Children of the new century: mental health findings from the Millennium Cohort Study

Full text

REPORT
Mental health findings from the Millennium Cohort Study
Leslie Morrison Gutman, Heather Joshi,
Michael Parsonage and Ingrid Schoon
Children of the new century

Centre for Mental Health REPORT Children of the new century
2
Acknowledgements
We are very grateful to members of our Expert Reference Group for helpful advice and support:
Dr Naomi Eisenstadt
Professor Barbara Maughan
Professor Lord Richard Layard
Contents
Executive summary 3
1 Introduction 6
2 Data sources and methods 7
3 Prevalence of mental health problems 10
4 Socio-demographic differences 13
5 Recent trends 21
6 Incidence and persistence 25
References 28
Appendix tables 29

3
Centre for Mental Health REPORT Children of the new century
This report sets out early ndings from a three-
year project on children’s mental health being
undertaken by the Institute of Education at UCL
and Centre for Mental Health, with funding from
the Economic and Social Research Council.
The ndings relate mainly to the mental health
of children around the age of 11 as recorded
in the Millennium Cohort Study (MCS), a multi-
purpose longitudinal study which is following
a large sample of children born in the UK at
the start of the 21st century. The data were
collected mainly in 2012.
Children’s mental health is assessed in the
MCS using the Strengths and Difculties
Questionnaire (SDQ), a widely used screening
instrument in which parents and teachers report
on a child’s mental health in the previous six
months.
The SDQ distinguishes between four broad
groups or “subscales” of mental health
difculties (conduct problems, hyperactivity/
inattention, emotional problems and peer
problems). A combined score across all four
subscales which is above a dened cut-off is
indicative of a clinically diagnosable disorder.
Children with scores at a level where a disorder
is likely to be present are described in this
report as having “severe” mental health
problems.
Prevalence
Based on the SDQ scores reported by parents,
the proportion of 11-year-old children in the UK
with severe mental health problems in 2012
was just over 10%.
There were markedly more boys reported as
having severe problems than girls (13% vs
8%). There were also some differences by
gender in the nature of problems: among boys
‘externalising’ problems (conduct problems
and hyperactivity/inattention) were relatively
more common than ‘internalising’ problems
(emotional and peer problems), while among
girls the opposite was the case.
The level of severe mental health problems
among 11-year-olds as rated by teachers was
somewhat lower at 8%, with an even wider gap
between boys and girls than suggested by the
parents’ scores.
As many as 28% of all 11-year-old children
were assessed by parents as having severe
difculties on at least one of the four individual
SDQ subscales.
Social and demographic differences
The prevalence of severe mental health
problems in children is strongly related to
parental education, parental occupation and
family income. For example, 17% of 11-year-
olds from families in the bottom fth of the
income distribution were identied as having
severe mental health problems in 2012,
compared with only 4% among those from
families in the top fth. This income-related
gradient in prevalence appears to have become
steeper in recent years and to be much steeper
among children than it is among adults.
Looking at 11-year-old-children living in the
four countries of the UK, the main feature is that
Scotland showed a somewhat lower prevalence
of severe mental health problems in 2012 than
England, Wales or Northern Ireland.
Comparisons of ethnic groups suggest that
severe problems were most common in 11-year-
old children of Mixed ethnic background
(particularly girls), followed by those classied
as White (particularly boys). Prevalence was
lowest among children of Indian origin.
When we look at the type of family, the
prevalence of severe mental health problems
among 11-year-old children living with both
their natural parents was well under half the
level found among children in other family types
(single-parent families, step-families etc.).
Recent trends
Recent changes have been assessed by
comparing data on the mental health of children
as recorded in the 2012 sweep of the MCS with
similar data on the mental health of 10- and 11-
Executive Summary

Centre for Mental Health REPORT Children of the new century
4
year-old children in the national Child and
Adolescent Mental Health Surveys carried out in
1999 and 2004.
Overall, the analysis suggests that the mental
health of children in this age group improved
slightly between 1999 and 2012. The picture
is however somewhat mixed and ndings vary
between boys and girls, by the type of problem
concerned and whether the rating is made by
parents or teachers.
As reported by parents, there appears, if
anything, to have been a small increase
between 1999 and 2012 in the overall
prevalence of severe mental health problems
in this age group. In contrast, teachers’ ratings
suggest somewhat larger changes going in the
opposite direction. The improvements noted by
teachers applied in broadly equal measure to
both boys and girls.
Incidence and persistence
The Millennium Cohort Study has so far
collected information on the mental health of
children in the sample at four different ages
(3, 5, 7 and 11). This can be analysed to show
how many children ever recorded severe mental
health problems between the ages of 3 and 11
(incidence) and how many with severe problems
at age 11 had these problems repeatedly
(persistence).
Just over a fth of all children in the MCS were
assessed with severe mental health problems
at one or more of the four ages surveyed. This
implies that the incidence of severe problems
at some point during childhood is roughly twice
as high as the prevalence of these problems at
any single age. On the other hand, only 1.5% of
children had severe problems at all four ages.
Dening persistent cases as those children who
had severe problems at three or four surveys
including at age 11, it is estimated that 3.6% of
all 11-year-old children fall into this category.
Persistence is nearly three times as common
among boys as among girls.
Looking ahead
We are charting the extent of mental health
problems in children of the new century as they
face adolescence, particularly in the context of
assessing the need for mental health services.
It is beyond the purpose of this report to explain
the causes of these problems or how they might
be prevented. These important questions
require further exploration of the evidence,
something we intend to do next.

5
Centre for Mental Health REPORT Children of the new century
Children of the new century:
Mental health ndings from the Millennium Cohort Study
Key
points
from the
report
Mental health
problems are
twice as common
in boys as girls
X2
Over 20% of children experience
a mental health problem at some
point between ages 3 and 11
20%
Overall, the mental health of 11-
year-old children was broadly the
same in 2012 as in 1999
0
500
1000
1500
2000
2500
1999 2012
Bottom
20%
Top
20%
Children from low-income families
are four times more likely
to experience mental health problems
than children from higher-income families
10% of 11-year-old children experienced a mental
health problem in 2012

Centre for Mental Health REPORT Children of the new century
6
1. Introduction
on four main topics:
• The prevalence of mental health problems
among 11-year-old children in the UK,
including analysis of the extent to which
children may be displaying two or more
problems at the same time.
• Socio-demographic differences in the
prevalence of mental health problems
among 11-year-olds, examining the extent
of variation in prevalence according to a
range of factors such as country of residence
within the UK, ethnicity, family type,
parental education, parental occupation and
family income.
• Trends in the prevalence of mental health
problems among 10- and 11-year-old
children between 1999 and 2012, based on
a comparison of ndings in the 1999 and
2004 BCAMHS and the MCS.
• Finally, information on patterns of incidence
and recurrence in mental health problems
within the same children, drawing on data
collected in successive surveys of the MCS
at child ages 3, 5, 7 and 11.
The report is written with a general readership
in mind and, wherever possible, quantitative
information is illustrated by relatively simple
diagrams rather than detailed tabular
presentations. For those with a more specialist
interest, detailed supporting information,
including sample numbers and statistical
signicance tests, is provided in a series of
appendix tables at the end of the report.
This report sets out some early ndings from a
three-year project on children’s mental health
being undertaken jointly by the Institute of
Education, now part of University College
London, and Centre for Mental Health, with
funding from the Economic and Social Research
Council1.
The primary aim of the joint project is to analyse
trajectories of mental health problems during
childhood using data from the Millennium
Cohort Study (MCS), a multi-purpose
longitudinal study which is following a large
sample of children born in the UK in 2000/01.
Early work on the project entailed the extraction
of substantial amounts of data on the mental
health of children in the MCS from successive
surveys carried out at child ages 3, 5, 7 and 11.
This was a necessary rst step in the analysis
of trajectories, but the work has also generated
much information on overall levels of mental
health in children born at the start of the 21st
century which may be of wider interest and
relevance. For example, it is now more than ten
years since the last ofcial British Child and
Adolescent Mental Health Survey (BCAMHS) was
in the eld, implying that decisions on policy
or service planning which rely on this source
are using information which is increasingly
out of date. In contrast, the most recent
survey of children in the MCS was carried out
mainly in 2012 and so provides data of a more
contemporary nature, albeit conned to children
born in a specic year. Another advantage of
the MCS is that it is based on a substantially
larger sample of children than the BCAMHS,
allowing more scope for detailed sub-group
analysis, for example by ethnicity.
The main, though not exclusive, focus of this
report is on the mental health of children around
the age of 11, which reects experiences at the
onset of adolescence and at an important time
of transition for most children because of the
move from primary to secondary education.
Following a short review of data sources and
methods, we present and discuss information
1 Research grant ES/L0008211/1 to Leslie Morrison Gutman.

7
Centre for Mental Health REPORT Children of the new century
Chapter 2: Data sources and methods
Data
The Millennium Cohort Study (MCS) is a
longitudinal study following a sample of
children born in all four countries of the United
Kingdom between September 2000 and January
2002. Unlike previous national birth cohort
studies, which attempted to recruit all children
born in a particular week, this survey drew on a
sample using a complex clustered and stratied
design, with deliberate over-sampling of
children in areas of high child poverty, minority
ethnic populations and the three smaller
countries of the UK (Plewis, 2007).
There have so far been ve rounds of interviews
completed with MCS families. The rst survey,
MCS1, was in the eld in 2001 and 2002, at
about nine months after birth of the sampled
children; MCS2 (child age 3) was mainly
undertaken during 2004, MCS3 (child age 5)
mainly during 2006 and MCS4 (child age 7)
mainly during 2008, while the most recent
completed survey, MCS5, at child age 11,
collected data mainly in 2012. A sixth survey is
in the eld in 2015, and another one planned
for 2018.
The number of families who have been
interviewed at least once is 19,244, which
includes some 692 families in England who
were not interviewed until MCS2. If these cases
are included, the initial response rate to the
survey out of all those eligible was 71%. As is
typical of longitudinal studies, there has been
some loss to follow-up, or ‘attrition’. Including
some former drop-outs who had re-joined,
the number of families interviewed at MCS5
in 2012 was 13,287, corresponding to 69% of
those ever interviewed (Mostafa, 2014). Finally,
the number of families interviewed at all ve
surveys undertaken to date was 10,448.
The fth survey is referred to as the ‘age
11’ survey and the children it surveys as
(for convenience) ‘11-year-olds’, although
interviewing took place at ages around, rather
than necessarily after, the 11th birthday. One
third were still 10, though approaching 11, and
most of the rest were closer to their 11th than
12th birthday.
Note also that parents and teachers were asked
to rate problems over the last six months, which
would have included time before the 11th
birthday for most of those who had already
reached that birthday.
Cases studied
In our analyses we examined one child per
family present at age 11, excluding 182 children
in MCS5 who were the second or third in sets
of twins or triplets. (There were not enough
of these to analyse separately.) Our sample
was further reduced because about 4% of
parents in MCS5 were not able to provide
usable information on their child’s mental
health, leaving a total of 12,798 cases for
analysis. Almost all of the parental informants
were mothers. They used computer-assisted
self-completion to answer the questions on
children’s mental health.
We also drew on responses from teachers in
England and Wales who answered a postal
paper questionnaire including questions on
children’s mental health in the 2012 survey.
(The questionnaire was not sent out in Scotland
or Northern Ireland). Teachers provided usable
information on mental health for 7,085 children.
Weighting procedures were used to correct for
the over-sampling of certain groups of children
in the MCS and for sample attrition, aiming to
ensure that the socio-demographic prole of the
whole cohort is nationally representative.
Measurement of mental health
As in the earlier surveys of children aged 3, 5
and 7 years, children’s mental health around
age 11 was assessed using the Strengths and
Difculties Questionnaire (SDQ) (Goodman,
1997 and 2001). This is a widely tested and
validated screening questionnaire in which
parents and/or teachers report on a child’s
mental health in the previous six months.
The SDQ includes 25 items divided into ve
groups or subscales relating to:
• conduct problems (e.g. often has temper

