Articles
Is there a Gender Difference in False
Belief Development?
Tony Charman, Institute of Child Health and Ted Ruffman and
Wendy Clements, University of Sussex
Abstract
The contribution of children’s social environment to their acquisition of theory of mind
skills, combined with the well documented advantage for girls in mental state talk with
siblings, peers and mothers, might lead to a female advantage on false belief tasks.
We present a post-hoc analysis of large datasets from two independent laboratories.
A slight advantage for girls on false belief task performance was found in both datasets
and was only apparent in younger but not older children. Language ability could be
controlled for only in a smaller subsample of one dataset and cannot be ruled out as
a potential mediator of this effect. However, if there is an age-specific advantage for
girls in false belief acquisition it is a weak effect only.
Keywords: theory of mind; false belief; gender
Research into individual differences in theory of mind performance has shown several
effects of the child’s social environment. Children with siblings show an advantage
over singletons (Perner, Ruffman, & Leekam, 1994), this advantage may be restricted
to the presence of older but not younger siblings (Ruffman et al., 1998), and may only
benefit verbally less able children (Jenkins & Astington, 1996). In addition, contact
with extended family members may confer an additional advantage over the presence
of siblings (Lewis et al., 1996). Parenting style is also associated with theory of mind
ability (Ruffman, Perner, & Parkin, 1999), and this effect shows some gender specificity
(Hughes, Deater-Deckard, & Cutting, 1999). Measures of attachment security
are also associated with theory of mind ability, although the direction of any causal
association remains to be determined (Fonagy, Redfern, & Charman, 1997; Meins et
al., 1998). It has been suggested that the facilitative effects of (older) siblings, family
members, parenting style and attachment relationships may operate via opportunities
for shared experiences of pretend play and deception, and talk about feelings and internal
mental states. Indeed, some specific evidence exists that co-operative sibling interaction
and frequency of mental state talk in dyadic play with friends is associated
© Blackwell Publishers Ltd 2002. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street,
Malden, MA 02148, USA.
Please address correspondence to either: Tony Charman, Behavioural and Brain Sciences Unit, Institute
of Child Health, 30 Guilford Street, London, WC1N 1EH, UK, E-mail: t.charman@ich.ucl.ac.uk, or: Ted
Ruffman, Experimental Psychology, University of Sussex, Brighton, East Sussex, BN1 9QG, UK. E-mail:
tedr@biols.susx.ac.uk
2 Tony Charman, Ted Ruffman and Wendy Clements
© Blackwell Publishers Ltd 2002 Social Development, 11, 1, 2002
longitudinally with false belief performance (Dunn et al., 1991; Hughes & Dunn,
1998). Further, the effects of such social interactions are not merely artefacts of language
ability, although this clearly plays a role (Dunn et al., 1991; Hughes & Dunn,
1998; Jenkins & Astington, 1996).
We know that gender differences exist in the amount and type of emotion and mental
state talk that occurs between parents (particularly mothers) and children, and between
child dyads. Mothers talk more to girls than to boys and more of this talk consists of
supportive speech acts, although this advantage may be restricted to toddlerhood rather
than the pre-school and early school age years (Leaper, Anderson, & Saunders, 1998).
Mothers talk more to daughters than sons about emotions at age 2, and the child’s own
level of emotion talk positively influences this association (Cervantes & Callanan,
1998; Dunn, Bretherton, & Munn, 1987). Older siblings also mention feeling states
more frequently to girls than boys, and both mother and sibling emotion and mental
state talk is associated with increased feeling state talk in 2-year-old girls compared
to boys (Brown, Donelan-McCall, & Dunn, 1996). The association between early
sibling interaction and later emotion understanding is stronger for girls than for boys
(Brown & Dunn, 1996).
Combining these two strands of research might lead us to expect an advantage for
girls over boys on theory of mind performance, at least in the earlier stages of acquisition.
Is there any evidence to support such a thesis? Some studies have shown an
advantage for girls on measures of mental state talk (Hughes & Dunn, 1998) and
emotion understanding (Dunn et al., 1991). Few studies have investigated gender influence
on theory of mind development as their primary aim. In a rare exception,
Banerjee (1997) found that girls outperformed boys in one of two conditions on an
emotion ‘appearance-reality’ task and on a task measuring understanding of emotion
display rules. Likewise, Bosacki and Astington (1999) found that 11-year-old girls
were better at assessing story characters’ motives and feelings than were boys. There
is some evidence of a female advantage in mindreading abilities in adulthood (BaronCohen
& Hammer, 1997; Baron-Cohen et al., 1997) but other studies have failed to
replicate this female advantage (Jarrold et al., 2000). However, these studies did not
examine false belief understanding, considered to be the ‘litmus test’of theory of mind
reasoning. Most studies that report a post-hoc gender comparison show no significant
advantage for girls or boys (e.g. Holmes, Black, & Miller, 1996; Jenkins &
Astington, 1996), with only a few exceptions (Cutting & Dunn, 1999; Happé, 1995).
