age 11 .7 .3 .5 Age 3/Age 7 1.7 1.3 1.5 Age 3/Age 11 1.7 1.3 1.5 Age 3/Age 5/Age 7 1.4 .9 1.2 Age 3/Age 5/Age 11 .7 .7 .7 Age 3/Age 7/Age 11 1.4 .7 1.0 Age 5/Age 7/Age 11 1.0 .3 .7 Age 3/Age 5/Age 7/Age 11 2.7 1.2 2.0 N (5,941) (5,739) (11,680) 4,959 4,932 9,891

Note: this table is based on cases with valid data in the longitudinal sample. Weighted counts are presented in brackets, showing the data on which each percentage is based; the unweighted number of observations is given below this.

47 .9 .5 .5 .4 1.5 2.3 6.4 2.0 3.1 3.0 1.6 1.4 1.2 1.1 1.2 72.9 9,760 (11,539) All All children .5 .8 .9 .2 .3 .8 .3 1.1 5.9 1.9 2.3 2.4 1.3 1.2 1.0 79.1 Girls 4,869 (5,667) .7 .7 .5 1.3 2.1 3.5 7.0 2.2 3.9 3.5 1.9 1.9 1.2 1.1 1.6 66.9 Boys 4,891 (5,872) N Age 3/Age 5/Age 7 5/Age 3/Age Age 11 5/Age 3/Age Age 11 7/Age 3/Age Age 11 7/Age 5/Age Age 11 7/Age 5/Age 3/Age Age Age 5/Age 7 5/Age Age 11 5/Age Age 7 3/Age Age 11 3/Age Age Age 7 Age 11 Age 5 3/Age Age 11 7/Age Age None 3 Age 5 Age number of observations is given below this. this. below given is observations number of Note: this table is based on cases with valid data in the longitudinal sample. Weighted counts counts Weighted sample. in the longitudinal data valid with on cases based is table this Note: the unweighted based; is percentage each on which the data showing in brackets, presented are Appendix Table A.17. Table Appendix problems hyperactivity/inattention with severe of MCS children Percentages Kingdom 3 to 11: United ratings) at ages (parent

Centre for Mental Health REPORT Children of the new century 48 Centre for Mental Health Appendix Table A.18. Percentages of MCS children with severe emotional problems (parent ratings) at ages 3 to 11: United Kingdom

Boys Girls All children None 78.0 78.2 78.1 Age 3 4.4 3.9 4.1

Age 5 1.4 1.8 1.6 REPORT Age 7 3.1 1.8 2.5 Children of the new century Age 11 5.6 6.9 6.3 Age 3/Age 5 .6 .6 .6 Age 7/Age 11 1.7 1.7 1.7 Age 5/Age 7 .7 .6 .6 Age 5/Age 11 .5 .4 .4 Age 3/Age 7 .5 .8 .7 Age 3/Age 11 1.1 .9 1.0 Age 3/Age 5/Age 7 .4 .4 .4 Age 3/Age 5/Age 11 .2 .5 .3 Age 3/Age 7/Age 11 .3 .5 .4 Age 5/Age 7/Age 11 .9 .5 .7 Age 3/Age 5/Age 7/Age 11 .6 .5 .6 N (5,913) (5,709) (11,622) 4,933 4,908 9,841

Note: this table is based on cases with valid data in the longitudinal sample. Weighted counts are presented in brackets, showing the data on which each percentage is based; the unweighted number of observations is given below this.

49 .6 .7 .5 .8 .6 .4 1.0 1.0 6.4 1.9 2.9 5.2 1.0 1.6 1.2 74.2 9,795 (11,579) All children All .6 .6 .5 .7 .3 .8 .3 .4 5.7 1.7 2.6 5.0 1.0 1.2 1.3 77.3 Girls 4,899 (5,698) .6 .7 .7 .8 .8 .4 1.3 1.5 7.1 2.2 3.1 5.3 1.0 2.0 1.2 71.3 Boys 4,896 (5,881) N Age 3/Age 5/Age 7 5/Age 3/Age Age 11 5/Age 3/Age Age 11 7/Age 3/Age Age 11 7/Age 5/Age Age 11 7/Age 5/Age 3/Age Age Age 5/Age 7 5/Age Age 11 5/Age Age 7 3/Age Age 11 3/Age Age Age 7 Age 11 Age 5 3/Age Age 11 7/Age Age None 3 Age 5 Age number of observations is given below this. this. below given is observations number of Note: this table is based on cases with valid data in the longitudinal sample. Weighted counts counts Weighted sample. in the longitudinal data valid with on cases based is table this Note: the unweighted based; is percentage each on which the data showing in brackets, presented are Appendix Table A.19. Table Appendix 3 ratings) at ages (parent peer problems with severe of MCS children Percentages Kingdom to 11: United

Centre for Mental Health REPORT Children of the new century 50 Centre for Mental Health REPORT Children of the new century

51 Children of the new century

Published November 2015 Photograph: istockphoto.com/OJO_Images

£10 where sold © Centre for Mental Health, 2015 Recipients (journals excepted) are free to copy or use the material from this paper, provided that the source is appropriately acknowledged. Register for our emails at www.centreformentalhealth.org.uk

Centre for Mental Health 134-138 Borough High Street, London SE1 1LB Tel 020 7827 8300 Fax 020 7827 8369 www.centreformentalhealth.org.uk Charity registration no. 1091156. A company limited by guarantee registered in England and Wales no. 4373019.