Centre for Mental Health REPORT Children of the new century
8
• tantrums…often ghts with other children…
steals from home, school or elsewhere);
• hyperactivity/inattention (e.g. restless,
overactive…constantly dgeting or
squirming…easily distracted, concentration
wanders);
• emotional problems (e.g. many worries…
often unhappy, downhearted…nervous or
clingy in new situations);
• peer problems (e.g. rather solitary, tends to
play alone…picked on or bullied…gets on
better with adults than other children); and
• pro-social (e.g. considerate of other
people’s feelings…shares readily with other
children…kind to younger children).
Each item is marked on a three-point scoring
system in which 0 = not true, 1 = somewhat
true and 2 = very true. The ve items in each
group or subscale are then totalled, giving
scores potentially ranging from 0 to 10. A
total difculties (or total problems) score is
calculated as the sum of the scores of the
conduct, hyperactivity/inattention, emotional
and peer problem subscales; the pro-social
subscale is not used in this calculation.
A distinction is sometimes drawn between
‘externalising’ and ‘internalising’ scores,
where ‘externalising’ is the sum of the conduct
and hyperactivity/inattention subscales, and
‘internalising’ is the sum of the emotional
and peer problem subscales. There is some
evidence that using these two combined scales
may be preferable to using the four separate
subscales in whole-population samples,
whereas the four separate subscales may
provide better discrimination in high-risk
samples (Goodman et al., 2010).
Although SDQ scores can be treated as
reecting a continuum, it is often convenient
to categorise or group the scores. The original
groupings presented for the SDQ scores were
‘normal’, ‘borderline’ and ‘abnormal’. These
bandings were established in a population-
based national survey, attempting to choose
cut-off points such that 80% of children scored
‘normal’, 10% ‘borderline’ and 10% ‘abnormal’.
Cut-off points are somewhat different based on
the reporter and age of the child. (The cut-offs
are to an extent arbitrary, a point particularly
relevant to the analysis in Chapter 6 of the
persistence and recurrence of mental health
problems during childhood, where some
apparent discontinuities may simply reect
children being rated just above a cut-off point
on one occasion and just below on another.)
To illustrate with actual numbers, ‘normal’
corresponds to a parent-rated total difculties
score which is in the range 0-13, ‘borderline’
corresponds to a score in the range 14-16 and
‘abnormal’ to a score in the range 17-40. An
‘abnormal’ score for total difculties – but
not necessarily for the individual subscales
– is understood to be indicative of a clinically
diagnosable mental disorder (Goodman and
Goodman, 2009). It is in practice very rare for a
child’s total difculties score to exceed 20, but
quite common for the total to be zero, where
absolutely no difculties are reported.
We retain this grouping system but modify the
terminology such that children in the three
bands are described as having ‘no problems’,
‘moderate problems’ and ‘severe problems’
respectively.
Assessments by parents and teachers
As noted above, the Millennium Cohort Study
has collected SDQ reports from teachers as
well as parents, and ndings from both sets of
assessments are presented at various points
in this report, sometimes with apparently
discrepant results.
The use of multiple informants is common
practice in studies of child mental health,
whatever specic instrument or measure is
used, and it has been widely observed that
informants often disagree about the presence
or absence of symptoms (De Los Reyes et al.,
2015). The SDQ is no exception to this. While
it has been found that agreement between
informants using this measure is “moderate”,
it is also “substantially higher than the
average reported for other measures of child
psychopathology” (Collishaw et al., 2009).
Various factors may contribute to divergence
of parent and teacher reports, but particular
importance attaches to contextual differences,
partly in the sense that children may behave

9
Centre for Mental Health REPORT Children of the new century
differently in different contexts (e.g. home
v. school) and partly because some sorts of
difculties 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 sufcient for a comprehensive assessment.
Statistical signicance
Although the MCS is a large-scale survey, there
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 signicance tests, this
report particularly highlights ndings which
have a likelihood of 95% or more that they
represent genuine differences.

Centre for Mental Health REPORT Children of the new century
10
Chapter 3: Prevalence of mental health problems
Prevalence based on parent ratings
The overall prevalence of mental health
problems among 11-year-old children in the UK,
based on the SDQ scores for total difculties
given by parents, is shown in Figure 3.1 below.
This indicates that, among all children, 82.8%
are classied as having no mental health
problems while 6.9% have moderate problems
and 10.3% have severe problems, indicative in
the latter case of a clinically diagnosable mental
disorder. The prevalence of both moderate and
severe problems is markedly higher among
boys than girls. For example, 12.7% of boys
are classied as having severe mental health
problems compared with only 7.7% of girls.
Detailed results for prevalence broken down
by type of problem, using the four subscales of
the SDQ, are given in Appendix Table A.1. This
shows that, among all 11-year-old children,
the proportions with severe problems are
much the same in all four SDQ categories,
ranging from 11.5% for both conduct problems
and hyperactivity/inattention to 11.9% for
emotional problems and 12.3% for peer
problems.
There are, however, differences when boys and
girls are analysed separately. Among girls, the
prevalence of severe internalising problems,
particularly emotional problems, is higher than
the prevalence of severe externalising problems,
whereas among boys the opposite is broadly
true, albeit with less overall variation between
the four subscales.
For three of the four subscales, the prevalence
of severe mental health problems is higher
among boys than girls, the one exception being
emotional problems, where the prevalence
of severe problems is much the same in both
genders. More than twice as many boys as
girls are identied by parents as having severe
hyperactivity/inattention problems.
Prevalence based on teacher ratings
The overall prevalence of mental health
problems based on the SDQ scores for total
difculties given by teachers is shown in Figure
3.2. It should be noted that these data relate
to England and Wales and are not therefore
directly comparable with those in Figure 3.1,
which relate to all of the UK.
Figure 3.1: Percentages of 11-year-old children with mental health problems: parent ratings,
UK, 2012
0
20
40
60
80
100 Severe
Moderate
None
Girls
BoysAll children
%

11
Centre for Mental Health REPORT Children of the new century
Figure 3.2: Percentages of 11-year-old children with mental health problems: teacher ratings,
England and Wales, 2012
0
20
40
60
80
100 Severe
Moderate
None
All children Girls
Boys
%
On the teacher ratings, 84.2% of all children
in England and Wales have no mental health
problems, 7.8% have moderate problems and
8.0% have severe problems. The proportion
with severe problems is thus somewhat lower
than the corresponding estimate for the UK
based on parent ratings; on the other hand, the
proportion with moderate problems is slightly
higher. More strikingly, the teachers identify
well over twice as many severe problems
among boys as girls, a noticeably wider gender
difference than in the parents’ scores.
A detailed breakdown of the teacher-rated
scores by type of problem is given in Appendix
Table A.2. In line with the parent data, this
shows that among girls the prevalence of
severe internalising problems is higher than the
prevalence of severe externalising problems,
whereas among boys the opposite is the case.
But there are also differences. For example,
the teacher ratings show that the prevalence
of severe conduct problems is three times as
high in boys as in girls, while the prevalence of
severe hyperactivity/inattention problems is
nearly ve times as high. These are much wider
gender differences than in the parent reports.
Combinations of mental health
problems
This section provides information on the extent
to which some children exhibit two or more
severe mental health problems at the same
time. Findings based on parent ratings are
given in Figure 3.3.
This shows that 28.1% of all children have
severe difculties in at least one of the SDQ
subscales. The total breaks down as follows:
15.7% of children are classied as having one
severe problem, 7.2% have two such problems,
3.6% have three and 1.6% have severe ratings
in all four subscales.
There are also gender differences in the extent
to which types of problem are combined. For
11-year-old boys, 31.4% are classied with
one or more severe problems, including 16.4%
with just one problem, 8.5% with two, 4.3%
with three and 2.2% with severe ratings in all
four subscales. All of these rates are lower for
girls, with 24.6% having at least one problem,
including 15.0% with one problem only, 5.8%
with two problems, 2.5% with three and 1.0%
with four.