However, if we assume that any theory of mind gender effect would be expected to
be a weak effect only, it may be that most previous research has lacked the statistical
power to demonstrate such an effect. A power analysis, employing a t-test calculation
with a variance inflation factor to model a power calculation for a sequential logistic
regression (Cohen, 1988; Hsieh, Bloch, & Larsen, 1998), demonstrates that a weak
effect would only have a 51% chance of being detected with alpha set at .05 in a total
sample of 200 boys and 200 girls with age entered first into the regression equation.
Power only reaches 88% with samples of 1000 children (500 boys, 500 girls). Theory
of mind research does not usually employ samples larger than 50 subjects (though see
Hughes and Cutting (1999) for an exception where in 119 same-sex twin pairs no
theory of mind advantage was found for girls). We wished to take advantage of the
large pool of false belief data residing in our laboratories in order to examine whether
post-hoc analysis of a sufficiently large sample would reveal a theory of mind advantage
for girls that might be expected from the argument set out above.
Data from each laboratory were analysed separately, for several reasons:
Gender Effects in False Belief 3
1. The task employed in one laboratory [Dataset 1—TC] was the deceptive box task
(Perner, Leekam, & Wimmer, 1987) and in the other [Dataset 2—TR, WC] it was
the location change task (Wimmer & Perner, 1983).
2. The two laboratories are located in different centres with different socio-economic
characteristics (London and Brighton, UK). Systematic biases in the samples from
the two laboratories might confound the results in any attempt to combine the two
datasets.
3. In one laboratory [Dataset 1], due to differences in the use and scoring of memory
and control questions across studies, it was possible only to record a pass/fail, with
the criterion adopted that in order to pass the child had to answer naming and reality
control questions correctly. In the other laboratory [Dataset 2] several other measures
were available which might help to establish whether any female advantage
on false belief tasks was independent of language ability. First, a direct measure of
verbal ability was also available on approximately one half of cases. Second, data
on three memory control questions was recorded separately. Memory questions are
of similar verbal complexity as the test question so that if the gender advantage is
linguistically mediated we would expect girls to show a substantial advantage on
memory questions as well as the test question.
In order to be able to compare gender effects independently in the two datasets but
also to examine any systematic age by gender interaction across the two datasets,
gender effects on false belief performance were studied independently in each dataset
split into age quartiles. If an advantage for girls was to be found we would expect it
at one agepoint only, since the advantages offered by the differential socialisation
effects outlined above would be transient—at least as measured by a single false belief
task. Boys would be expected to catch up once they had accrued the necessary social
experience and any continuing gender effects in theory of mind ability would be
masked by ceiling effects when measured by a single false belief task.
Method
Dataset 1
(i) Participants. 375 children (183 girls, 192 boys) aged between 2.33 years and 6.25
years (M = 4.24 years, SD = .82 years) from a number of playgroups, nurseries and
primary schools in the London area participated. Systematic data on ethnicity and
social background was not collected in all studies but the samples were approximately
60% Caucasian, 20% Afro-Caribbean and 20% Asian, and were predominantly from
lower middle-class and working class areas of inner and outer London.
For analysis the sample subdivided into age quartiles as follows: 1st quartile: 93
children (46 girls, 47 boys), mean age 3.20 years (SD = .29, range 2.33 to 3.50); 2nd
quartile: 95 children (45 girls, 50 boys), mean age 3.96 years (SD = .22, range 3.58
to 4.25); 3rd quartile: 86 children (39 girls, 47 boys), mean age 4.50 years (SD = .13,
range 4.33 to 4.67); 4th quartile: 101 children (53 girls, 48 boys), mean age 5.25 years
(SD = .42, range 4.75 to 6.25).
(ii) Deceptive box task. Children received a deceptive box false belief task, similar to
that developed by Perner et al. (1987)1
. The child was shown a familiar cereal carton
and asked the Naming Question: ‘What do you think is inside this?’ All the children
said ‘cereal’ or named the cereal brand. The carton was then opened and the child was
© Blackwell Publishers Ltd 2002 Social Development, 11, 1, 2002
4 Tony Charman, Ted Ruffman and Wendy Clements
© Blackwell Publishers Ltd 2002 Social Development, 11, 1, 2002
shown it really contained a mirror. The carton was then closed again, and the child
was asked the Reality Question: ‘Now what do you think is inside this carton?’ Most
children said ‘a mirror’, confirming that their belief had now changed. If the child was
unable to name the mirror the experimenter informed the child that there was a mirror
in the carton and asked them to name the contents again. The child was then asked
the Belief Question: ‘When I first showed you the carton, before we opened it, what
did you think was inside?’ If the child did not answer, the question was repeated but
took the form of ‘When I first showed you the carton, before we opened it, did you
think there was a mirror or cereal inside?’ (order of mirror and cereal counterbalanced
across children). In order to pass the task the child had to answer the Belief Question
and both Naming and Reality Control Questions correctly.
Dataset 2
(iii) Participants. 1093 children (558 girls, 535 boys) aged between 2.25 years and
6.17 years (M = 3.92 years, SD = .79 years) from a number of playgroups, nurseries
and primary schools in the Brighton area participated. Systematic data on ethnicity
and social background was not collected in all studies but the large majority of children
were Caucasian, and were from middle-class and lower middle class suburban
backgrounds.