Centre for Mental Health REPORT Children of the new century
12
Detailed information on all possible
combinations of problems is given in Appendix
Table A.3. This shows, for example, that for
children classied as having severe problems
in two separate subscales, the most common
grouping is emotional problems + peer
problems (i.e. the two internalising conditions),
followed by conduct problems + hyperactivity/
inattention (i.e. the two externalising
conditions). The internalising grouping is
more common among girls than boys and vice
versa for the externalising grouping. Among
all children, 8.9% have at least one problem in
both internalising and externalising groups.
Figure 3.3: Percentages of 11-year-old children with different numbers of severe mental
health problems: parent ratings, UK, 2012
Four problems
Three problems
Two problems
No problems
All children
Boys Girls
One problem

13
Centre for Mental Health REPORT Children of the new century
Chapter 4: Socio-demographic differences
Introduction
This chapter explores differences in the
proportions of 11-year-old children with severe
mental health problems according to a range of
socio-demographic factors such as ethnicity and
family income. All the ndings are based on
SDQ scores provided by parents. It is important
to note that the associations shown below do
not necessarily imply causation, as many other
factors may also be at play and further analysis
is needed to disentangle their separate effects.
Country of residence within the UK
The relative proportions of 11-year-old children
in the four constituent countries of the UK
who have severe mental health problems,
as measured by their SDQ scores for total
difculties, are given in Figure 4.1.
The main feature of Figure 4.1 is the lower
prevalence of severe mental health problems in
Scotland compared with the rest of the UK; thus
the relative numbers in Scotland with severe
problems are below those elsewhere by about
a quarter among both boys and girls. It should,
however, be noted that this difference falls just
short of statistical signicance. In the UK as a
whole the prevalence of severe problems among
boys is highest in Wales, while among girls it
is highest in England. Prevalence in Northern
Ireland is virtually identical to the UK average,
both for children taken together and for boys
and girls separately.
Detailed information on the proportions with
severe problems in each of the four SDQ
subscales is given by country in Appendix Table
A.4. Among other things this suggests that the
lower prevalence of severe problems in Scotland
is somewhat more pronounced for internalising
problems than it is for externalising problems.
Within the UK as a whole, the prevalence of
severe problems is highest in England for
conduct problems and peer problems, in Wales
for hyperactivity/inattention, and in Northern
Ireland for emotional problems.
Ethnicity
A breakdown by ethnicity of the prevalence of
severe mental health problems, based on SDQ
scores for total difculties, is given in Figure 4.2
(overleaf).
Figure 4.1: Percentages of 11-year-old children
with severe mental health problems (total
difculties) by country of residence: parent
ratings, UK, 2012
0
3
6
9
12
15
N. IrelandScotlandWalesEngland
0
3
6
9
12
15
N. IrelandScotlandWalesEngland
0
3
6
9
12
15
N. IrelandScotlandWalesEngland
All children
Girls
Boys
%
%
%

Centre for Mental Health REPORT Children of the new century
14
Figure 4.2: Percentages of 11-year-old children with severe mental
health problems (total difculties) by ethnicity: parent ratings, UK, 2012
0
3
6
9
12
15
OtherBlack Pakistani
/Bangladeshi
IndianMixedWhite
0
3
6
9
12
15
OtherBlack Pakistani
/Bangladeshi
IndianMixedWhite
0
3
6
9
12
15
OtherBlack Pakistani
/Bangladeshi
IndianMixedWhite
All children
Girls
Boys
%
%
%

15
Centre for Mental Health REPORT Children of the new century
This shows that, among all children, prevalence
is highest in the Mixed group, followed by
those classied as White. In all the other four
ethnic groups for which information is given,
the prevalence of severe problems is below the
national average and is particularly low among
children of Indian origin. These differences are
not, however, at a level which reaches statistical
signicance.
Patterns for boys and girls within ethnic groups
are broadly similar but with some exceptions.
In particular, the prevalence of severe problems
among boys in the Mixed group is below rather
than above the national average, meaning that
the high overall prevalence of severe problems
in this group is fully explained by an extremely
high level of problems among girls. Severe
mental health problems are most common
among White boys, closely followed by those
classied as Black. Among both boys and girls,
prevalence is lowest in Indians, with the rate for
boys being particularly low.
Detailed information on severe problems in
each of the four SDQ subscales by ethnicity is
given in Appendix Table A.5. Features of note
include: very low levels of conduct problems -
but not hyperactivity/inattention - among Indian
children; high levels of emotional problems
among children in the Mixed group, particularly
girls; exceptionally low levels of conduct
problems among girls in the Black and Other
groups; and very high levels of peer problems
among boys in the small and heterogeneous
Other group.
The pattern of severe problems across ethnic
groups just described contrasts with earlier
waves of the survey, where Indians and Mixed
had similar total difculties scores to Whites,
while Pakistanis and Bangladeshis, particularly
the former, had signicantly higher levels of
problems (George et al., 2006). The earlier
surveys show a contrast between a high level
of problems among Black Caribbeans and a
particularly low one for Black Africans.
Partnership status of parents
The prevalence of severe mental health
problems on the total difculties measure
according to parents’ partnership status is given
in Figure 4.3, with matching information
for each of the four SDQ subscales given in
Appendix Table A. 6.
The prevalence of severe mental health
problems among children living with both their
natural parents is about a half to a third of the
level found among children in other family types
(single-parent families, step-families etc.). The
differential is broadly the same among boys as
among girls and is also observed across all four
0
5
10
15
20
25
Both natural
parents
Lone natural
parent Step and other
family types
0
3
6
9
12
15
Step and other
family types
Lone natural
parent
Both natural
parents
0
5
10
15
20
Step and other
family types
Lone natural
parent
Both natural
parents
All children
Girls
Boys
%
%
%
Figure 4.3: Percentages of 11-year-old children
with severe mental health problems (total
difculties) by partnership status of parents:
parent ratings, UK

Centre for Mental Health REPORT Children of the new century
16
higher the prevalence of severe mental health
problems in their children. For example, using
SDQ scores for total difculties, the children of
parents with no or very low qualications are
about three times as likely to exhibit severe
mental health problems as those children
whose parents have a university education up to
rst-degree level and about four times as likely
as those whose parents have higher-degree
qualications.
The inverse association appears to be
somewhat stronger among girls than among
boys, while the detailed analysis by SDQ
subscale given in the appendix table suggests
that the relationship is more pronounced
for externalising problems than it is for
internalising problems.
types of mental health problems, although to
a more pronounced degree for externalising as
opposed to internalising problems.
Parental education
Figure 4.4 below shows the proportions of
children with severe mental health problems
analysed according to highest academic
qualication held by the parents (or parent if
only one present). Appendix Table A.7 provides
detailed information relating to each of the four
SDQ subscales.
All the plots show a very strong inverse
association between parental education and
child mental health problems, i.e. the lower the
level of parents’ educational qualications, the
Figure 4.4: Percentages of 11-year-old children with severe mental health problems (total
difculties, parent ratings) by highest parent academic qualication: UK, 2012
All children
%
0
5
10
15
20
25
Higher
degree
Degree
/Diploma
AS/
A-level
GCSE A-CLow
qual.
No
academic
qualifications
Boys
%
0
5
10
15
20
25
Higher
degree
Degree
/Diploma
AS/
A-level
GCSE A-CLow
qual.
No
academic
qualifications
Girls
%
0
5
10
15
20
25
Higher
degree
Degree
/Diploma
AS/
A-level
GCSE A-CLow
qual.
No
academic
qualifications

17
Centre for Mental Health REPORT Children of the new century
Parental occupation
Figure 4.5 below shows the proportions of
children with severe mental health problems
according to the occupational status of their
parents, taking the highest ranked current
occupation of a couple, where there are two
parents present at child age 11. Detailed
supporting information relating to the four SDQ
subscales is given in Appendix Table A.8.
Figure 4.5: Percentages of 11-year-old children with severe mental health problems
(total difculties) by parental occupation at child age 11: parent ratings, UK, 2012
0
5
10
15
20
25
30
0
5
10
15
20
25
30
0
5
10
15
20
25
30
All Children
Girls
Boys
%
%
%
Professional/
Managerial
IntermediateSmall EmployerLower
Supervisory
Routine/
Semi-Routine
Workless
Professional/
Managerial
IntermediateSmall EmployerLower
Supervisory
Routine/
Semi-Routine
Workless
Professional/
Managerial
IntermediateSmall EmployerLower
Supervisory
Routine/
Semi-Routine
Workless

Centre for Mental Health REPORT Children of the new century
18
Given the close links between education and
occupation, it is perhaps not surprising to nd
that the inverse association identied above
between child mental health problems and
parental education is repeated in the case of
the highest ranked occupation of any parent
who was working at child age 11. As before,
the gap in prevalence is strikingly wide,
with the prevalence of severe mental health
problems based on the SDQ total difculties
score being about ve times as high among
the children of workless parents as it is among
the children of parents in the professional/
managerial group. Measured in relative
terms, the difference in prevalence between
children with parents at opposite ends of the
occupational scale is much the same among
boys as it is among girls, but it is larger among
boys when measured in absolute percentage
points, reecting the higher overall prevalence
of severe mental health problems in this
group. Thus the prevalence of parent-reported
severe mental health problems among boys
with workless parents is 27.2% compared with
5.8% among boys whose parents are in the
occupational/managerial group, whereas the
corresponding rates among girls are 15.2% and
2.9% respectively. Analysed by SDQ subscale,
the prevalence gap is somewhat wider for
externalising problems, particularly conduct
problems, than it is for internalising problems.
Family income
Figure 4.6 shows the prevalence of severe
mental health problems on the total difculties
measure analysed according to family income,
with detailed information relating to the four
SDQ subscales given in Appendix Table A.9.
In line with the strong links between
occupational status and income, these gures
again show a steep gradient in prevalence
across the income scale. For example, 17.0%
of all children from families in the bottom fth
of the income distribution are identied by
their parents as having severe mental health
problems, compared with only 4.1% among
those from families in the top fth. The
proportions of children with severe problems in
the middle quintile are twice as large as among
those in the top quintile and half as large as
among those in the bottom quintile. As before,
the slope of the gradient is broadly the same
among boys as it is among girls and, analysed
according to the four SDQ subscales, is most
pronounced among children with conduct
problems, irrespective of gender.
The fact that the risks of severe problems are
greater for the lower income groups does not
mean that they are totally absent from other
groups in the population. As can be seen in the
gures, there are some children in the middle
and upper income groups who display severe
problems. If we consider how many of those
with severe problems are in each income group
(rather than the proportion in each income
group who have problems), it turns out that
around two-thirds of the problems are found
in the bottom two-fths of the income range.
Targeting the bottom two quintiles would miss
the one third of cases spread (in diminishing
proportion) over the top three bands. Targeting
services at the lowest quintile only would miss
two-thirds of the cases.
This illustrates Rose’s Paradox (Rose, 1992) that
the highest risk group in a population does not
necessarily account for the majority of the risk,
which calls for a strategy of prevention in the
population as a whole, as well as or instead of a
concentration of effort on high risk groups. The
pattern of risk concentration in total difculties
is similar in girls and boys, and fairly similar
for the subscales. Conduct problems are more
concentrated in the bottom two income groups
(70%) and the others less so; hyperactivity/
inattention and peer problems both have 60%
of the cases from the bottom 40% of income,
and emotional problems have 50%.
Information on the relationship between
childhood mental health problems and family
income is also given in the 2004 Child and
Adolescent Mental Health Survey (Green et
al., 2005), although income is measured in
ranges of £ per week rather than quintiles.
Converting the data to quintiles on the basis of
interpolation, the gures suggest that in

19
Centre for Mental Health REPORT Children of the new century
2004 the prevalence of severe mental health
problems in children aged 11-16 was about
three times as high among those in the bottom
quintile of family income as among those in the
top quintile. This compares with a more than
fourfold difference among 11-year-olds in 2012,
suggesting that the income-related gradient in
the prevalence of severe mental health
Figure 4.6: Percentages of 11-year-old children with severe mental health
problems (total difculties) by quintile of family income: parent ratings, UK, 2012
0
4
8
12
16
20
0
4
8
12
16
20
0
4
8
12
16
20
All children
Third
20%
Top
20%
Fourth
20%
Bottom
20%
Second
20%
Girls
Boys
%
%
%
Third
20%
Top
20%
Fourth
20%
Bottom
20%
Second
20%
Third
20%
Top
20%
Fourth
20%
Bottom
20%
Second
20%

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

21
Centre for Mental Health REPORT Children of the new century
Chapter 5: Recent trends
Introduction
This chapter examines recent changes over
time in children’s mental health using parent
and teacher reports from three nationally
representative data sets: the British Child
and Adolescent Mental Health Survey of 1999
(Meltzer et al., 2000), the repeat of this survey
undertaken in 2004 (Green et al., 2005) and the
survey of the MCS sample carried out in 2012.
Because the 1999 and 2004 surveys did not
cover Northern Ireland, parent reports relate to
children living in Great Britain. Furthermore, as
the 2012 teacher survey did not cover Scotland,
teacher reports relate to children living in
England and Wales.
More importantly, because the 1999 and 2004
surveys covered children of all ages between 5
and 16, the numbers of 11-year-olds in these
surveys are relatively small, reducing the
likelihood that observed changes over time
will be found to be statistically signicant. To
address this, the samples of children from the
1999 and 2004 surveys have been expanded
to cover 10-year-olds as well as 11-year-olds,
taking advantage of the fact that about a third
of all children in the MCS sample were aged
10 at the time they were surveyed in 2012.
To minimise the extent of possible age bias,
additional weights were applied to the SDQ
data in the MCS in order to have an equal
representation of 10- and 11-year-olds. Socio-
demographic characteristics of the three
samples are given in Appendix Table A.10. The
proles are fairly similar, though they also
reect a general trend for fewer two-parent
families to be married rather than cohabiting
and for a decline in owner occupation.
Changes in mental health at a
population level
The analyses presented so far in this report
have focused mainly on children with severe
mental health difculties. However, in looking
at trends over time, there may also be interest
in taking a wider perspective, using SDQ scores
averaged across the whole population of 10-
and 11-year-olds in each of the three samples,
in order to identify whether there has been any
general improvement or deterioration in the
mental health of children in this age group.
Average SDQ scores for total difculties in
1999, 2004 and 2012 are shown in Figure 5.1
(parent ratings) and Figure 5.2 (teacher ratings).
Detailed information relating to the four SDQ
subscales is given in Appendix Tables A.11
and A.12. In all cases, higher scores represent
worse mental health.
Figure 5.1: Average SDQ scores for total difculties among 10- and 11-year-old children:
parent ratings, Great Britain
0
2
4
6
8
10
MCS 2012
BCAMHS 2004
BCAMHS 1999
Girls
Boys
All children
SDQ
scores

Centre for Mental Health REPORT Children of the new century
22
Looking rst at the data for total difculties,
the reports by parents suggest little change
comparing 1999 and 2012. There was a
statistically signicant improvement in average
SDQ scores for total difculties, particularly
among girls, between 1999 and 2004, but much
of this was lost by 2012.
The reports by teachers paint a more optimistic
picture, showing statistically signicant
improvements in the average score of total
difculties for all children, and for both boys
and girls considered separately, between 1999
and 2012. For all children, and for girls but not
boys on their own, there were also statistically
signicant improvements in the sub-period
2004 to 2012.
Detailed data on the four SDQ subscales also
show contrasting results between parents
and teachers. On parent reports, the only
signicant change between 1999 and 2012 was
an improvement in scores for hyperactivity/
inattention among boys. While parent reports
show a decrease in hyperactivity/inattention
and peer problems from 1999 to 2004 for girls,
these improvements were lost by 2012. In fact,
there was a statistically signicant increase in
girls’ hyperactivity/inattention from 2004.
Teacher reports, on the other hand, show
improvements among both boys and girls in
most areas of mental health, the main exception
being emotional problems.
Changes in the prevalence of severe
mental health problems
Returning to a focus on the group of children
showing the worst SDQ symptoms, changes
in the prevalence of severe mental health
problems between 1999 and 2012 are shown
in Figure 5.3 (parent ratings) and Figure 5.4
(teacher ratings.) Detailed data on the four SDQ
subscales is given in Appendix Tables A.13 and
A.14.
Figure 5.2: Average SDQ scores for total difculties among 10- and 11-year-old
children: teacher ratings, England and Wales
0
2
4
6
8
10
MCS 2012
BCAMHS 2004
BCAMHS 1999
Girls
Boys
All children
SDQ
scores

23
Centre for Mental Health REPORT Children of the new century
Figure 5.3: Percentages of 10- and 11-year-old children with severe mental health
problems (total difculties) in 1999, 2004 and 2012: parent ratings, Great Britain
0
4
8
12
16
MCS 2012
BCAMHS 2004
BCAMHS 1999
Girls
BoysAll children
%
0
4
8
12
16
MCS 2012
BCAMHS 2004
BCAMHS 1999
Girls
BoysAll children
%
Figure 5.4: Percentages of 10- and 11-year-old children with severe mental health
problems (total difculties) in 1999, 2004 and 2012: teacher ratings, England and Wales

Centre for Mental Health REPORT Children of the new century
24
Looking rst at the data on total difculties,
we again nd conicting results depending
on whether the ratings are made by parents
or teachers. According to parents’ reports,
the proportion of 10- and 11-year-old children
with severe problems increased somewhat
between 1999 and 2012, but not enough to
reach statistical signicance, whether among all
children combined or among boys or girls taken
separately.
In contrast, teacher reports suggest that the
prevalence of severe mental health problems
fell signicantly between 1999 and 2012.
This applies to children as a whole and also to
boys on their own but not to girls. In addition,
there were signicant falls in prevalence in all
groupings including girls in the sub-period 2004
to 2012.
Turning to the individual SDQ subscales,
the only signicant change in parent reports
between 1999 and 2012 was a decline in
the prevalence of severe problems relating
to hyperactivity and inattention, particularly
among boys. There was also an increase in
the prevalence of severe peer problems for
girls from 2004. In contrast, the assessments
by teachers again show more widespread
improvements, covering conduct problems
as well as hyperactivity/inattention among
both boys and girls and also a decline in the
prevalence of emotional problems among girls
between 2004 and 2012.
Discussion
Overall, the ndings set out in this chapter
suggest that the mental health of 10- and
11-year-old children in this country improved
slightly rather than deteriorated between 1999
and 2012. The picture is, however, somewhat
mixed and it is also apparent that perceptions
vary according to the gender of the child, the
type of mental health problem concerned and
whether assessments are provided by parents
or teachers.
The last of these points is particularly important.
For example, reports by parents identify only
one area of consistent improvement between
1999 and 2012, whether this relates to average
SDQ scores or the prevalence of severe
problems, namely hyperactivity/inattention
among boys. In contrast, the assessments by
teachers suggest improvements on a wider
front, including conduct problems as well as
hyperactivity/inattention and applying to girls
as well as boys. The parents’ scores for total
difculties suggest a small and non-signicant
increase in the prevalence of severe problems,
whereas teachers’ ratings indicate a signicant
fall in prevalence, particularly among boys as
well as girls from 2004.
The mixed nature of these ndings is in contrast
to those reported in an earlier study which
compared the mental health of children in the
MCS when they were 7 years old with that of
children of the same age in the 1999 and 2004
BCAMHS (Sellers et al., 2014). This found
strong evidence of improving mental health
over the period studied (1999-2008), including
both a decline in average problem scores
among all 7-year-olds and a fall in the relative
numbers with severe problems for all symptom
types. These improvements were observed in
both parent and teacher reports and were more
marked for boys than girls.
Various factors might explain the more varied
picture found among 10- and 11-year-old
children, including for example the impact
on families of the economic recession which
followed the nancial crash of 2008/09.
Another factor may be an increased focus on
behaviour management in the classroom, which
may further explain the differences between
parent and teacher reports. It may also be
the case that 10- and 11-year-old children
show fewer mental health problems in the
classroom than at home. Age 10 to 11 marks
the beginning of puberty, and early adolescents
might demonstrate different behaviours in front
of their parents, as opposed to their teachers
and peers. It will be of interest to assess
whether or not the earlier and more clear-cut
trend towards improving mental health among
children has been maintained when data from
the most recent sweep of the MCS (at child age
14) become available for comparison with the
earlier surveys.

25
Centre for Mental Health REPORT Children of the new century
Chapter 6: Incidence and persistence
Introduction
This chapter provides an analysis of severe
mental health problems among those children
in the MCS for whom valid data on mental
health are available at all four of the surveys
undertaken when the child was aged 3, 5,
7 and 11. We describe this data set as the
longitudinal sample. From these records we
can see how many children ever recorded a
particular type of severe problem between the
ages of 3 and 11 (incidence) and also how many
with severe problems at age 11 also had these
problems repeatedly before that (persistence).
This is the sort of picture that can only be
provided by longitudinal data.
A drawback with our longitudinal sample is that
the picture has some pieces missing, in the form
of nearly three thousand cases with incomplete
records. We can tell that these exclusions have
a somewhat different prole from those who
are observed on all occasions, because the
longitudinal sample understates the prevalence
of severe mental health problems at age 11
compared with the estimates given elsewhere
in this report. For example, the prevalence
of severe problems at age 11 on the total
difculties score is 10.3% in the full sample but
only 8.9% in the longitudinal sample. For the
purpose of plotting the longitudinal results in
graphs, we have made a crude adjustment for
the missing cases, distributing them pro rata
over those with and without severe problems
at age 11 according to the information we
have about them at that point. It is likely that
children with recurrent severe problems are still
under-represented after this adjustment.
Incidence
Figure 6.1 provides information on the number
of different ages at which children in the
longitudinal sample were assessed as having
severe mental health problems based on the
total difculties score on parent's reports. A
more detailed but unadjusted breakdown of the
longitudinal sample is given in Appendix Table
A.15.
Figure 6.1: Percentages of children with severe mental health problems (total
difculties) at different ages: parent ratings, UK
Problems at four ages
Problems at three ages
Problems at two ages
Problems at one age
No problems
All Children
Boys Girls

Centre for Mental Health REPORT Children of the new century
26
Figure 6.1 indicates that among all children
in the longitudinal sample just over a fth
(21.9%) were assessed with severe mental
health problems in at least one of the four MCS
surveys undertaken at child ages 3, 5, 7 and 11.
One implication of this is that the incidence of
severe problems during the course of childhood
is considerably higher than the prevalence of
these problems at any single age. For example,
it was noted above that the prevalence of severe
problems at age 11 was 10.3%. This is less than
half the gure of 21.9% for incidence between
ages 3 and 11. A major reason for the excess
of incidence over prevalence is that as many as
7% of all children in the MCS were classied as
having severe problems at age 3 only.
The estimate of 21.9% for incidence breaks
down as follows: 12.7% of children had severe
mental health problems at one age only, 5.1%
had these problems at two different ages, 2.7%
had problems at three ages and just 1.5% had
severe problems at all four ages.
There was a higher incidence of severe mental
health problems among boys than girls on
all measures. For example, 26.0% of boys
in the longitudinal sample are estimated to
have experienced severe problems at least
once, compared with 17.8% of girls. Twice as
many boys as girls were assessed with severe
problems at three different ages between 3 and
11, and over three times as many had severe
problems at all four ages.
These features broadly apply to each of the
four SDQ subscales, as shown in Appendix
Tables A.16-19. Incidence is somewhat higher
for externalising problems (conduct problems
and hyperactivity/inattention) than it is for
internalising problems (emotional problems
and peer problems) but is always at least twice
as high as prevalence. With the exception of
emotional problems, where the pattern is very
similar by gender, incidence is higher among
boys than among girls on all measures.
Persistence
Appendix Tables A.15-19 provide a good
deal of detailed information on patterns of
persistence and recurrence in severe mental
health problems during childhood. Because
the main focus of this report is on the mental
health of 11-year-old children, we look here at
persistence and recurrence specically among
those in the longitudinal sample who were
assessed with severe problems at this age.
We have also chosen to present as ‘persistent’
cases all children who had severe problems at
three or four surveys including at age 11. For
these with severe problems at three surveys, it
remains to be checked how far the apparently
problem-free year had a score close to the
threshold, as we suggested above might be the
case.
The persistence of severe problems using the
total difculties score is summarised in Figure
6.2 below.
Figure 6.2: Incidence of severe mental health problems (total difculties) by age 11:
parent ratings, UK
0
5
10
15
20
25
30 11 and 2 or 3 other ages
11 alone or + one other
Some but not age 11
GirlsBoysAll children
%

27
Centre for Mental Health REPORT Children of the new century
This shows that 3.6% of all 11-year-old children
are classied 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 dened it is
nearly three times as common among boys as
among girls (5.1% against 1.8%). The numbers
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:
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.
Boys Girls All
Children
Conduct
Problems
6.3 3.4 4.9
Hyperactivity/
inattention
8.1 3.0 5.6
Emotional
Problems
2.1 2.0 2.1
Peer Problems 4.2 2.3 3.2

Centre for Mental Health REPORT Children of the new century
28
T. and Goodman, R. (2005) Mental health of
children and young people in Great Britain,
2004. London: Palgrave Macmillan.
McManus, S., Meltzer, H., Brugha, T.,
Bebbington, P. and Jenkins, R. (eds.) (2009)
Adult psychiatry morbidity in England, 2007.
London: NHS Information Centre for Health and
Social Care.
Meltzer, H., Gatward, R., Goodman, R. and Ford,
T. (2000) The mental health of children and
adolescents in Great Britain. London: TSO.
Mostafa, T. (2014) Millennium Cohort Study:
technical report on response in sweep 5 (age
11). London: Centre for Longitudinal Studies.
Plewis, I. (ed.) (2007 The Millennium Cohort
Study: technical report on sampling (4th
edition). London: Centre for Longitudinal
Studies.
Rose, G. (1992). The strategy of preventive
medicine. Oxford: Oxford Medical publications.
Sellers, R., Maughan, B., Pickles, A., Thapar, A.
and Collishaw,S. (2014) Trends in parent- and
teacher-rated emotional, conduct and ADHD
problems and their impact in prepubertal
children in Great Britain: 1999-2008. Journal of
Child Psychology and Psychiatry, 56(1), 49-57.
Collishaw, S., Goodman, R., Ford, T., Rabe-
Hesketh, S. and Pickles, A. (2009) How far
are associations between child, family and
community factors and child psychopathology
informant-specic and informant-general?
Journal of Child Psychology and Psychiatry,
50(5), 571-580.
De Los Reyes, A., Augenstein, T., Wang, M. et
al. (2015) The validity of the multi-informant
approach to assessing child and adolescent
mental health. Psychological Bulletin, 141(4),
858-900.
George, A., Hansen, K., Schoon, I. (2006).
Child Development. In: S. Dex & H. Joshi
(Eds.). Millennium Cohort Study Second
Survey. Descriptive Report. London: Centre for
Longitudinal Studies.
Goodman, Alissa, Joshi, H., Nasim, B. and
Tyler, C. (2015) Social and emotional skills
in childhood and the long-term effects on
adult life. Early Intervention Foundation
Review, March 2015. London: UCL Institute of
Education.
Goodman, Anna, Lamping, D. and Ploubidis,
G. (2010) When to use broader internalising
and externalising subscales instead of the
hypothesised ve subscales on the Strengths
and Difculties Questionnaire (SDQ): data from
British parents, teachers and children. Journal
of Abnormal Child Psychology, 38, 1179-1191.
Goodman, Anna, and Goodman, R. (2009)
Strengths and Difculties Questionnaire as a
dimensional measure of child mental health.
Journal of the American Academy of Child and
Adolescent Psychiatry, 48(4), 400-403.
Goodman, R. (1997) The Strengths and
Difculties Questionnaire: a research note.
Journal of Child Psychology and Psychiatry,
38(5), 581-586.
Goodman, R. (2001) Psychometric properties
of the Strengths and Difculties Questionnaire.
Journal of the American Academy of Child and
Adolescent Psychiatry, 40(11), 1337-1345.
Green, H., McGinnity, A., Meltzer, H., Ford,
References

29
Centre for Mental Health REPORT Children of the new century
Boys Girls All Children
SDQ None Moderate Severe n None Moderate Severe n None Moderate Severe N *χ2
p-value
Conduct
Problems
74.7 11.4 13.9 6,458 81.4 9.7 8.9 6,340 77.9 10.6 11.5 12,798
98.01
p<.001
Hyper-
activity/
inattention
76.4 8.1 15.5 6,439 87.4 5.5 7.1 6,322 81.7 6.8 11.5 12,761
273.14
p<.001
Emotional
Problems
81.1 7.2 11.7 6,449 80.4 7.5 12.1 6,333 80.8 7.3 11.9 12,782
1.17
p = .39
Peer
Problems
76.4 9.4 14.2 6,454 80.9 8.3 10.8 6,336 78.6 8.9 12.5 12,790
45.39
p <.001
Total
Problems
79.3 8.0 12.7 6,431 86.5 5.7 7.8 6,320 82.8 6.9 10.3 12,751
118.18
p<.001
Sample size unweighted. Statistical tests use corrections for sample design.
* Likelihood ratio χ2 test probability for comparison of boys and girls.
Appendix Table A.1.
Percentages of 11-year-old children with mental health problems (parent-rated)
by SDQ subscale: United Kingdom, 2012

Centre for Mental Health REPORT Children of the new century
30
Boys Girls All Children
SDQ None Moderate Severe n None Moderate Severe n None Moderate Severe N *χ2
p-value
Conduct
Problems
83.7 5.8 10.5 3,538 94.5 2.0 3.5 3,547 89.0 4.0 7.0 7,085
208.52
P<.001
Hyper-
activity/
inattention
78.6 6.6 14.8 3,538 94.6 2.1 3.3 3,547 86.5 4.4 9.1 7,085
388.70
P<.001
Emotional
Problems
90.8 4.4 4.8 3,538 91.5 4.2 4.3 3,547 91.1 4.3 4.6 7,085
1.34
P<.47
Peer
Problems
85.9 5.6 8.5 3,538 91.1 3.6 5.3 3,547 88.5 4.6 6.9 7,085
48.00
P<.001
Total
Problems
78.2 10.5 11.3 3,538 90.4 5.1 4.5 3,547 84.2 7.8 8.0 7,085
200.25
p<.001
Sample size unweighted. Statistical tests use corrections for sample design.
* Likelihood ratio χ2 test probability for comparison of boys and girls.
Appendix Table A.2.
Percentages of 11-year-old children with mental health problems (teacher-rated) by SDQ subscale: England and Wales, 2012

31
Centre for Mental Health REPORT Children of the new century
Boys Girls All Children
None 68.6 75.4 71.9
Conduct 3.8 3.3 3.6
Hyperactivity 5.0 2.4 3.7
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.
Appendix Table A.3.
Percentages of 11-year-old children with severe problems (parent-rated) in
combinations of SDQ subscales: United Kingdom, 2012

Centre for Mental Health REPORT Children of the new century
32
Conduct
Problems
Hyperactivity/
Inattention
Emotional
Problems
Peer
Problems
Total
Problems
All Children
England 11.6
(12,987)
11.7
(12,952)
12.2
(12,972)
13.0
(12,981)
10.6
(12,941)
Wales 11.3
(771)
12.8
(769)
11.3
(769)
12.2
(770)
10.6
(768)
Scotland 10.2
(1,376)
9.1
(1,376)
9.0
(1,376)
9.0
(1,376)
7.9
(1,376)
N. Ireland 11.4
(636)
11.9
(634)
13.0
(636)
11.5
(636)
10.4
(634)
χ² difference 3.66 5.90* 5.80 12.27** 5.31
Unweighted N 12,798 12,761 12,782 12,790 12,751
Boys
England 14.2
(6,677)
15.6
(6,654)
11.9
(6,688)
14.9
(6,673)
12.9
(6,644)
Wales 13.3
(414)
17.2
(413)
11.4
(413)
13.7
(414)
13.7
(413)
Scotland 11.9
(686)
13.5
(686)
9.4
(686)
9.1
(686)
10.1
(686)
N. Ireland 12.5
(342)
16.3
(341)
12.3
(342)
12.8
(342)
12.8
(341)
χ² difference 2.98 2.51 3.08 14.13*** 3.92
Unweighted N 6,458 6,439 6,449 6,454 6,431
Girls
England 8.8
(6,261)
7.5
(6,249)
12.5
(6,255)
11.1
(6,259)
8.1
(6,248)
Wales 10.6
(378)
8.0
(378)
11.3
(377)
10.6
(378)
7.2
(377)
Scotland 8.6
(704)
4.7
(704)
8.6
(704)
8.9
(704)
5.7
(704)
N. Ireland 10.3
(308)
7.2
(307)
13.7
(308)
10.0
(308)
7.7
(307)
χ² difference 1.94 6.27* 8.46* 2.78 4.26
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
reect sample losses (Mostafa, 2014).
Appendix Table A.4.
Percentages of 11-year-old children with severe problems (parent-rated) by
country of residence at MCS5 and SDQ subscale: United Kingdom, 2012

33
Centre for Mental Health REPORT Children of the new century
Conduct
Problems
Hyperactivity/
Inattention
Emotional
Problems
Peer
Problems
Total
Problems
All Children
White 12.0
(13,521)
11.9
(13,486)
12.0
(13,481)
12.4
(13,495)
10.6
(13,476)
Mixed 12.7
(553)
12.9
(551)
15.9
(553)
13.9
(553)
12.3
(551)
Indian 3.7
(287)
12.5
(285)
8.2
(287)
10.6
(287)
5.0
(285)
Pakistani/
Bangladeshi
8.2
(712)
7.3
(697)
11.4
(706)
14.0
(710)
7.8
(695)
Black 7.5
(519)
7.8
(505)
8.4
(517)
13.2
(517)
10.1
(502)
Other 5.0
(202)
7.4
(201)
13.2
(202)
17.9
(202)
7.9
(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
(6,916)
16.2
(6,912)
11.8
(6,910)
14.1
(6,914)
13.2
(6,906)
Mixed 12.0
(297)
14.1
(296)
12.5
(296)
12.7
(296)
10.9
(296)
Indian 3.6
(162)
16.9
(162)
9.5
(162)
12.8
(162)
4.5
(162)
Pakistani/
Bangladeshi
9.2
(352)
10.8
(344)
10.8
(349)
15.6
(352)
9.3
(342)
Black 10.9
(286)
8.4
(273)
8.8
(284)
15.8
(284)
12.9
(271)
Other 8.4
(104)
10.6
(104)
14.0
(104)
20.8
(104)
9.6
(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
Appendix Table A.5.
Percentages of 11-year-old children with severe problems (parent-rated) by
child ethnicity and SDQ subscale: United Kingdom

Centre for Mental Health REPORT Children of the new century
34
Girls
White 9.2
(6,574)
7.3
(6,572)
12.1
(6,572)
10.6
(6,573)
7.8
(6,571)
Mixed 13.6
(256)
11.6
(255)
19.9
(256)
15.4
(256)
13.8
(255)
Indian 3.9
(126)
6.7
(124)
6.5
(125)
7.7
(126)
5.5
(124)
Pakistani/
Bangladeshi
7.3
(360)
4.9
(354)
11.9
(357)
12.5
(359)
6.3
(354)
Black 3.4
(234)
6.1
(231)
7.9
(232)
10.0
(232)
6.9
(231)
Other 1.4
(98)
4.6
(98)
12.4
(98)
14.8
(98)
6.1
(98)
χ² difference 23.88* 60.71*** 18.66 8.73 13.15
Unweighted N 6,340 6,322 6,333 6,336 6,320
Conduct
Problems
Hyperactive/
Inattention
Emotional
Problems
Peer
Problems
Total
Problems
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 reect
sample losses (Mostafa, 2014).
Appendix Table A.5. continued

35
Centre for Mental Health REPORT Children of the new century
Girls
White 9.2
(6,574)
7.3
(6,572)
12.1
(6,572)
10.6
(6,573)
7.8
(6,571)
Mixed 13.6
(256)
11.6
(255)
19.9
(256)
15.4
(256)
13.8
(255)
Indian 3.9
(126)
6.7
(124)
6.5
(125)
7.7
(126)
5.5
(124)
Pakistani/
Bangladeshi
7.3
(360)
4.9
(354)
11.9
(357)
12.5
(359)
6.3
(354)
Black 3.4
(234)
6.1
(231)
7.9
(232)
10.0
(232)
6.9
(231)
Other 1.4
(98)
4.6
(98)
12.4
(98)
14.8
(98)
6.1
(98)
χ² difference 23.88* 60.71*** 18.66 8.73 13.15
Unweighted N 6,340 6,322 6,333 6,336 6,320
Conduct
Problems
Hyperactivity/
Inattention
Emotional
Problems
Peer
Problems
Total
Problems
All children
Both natural
parents
7.1
(9,530)
8.0
(9,511)
9.9
(9,520)
9.9
(9,528)
6.6
(9,506)
Lone natural
parent
17.4
(4,130)
15.4
(4,122)
15.1
(4,124)
15.9
(4,127)
15.0
(4,116)
Step and other
family types
19.5
(2,110)
20.0
(2,099)
14.8
(2,110)
17.9
(2,109)
18.1
(2,097)
χ2difference 366.13*** 264.10*** 78.55*** 129.17*** 303.00***
Unweighted N 12,798 12,761 12,782 12,790 12,751
Boys
Both natural
parents
8.6
(4,844)
11.0
(4,833)
9.8
(4,837)
11.0
(4,843)
8.2
(4,828)
Lone natural
parent
20.5
(2,157)
20.0
(2,154)
13.9
(2,153)
18.0
(2,155)
17.9
(2,149)
Step and other
family types
24.0
(1,118)
26.7
(1,107)
15.7
(1,118)
21.1
(1,116)
22.5
(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
parents
5.6
(4,686)
4.8
(4,678)
9.9
(4,683)
8.8
(4,685)
5.0
(4,678)
Lone natural
parent
14.0
(1,973)
10.4
(1,968)
16.5
(1,970)
13.7
(1,972)
11.9
(1,967)
Step and other
family types
14.3
(992)
12.6
(992)
13.8
(992)
14.4
(992)
13.1
(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 reect sample losses (Mostafa, 2014).
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

Centre for Mental Health REPORT Children of the new century
36
Conduct
Problems
Hyperactivity/
Inattention
Emotional Problems Peer Problems Total
Problems
All children
GCSE at grade
D-G
18.5
(847)
18.7
(846)
16.7
(846)
18.0
(846)
17.5
(846)
GCSE at grade
A*-C
11.8
(3,721)
12.8
(3,719)
12.0
(3,718)
12.8
(3,721)
10.6
(3,717)
AS and
A-level
9.7
(1,079)
11.7
(1,078)
12.0
(1,078)
9.7
(1,079)
9.2
(1,078)
Degree/diploma 6.7
(4,067)
7.3
(4,065)
8.1
(4,064)
8.1
(4,066)
5.5
(4,062)
Higher degree 2.9
(1,614)
5.4
(1,614)
7.7
(1,614)
8.4
(1,614)
3.9
(1,614)
Other overseas
qualications
14.8
(597)
11.5
(595)
12.6
(595)
13.7
(596)
11.0
(593)
No academic
qualications
18.6
(3,139)
15.0
(3,111)
16.4
(3,133)
18.7
(3,136)
16.3
(3,106)
χ2difference 346.11*** 179.47*** 134.63*** 196.81*** 288.08***
Unweighted N 12,346 12,311 12,330 12,338 12,301
Boys
GCSE at grade
D-G
19.1
(435)
23.0
(435)
11.2
(435)
16.0
(435)
16.4
(435)
GCSE at grade
A*-C
14.4
(1,943)
17.4
(1,942)
11.3
(1,940)
14.7
(1,943)
12.8
(1,940)
AS and
A-level
12.0
(537)
16.2
(537)
12.3
(537)
11.5
(537)
12.1
(537)
Degree/diploma 8.1
(2,126)
10.5
(2,125)
7.8
(2,124)
8.9
(2,124)
6.9
(2,122)
Higher degree 2.6
(821)
7.6
(821)
7.9
(821)
10.7
(821)
5.5
(821)
Other overseas
qualications
19.0
(312)
15.8
(312)
13.7
(310)
16.7
(311)
15.2
(310)
No academic
qualications
22.7
(1,570)
20.5
(1,549)
17.7
(1,566)
21.7
(1,568)
20.8
(1,544)
χ2difference 219.77*** 108.07*** 81.78*** 109.50*** 166.47***
Unweighted N 6,222 6,204 6,213 6,218 6,196
Appendix Table A.7.
Percentages of 11-year-old children with severe problems (parent-rated) by highest
parent academic qualification and SDQ subscale: United Kingdom, 2012

37
Centre for Mental Health REPORT Children of the new century
Conduct
Problems
Hyperactivity/
Inattention
Emotional
Problems
Peer
Problems
Total
Problems
Girls
GCSE at grade
D-G 17.7
(412) 14.1
(411) 22.6
(411) 20.2
(411) 18.6
(411)
GCSE at grade
A*-C 8.8
(1,978) 7.8
(1,977) 12.8
(1,977) 10.7
(1,978) 8.2
(1,977)
AS and
A-level 7.4
(542) 7.3
(541) 11.7
(541) 7.9
(542) 6.3
(541)
Degree/di-
ploma 5.1
(1,941) 3.7
(1,941) 8.5
(1,941) 7.1
(1,941) 4.0
(1,940)
Higher degree 3.2
(793) 3.2
(793) 7.4
(793) 6.1
(793) 2.2
(793)
Other overseas
qualications 10.2
(285) 6.8
(284) 11.4
(285) 10.4
(285) 6.3
(284)
No academic
qualications 14.4
(1,570) 9.5
(1,562) 15.0
(1,567) 15.6
(1,569) 11.8
(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 reect sample losses (Mostafa, 2014).
(ii) ‘Other overseas qualications’ are included with ‘GCSE at grade D-G’ as ‘Low qualications’ in
Figure 4.4.

Centre for Mental Health REPORT Children of the new century
38
Conduct
Problems
Hyperactivity/
Inattention
Emotional
Problems
Peer
Problems
Total
Problems
All children
Professional/
Managerial
5.3
(3,383)
6.8
(3,382)
7.7
(3,382)
7.9
(3,382)
4.4
(3,381)
Intermediate 7.3
(1,939)
8.9
(1,937)
8.4
(1,937)
9.6
(1,939)
6.3
(1,937)
Small
Employer
6.5
(1930)
8.4
(1927)
9.3
(1928)
9.8
(1930)
5.7
(1926)
Lower
Supervisory
12.5
(873)
10.1
(871)
12.5
(873)
13.2
(872)
11.0
(871)
Routine/
Semi-
Routine
11.1
(3,772)
11.4
(3,756)
12.5
(3,769)
11.8
(3,770)
10.7
(3,753)
Workless 24.8
(2,575)
20.6
(2,564)
19.5
(2,567)
22.7
(2,573)
21.5
(2,558)
χ2difference 553.42*** 264.38*** 194.54*** 281.72*** 468.97***
Unweighted N 11,843 11,811 11,829 11,835 11,801
Boys
Professional/
Managerial
6.0
(1,741)
9.6
(1,740)
7.4
(1,740)
9.3
(1,740)
5.8
(1,739)
Intermediate 8.3
(979)
11.9
(979)
8.7
(979)
11.6
(979)
7.2
(979)
Small
Employer
9.4
(992)
11.4
(992)
9.0
(992)
11.7
(992)
7.6
(991)
Lower
Supervisory
15.5
(455)
14.5
(454)
12.8
(455)
15.4
(455)
13.4
(454)
Routine/
Semi-
Routine
12.5
(1,936)
15.7
(1,922)
12.4
(1,934)
13.0
(1,934)
12.6
(1,919)
Workless 30.3
(1,354)
27.6
(1,349)
18.8
(1,348)
25.3
(1,353)
27.2
(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
Appendix Table A.8.
Percentages of 11-year-old children with severe problems (parent-rated) by highest
occupation held by parents at MCS5 (if working) and SDQ subscale: United Kingdom,
2012

39
Centre for Mental Health REPORT Children of the new century
Conduct
Problems
Hyperactivity/
Inattention
Emotional
Problems
Peer
Problems
Total
Problems
All children
Professional/
Managerial
5.3
(3,383)
6.8
(3,382)
7.7
(3,382)
7.9
(3,382)
4.4
(3,381)
Intermediate 7.3
(1,939)
8.9
(1,937)
8.4
(1,937)
9.6
(1,939)
6.3
(1,937)
Small
Employer
6.5
(1930)
8.4
(1927)
9.3
(1928)
9.8
(1930)
5.7
(1926)
Lower
Supervisory
12.5
(873)
10.1
(871)
12.5
(873)
13.2
(872)
11.0
(871)
Routine/
Semi-
Routine
11.1
(3,772)
11.4
(3,756)
12.5
(3,769)
11.8
(3,770)
10.7
(3,753)
Workless 24.8
(2,575)
20.6
(2,564)
19.5
(2,567)
22.7
(2,573)
21.5
(2,558)
χ2difference 553.42*** 264.38*** 194.54*** 281.72*** 468.97***
Unweighted N 11,843 11,811 11,829 11,835 11,801
Boys
Professional/
Managerial
6.0
(1,741)
9.6
(1,740)
7.4
(1,740)
9.3
(1,740)
5.8
(1,739)
Intermediate 8.3
(979)
11.9
(979)
8.7
(979)
11.6
(979)
7.2
(979)
Small
Employer
9.4
(992)
11.4
(992)
9.0
(992)
11.7
(992)
7.6
(991)
Lower
Supervisory
15.5
(455)
14.5
(454)
12.8
(455)
15.4
(455)
13.4
(454)
Routine/
Semi-
Routine
12.5
(1,936)
15.7
(1,922)
12.4
(1,934)
13.0
(1,934)
12.6
(1,919)
Workless 30.3
(1,354)
27.6
(1,349)
18.8
(1,348)
25.3
(1,353)
27.2
(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
Girls
Professional/
Managerial
4.5
(1,642)
3.8
(1,642)
8.1
(1,642)
6.4
(1,642)
2.9
(1,642)
Intermediate 6.3
(960)
5.9
(959)
8.0
(959)
7.5
(960)
5.3
(958)
Small
Employer
3.5
(938)
5.2
(935)
9.5
(936)
7.9
(938)
3.6
(935)
Lower
Supervisory
9.2
(418)
5.3
(417)
12.3
(418)
10.8
(417)
8.6
(417)
Routine/
Semi-Routine
9.6
(1,837)
7.0
(1,834)
12.5
(1,835)
10.5
(1,836)
8.8
(1,834)
Workless 18.7
(1,221)
12.9
(1,215)
20.2
(1,218)
19.9
(1,219)
15.2
(1,214)
χ2difference 194.38*** 83.69*** 103.11*** 132.95*** 158.20***
Unweighted N 5,864 5,847 5,857 5,860 5,845
Conduct
Problems
Hyperactivity/
Inattention
Emotional
Problems
Peer
Problems
Total
Problems
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 reect
sample losses (Mostafa, 2014).
(ii) This is not a standard social class variable as the workless are not classied to their previous
occupation.

Centre for Mental Health REPORT Children of the new century
40
Conduct
Problems Hyperactivity/
Inattention Emotional
Problems Peer Problems Total
Problems
All children
Bottom 20.8
(3,282) 17.2
(3,262) 17.0
(3,269) 19.0
(3,279) 17.0
(3,255)
Second 17.0
(3,407) 15.6
(3,404) 13.9
(3,405) 16.8
(3,404) 15.6
(3,399)
Third 8.9
(3,110) 9.9
(3,097) 11.2
(3,109) 9.8
(3,110) 8.3
(3,097)
Fourth 5.4
(3,003) 7.9
(3,000) 9.5
(3,002) 8.6
(3,003) 5.5
(3,000)
Top 3.7
(2,969) 6.1
(2,969) 7.1
(2,969) 7.5
(2,969) 4.1
(2,969)
χ2difference 561.01*** 234.05*** 145.35*** 252.31*** 379.78***
Unweighted N 12,798 12,761 12,782 12,790 12,751
Boys
Bottom 25.5
(1,678) 23.8
(1,667) 17.0
(1,670) 21.3
(1,678) 21.8
(1,661)
Second 20.3
(1,748) 19.8
(1,747) 14.2
(1,746) 19.6
(1,745) 18.7
(1,743)
Third 10.9
(1,599) 13.2
(1,588) 9.7
(1,598) 9.6
(1,599) 9.6
(1,588)
Fourth 6.5
(1,581) 10.1
(1,578) 9.6
(1,581) 10.7
(1,581) 6.6
(1,578)
Top 4.6
(1,513) 9.7
(1,513) 7.2
(1,513) 8.6
(1,513) 5.6
(1,513)
χ2difference 352.36*** 152.83*** 77.95*** 154.09*** 251.29***
Unweighted N 6,458 6,439 6,449 6,454 6,431
Girls
Bottom 15.9
(1,604) 10.2
(1,595) 17.1
(1,599) 16.5
(1,601) 12.0
(1,594)
Second 13.5
(1,660) 11.1
(1,657) 13.6
(1,659) 13.8
(1,659) 12.2
(1,657)
Third 6.7
(1,511) 6.4
(1,509) 12.8
(1,511) 10.0
(1,511) 6.9
(1,509)
Fourth 4.2
(1,422) 5.4
(1,422) 9.4
(1,421) 6.2
(1,422) 4.3
(1,421)
Top 2.7
(1,456) 2.4
(1,456) 6.9
(1,456) 6.4
(1,456) 2.6
(1,456)
χ2difference 212.48*** 96.69*** 72.95*** 17.82*** 135.71***
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 reect sample losses (Mostafa, 2014).
Appendix Table A.9.
Percentages of 11-year-old children with severe problems (parent-rated) by
equivalised income quintiles at MCS5 and SDQ subscale: United Kingdom, 2012

41
Centre for Mental Health REPORT Children of the new century
BCAMHS
1999
(N=1904)
BCAMHS
2004
(N=1348)
MCS
2012
(N=11,397)
Child Age
10 50.6 50.6 34.0
11 49.4 49.4 66.0
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.
Appendix Table A.10.
Socio-demographic characteristics of three study samples of children aged 10
and 11: Great Britain

Centre for Mental Health REPORT Children of the new century
42
SDQ BCAMHS
1999
Mean(SD)
Condence
Interval BCAMHS
2004
Mean(SD)
Condence
Interval MCS
2012
Mean(SD)
Condence
Interval Cohort
Difference
(p < .05)*
Cohen’s
d*
Conduct Problems
Boys 1.66(1.75) 1.54-1.77 1.59(1.87) 1.45-1.73 1.69(1.82) 1.63-1.76
Girls 1.30(1.55) 1.20-1.40 1.25(1.53) 1.13-1.37 1.33(1.50) 1.27-1.39
Total 1.48(1.66) 1.40-1.55 1.43(1.72) 1.34-1.52 1.52(1.68) 1.47-1.57
Hyperactivity/
inattention
Boys 3.96(2.72) 3.79-4.13 3.65(2.71) 3.45-3.85 3.76(2.61) 3.67-3.86 1>2, 3 .12, .08
Girls 2.85(2.41) 2.70-3.01 2.55(2.36) 2.36-2.73 2.79(2.30) 2.69-2.89 1,3 >2 .13, .09
Total 3.41(2.62) 3.29-3.53 3.12(2.61) 2.98-3.26 3.29(2.51) 3.22-3.36 1>2 .11
Emotional Problems
Boys 1.86(2.01) 1.73-1.98 1.98(2.00) 1.82-2.14 1.90(2.03) 1.83-1.97
Girls 2.13(2.09) 1.99-2.26 1.96(1.99) 1.82-2.07 1.99(2.03) 1.90-2.06
Total 1.99(2.05) 1.90-2.08 1.98(2.05) 1.87-2.09 1.94(2.02) 1.88-1.99
Peer Problems
Boys 1.59(1.80) 1.47-1.70 1.55(1.80) 1.42-1.69 1.57(1.86) 1.50-1.63
Girls 1.44(1.61) 1.34-1.54 1.24(1.45) 1.13-1.36 1.35(1.65) 1.28-1.41 1>2 .12
Total 1.51(1.71) 1.44-1.59 1.41(1.65) 1.32-1.49 1.47(1.77) 1.42-1.52
Total Problems
Boys 9.06(5.97) 8.68-9.44 8.76(6.37) 8.29-9.23 8.93(6.45) 8.69-9.16
Girls 7.72(5.44) 7.37-8.07 7.02(5.52) 6.59-7.45 7.45(5.66) 7.22-7.68 1>2 .13
Total 8.39(5.75) 8.14-8.65 7.93(6.04) 7.61-8.26 8.21(6.13) 8.03-8.39 1>2 .08
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.
MCS SDQ data were weighted so that 10- and 11-year-old children represented equivalent proportions in all analyses.
*Signicant effects only. Cohen’s d is an effect size used to indicate the standardised difference between two means.
Appendix Table A 11.
Parent-rated means and standard deviations of SDQ scores for 10- and 11-year-olds in 1999, 2004 and 2012: Great Britain

43
Centre for Mental Health REPORT Children of the new century
SDQ BCAMHS
1999
Mean(SD)
Condence
Interval BCAMHS
2004
Mean(SD)
Condence
Interval MCS
2012
Mean(SD)
Condence
Interval Cohort
Difference
(p < .05)*
Cohen’s
d*
Conduct Problems
Boys 1.66(1.75) 1.54-1.77 1.59(1.87) 1.45-1.73 1.69(1.82) 1.63-1.76
Girls 1.30(1.55) 1.20-1.40 1.25(1.53) 1.13-1.37 1.33(1.50) 1.27-1.39
Total 1.48(1.66) 1.40-1.55 1.43(1.72) 1.34-1.52 1.52(1.68) 1.47-1.57
Hyperactivity/
inattention
Boys 3.96(2.72) 3.79-4.13 3.65(2.71) 3.45-3.85 3.76(2.61) 3.67-3.86 1>2, 3 .12, .08
Girls 2.85(2.41) 2.70-3.01 2.55(2.36) 2.36-2.73 2.79(2.30) 2.69-2.89 1,3 >2 .13, .09
Total 3.41(2.62) 3.29-3.53 3.12(2.61) 2.98-3.26 3.29(2.51) 3.22-3.36 1>2 .11
Emotional Problems
Boys 1.86(2.01) 1.73-1.98 1.98(2.00) 1.82-2.14 1.90(2.03) 1.83-1.97
Girls 2.13(2.09) 1.99-2.26 1.96(1.99) 1.82-2.07 1.99(2.03) 1.90-2.06
Total 1.99(2.05) 1.90-2.08 1.98(2.05) 1.87-2.09 1.94(2.02) 1.88-1.99
Peer Problems
Boys 1.59(1.80) 1.47-1.70 1.55(1.80) 1.42-1.69 1.57(1.86) 1.50-1.63
Girls 1.44(1.61) 1.34-1.54 1.24(1.45) 1.13-1.36 1.35(1.65) 1.28-1.41 1>2 .12
Total 1.51(1.71) 1.44-1.59 1.41(1.65) 1.32-1.49 1.47(1.77) 1.42-1.52
Total Problems
Boys 9.06(5.97) 8.68-9.44 8.76(6.37) 8.29-9.23 8.93(6.45) 8.69-9.16
Girls 7.72(5.44) 7.37-8.07 7.02(5.52) 6.59-7.45 7.45(5.66) 7.22-7.68 1>2 .13
Total 8.39(5.75) 8.14-8.65 7.93(6.04) 7.61-8.26 8.21(6.13) 8.03-8.39 1>2 .08
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.
MCS SDQ data were weighted so that 10- and 11-year-old children represented equivalent proportions in all analyses.
*Signicant effects only. Cohen’s d is an effect size used to indicate the standardised difference between two means.
SDQ BCAMHS
1999
Mean(SD)
Condence
Interval BCAMHS
2004
Mean(SD)
Condence
Interval MCS
2012
Mean(SD)
Condence
Interval Cohort
Difference
(t-test)*
Cohen’s d*
Conduct Problems
Boys 1.27(1.97) 1.12-1.41 1.18(1.90) 1.02-1.35 1.07(1.74) .98-1.16 1>3 .10
Girls .58(1.36) .48-.68 .50(1.13) .40-.60 .44(1.10) .40-.51 1>3 .09
Total .92(1.72) .83-1.01 .86(1.62) .76-.96 .76(1.50) .72-.82 1>3 .09
Hyperactivity/
inattention
Boys 3.64(2.99) 3.42-3.86 3.40(2.96) 3.14-3.66 3.31(2.82) 3.16-3.46 1>3 .11
Girls 2.04(2.35) 1.86-2.21 1.86(2.30) 1.65-2.07 1.57(1.95) 1.47-1.67 1>3 .21
Total 2.83(2.81) 2.68-2.97 2.67(2.77) 2.50-2.85 2.45(2.57) 2.35-2.55 1>3 .14
Emotional Problems
Boys 1.43(1.92) 1.28-1.57 1.50(2.06) 1.32-1.68 1.44(1.91) 1.34-1.54
Girls 1.39(1.94) 1.25-1.54 1.54(2.06) 1.35-1.73 1.46(1.92) 1.36-1.56
Total 1.41(1.93) 1.31-1.51 1.52(2.06) 1.39-1.65 1.45(1.91) 1.38-1.52
Peer Problems
Boys 1.62(1.95) 1.47-1.76 1.53(2.00) 1.35-1.70 1.43(1.85) 1.33-1.53 1>3 .10
Girls 1.12(1.57) 1.01-1.24 1.17(1.76) 1.02-1.33 1.02(1.59) .94-1.10
Total 1.37(1.79) 1.29-1.44 1.36(1.85) 1.24-1.48 1.22(1.74) 1.15-1.29 1,2 >3 .08, .08
Total Problems
Boys 7.95(6.57) 7.47-8.43 7.61(6.85) 7.01-8.21 7.24(6.28) 6.91-7.57 1>3 .11
Girls 5.13(5.42) 4.74-5.53 5.08(5.33) 4.58-5.57 4.48(4.78) 4.23-4.73 1, 2>3 .15, .10
Total 6.53(6.18) 6.21-6.85 6.42(6.31) 6.02-6.82 5.88(5.76) 5.66-6.08 1,2>3 .11, .09
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.
*Signicant effects only. Cohen’s d is an effect size used to indicate the standardised difference between two means.
Appendix Table 12.
Teacher-rated means and standard deviations of SDQ scores for 10- and 11-year-olds in 1999, 2004 and 2012: England and Wales

Centre for Mental Health REPORT Children of the new century
44
SDQ BCAMHS
1999
BCAMHS
2004
MCS
2012
Cohort
Difference
(p < .05)
Conduct Problems
Boys 13.7 15.6 14.3
Girls 8.6 8.5 8.9
Total 11.2 12.2 11.7
Hyperactivity/ inattention
Boys 19.2 15.7 15.4 1>3
Girls 9.4 7.3 7.7
Total 14.4 11.7 11.8 1>2, 3
Emotional Problems
Boys 11.1 12.4 11.6
Girls 13.5 11.4 12.0
Total 12.3 11.9 11.9
Peer Problems
Boys 14.9 14.6 13.9
Girls 10.9 8.3 10.9 2<3
Total 12.9 11.6 12.8
Total Problems
Boys 11.9 13.0 13.1
Girls 6.9 6.2 8.0
Total 9.4 9.8 10.6
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. MCS SDQ data were weighted so that 10- and 11-year-old children
represented equivalent proportions in all analyses.
Appendix Table A.13.
Percentages of 10- and 11-year-olds with severe parent-rated SDQ scores in
1999, 2004 and 2012: Great Britain

45
Centre for Mental Health REPORT Children of the new century
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
1999
BCAMHS
2004
MCS
2012
Cohort
Difference
(p< .05)
Conduct Problems
Boys 13.5 12.5 10.5 1>3
Girls 5.7 3.3 3.2 1>3
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.

Centre for Mental Health REPORT Children of the new century
46
Boys Girls All children
None 75.4 83.1 79.3
Age 3 7.7 6.5 7.1
Age 5 0.8 0.6 0.7
Age 7 1.8 0.6 1.2
Age 11 3.5 3.1 3.3
Age 3/Age 5 1.2 0.7 0.9
Age 7/Age 11 1.3 0.8 1.0
Age 5/Age 7 0.3 0.2 0.3
Age 5/Age 11 0.5 0.2 0.4
Age 3/Age 7 1.0 1.0 1.0
Age 3/Age 11 1.3 1.0 1.1
Age 3/Age 5/Age 7 0.7 0.6 0.6
Age 3/Age 5/Age 11 0.5 0.2 0.4
Age 3/Age 7/Age 11 1.0 0.5 0.8
Age 5/Age 7/Age 11 1.0 0.3 0.6
Age 3/Age 5/Age 7/Age 11 2.0 0.6 1.3
N(5,758)
4,797
(5,586)
4,798
(11,344)
9,595
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.
Appendix Table A.15.
Percentages of MCS children with severe problems (total difficulties, parent ratings) at
ages 3 to 11: United Kingdom

47
Centre for Mental Health REPORT Children of the new century
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
Age 7 2.1 1.5 1.8
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)
4,959
(5,739)
4,932
(11,680)
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.
Appendix Table A.16.
Percentages of MCS children with severe conduct problems (parent ratings) at
ages 3 to 11: United Kingdom

Centre for Mental Health REPORT Children of the new century
48
Boys Girls All children
None 66.9 79.1 72.9
Age 3 7.0 5.9 6.4
Age 5 2.2 1.9 2.0
Age 7 3.9 2.3 3.1
Age 11 3.5 2.4 3.0
Age 3/Age 5 1.9 1.3 1.6
Age 7/Age 11 1.9 .9 1.4
Age 5/Age 7 1.2 1.2 1.2
Age 5/Age 11 .7 .2 .5
Age 3/Age 7 1.1 1.0 1.1
Age 3/Age 11 .7 .3 .5
Age 3/Age 5/Age 7 1.6 .8 1.2
Age 3/Age 5/Age 11 .5 .3 .4
Age 3/Age 7/Age 11 1.3 .5 .9
Age 5/Age 7/Age 11 2.1 .8 1.5
Age 3/Age 5/Age 7/Age 11 3.5 1.1 2.3
N(5,872)
4,891
(5,667)
4,869
(11,539)
9,760
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.
Appendix Table A.17.
Percentages of MCS children with severe hyperactivity/inattention problems
(parent ratings) at ages 3 to 11: United Kingdom

49
Centre for Mental Health REPORT Children of the new century
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
Age 7 3.1 1.8 2.5
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)
4,933
(5,709)
4,908
(11,622)
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.
Appendix Table A.18.
Percentages of MCS children with severe emotional problems (parent ratings) at
ages 3 to 11: United Kingdom

Centre for Mental Health REPORT Children of the new century
50
Boys Girls All children
None 71.3 77.3 74.2
Age 3 7.1 5.7 6.4
Age 5 2.2 1.7 1.9
Age 7 3.1 2.6 2.9
Age 11 5.3 5.0 5.2
Age 3/Age 5 1.0 1.0 1.0
Age 7/Age 11 2.0 1.2 1.6
Age 5/Age 7 .7 .7 .7
Age 5/Age 11 .7 .3 .5
Age 3/Age 7 .8 .8 .8
Age 3/Age 11 1.2 1.3 1.2
Age 3/Age 5/Age 7 .8 .3 .6
Age 3/Age 5/Age 11 .4 .4 .4
Age 3/Age 7/Age 11 .6 .6 .6
Age 5/Age 7/Age 11 1.3 .6 1.0
Age 3/Age 5/Age 7/Age 11 1.5 .5 1.0
N(5,881)
4,896
(5,698)
4,899
(11,579)
9,795
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.
Appendix Table A.19.
Percentages of MCS children with severe peer problems (parent ratings) at ages 3
to 11: United Kingdom

51
Centre for Mental Health REPORT Children of the new century

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.