For analysis the sample subdivided into age quartiles as follows: 1st quartile: 242
children (128 girls, 114 boys), mean age 2.97 years (SD = .24, range 2.25 to 3.32);
2nd quartile: 266 children (131 girls, 135 boys), mean age 3.52 years (SD = .12, range
3.33 to 3.72); 3rd quartile: 299 children (147 girls, 152 boys), mean age 4.04 years
(SD = .18, range 3.75 to 4.33); 4th quartile: 286 children (143 girls, 143 boys), mean
age 4.97 years (SD = .50, range 4.34 to 6.17).
Location change task:
Children received a location change false belief task, similar in structure to that
developed by Wimmer and Perner (1983)2
.The false belief story was enacted with dolls
in two model rooms. In the mouse story, Sam placed his cheese by the blue box and
then retreated through the nearby hole behind the wall and out of sight. In Sam’s
absence, a second mouse, Katy, moved the cheese to her red box. Before Sam returned
children were asked the Belief Question: ‘Where will Sam look for his cheese?’ and
three Memory Questions: Beginning ‘Where did Sam put his cheese in the beginning?’,
Reality ‘Where is the cheese now?’, See ‘Did Sam see Katy move the cheese?’
Data on answers to the Belief Question were available for all 1093 children, and
data on the three memory questions were available for 680 children (349 girls, 331
boys)3
. In addition, verbal mental age data from the British Picture Vocabulary Scale
(BPVS; Dunn et al., 1982) were available for 519 children (259 girls, 260 boys).
Results
Dataset 1
The number and percentage of girls and boys passing the deceptive box false belief
task at each age quartile are shown in Table 1. In order to test for the predicted agespecific
differences separate sequential logistic regressions were conducted on the
Gender Effects in False Belief 5
sample split by cumulative age quartile, with age entered into the prediction equation
first. Entering age first ensured that differences in mean age between boys and
girls were not responsible for gender effects. In the 1st
(youngest) quartile of the
sample neither age nor gender significantly predicted false belief performance (X2
(1, N = 93) = .018; X2
(1, N = 93) = .001, respectively, both p > .10). In the 1st
plus
2nd
quartiles combined age significantly predicted false belief performance and
there was a non-significant trend to gender predicting false belief performance (X2
(1, N = 188) = 15.0, p < .001; X2
(1, N = 188) = 3.35, p = .07, respectively). In
the 1st
, 2nd
and 3rd
quartiles combined age and gender significantly predicted
false belief performance (X2
(1, N = 274) = 37.6, p < .001; X2
(1, N = 274) = 5.94,
p < .05, respectively). In the whole sample age significantly predicted false belief
performance and there was a non-significant trend to gender predicting false belief
performance (X2
(1, N = 375) = 75.5, p < .001; X2
(1, N = 375) = 3.51, p = .06,
respectively).
Regression coefficients, odds ratios and 95% confidence intervals for odds ratios
for the predictor variables age and gender for Dataset 1 broken down by cumulative
age quartiles are shown in Table 2. The odds ratio indicates the relative size of the
effect of age and gender on false belief performance. That is, it shows the change in
the relative odds of a subject passing or failing the false belief task when the predictor
variable of interest changes by a unit of one (in this case one year for age, and
being a boy vs. a girl for gender). Comparison of the size of the odds ratios of the two
predictor variables indicates their relative strength, although this is more conceptually
difficult when one variable is continuous and the other is categorical, as in the present
case. In the 1st
, 2nd
and 3rd
age quartiles combined where the effect of both age and
gender significantly predicted false belief performance, the relative increase in likelihood
of a subject passing the false belief task was approximately 9 times higher for
each year of age than for being a girl rather than a boy.
© Blackwell Publishers Ltd 2002 Social Development, 11, 1, 2002
Table 1. The Number and Percentage (in Brackets) of Girls and Boys in Each
Dataset Broken Down by Age Quartiles Passing the False Belief Tasks
Girls Boys
Age quartile Fail N (%) Pass N (%) Fail N (%) Pass N (%)
Dataset 1
1st
quartile 38 (82.6%) 8 (17.4%) 39 (83.0%) 8 (17.0%)
2nd
quartile 19 (42.2%) 26 (57.8%) 32 (64.0%) 18 (36.0%)
3rd
quartile 11 (28.2%) 28 (71.8%) 21 (44.7%) 26 (55.3%)
4th
quartile 15 (28.3%) 38 (71.7%) 12 (25.0%) 36 (75.0%)
Dataset 2
1st
quartile 107 (83.6%) 21 (16.4%) 103 (90.4%) 11 (9.6%)
2nd
quartile 85 (64.9%) 46 (35.1%) 99 (73.3%) 36 (26.7%)
3rd
quartile 72 (49.0%) 75 (51.0%) 71 (46.7%) 81 (53.3%)
4th
quartile 38 (26.6%) 105 (73.4%) 31 (21.7%) 112 (78.3%)
6 Tony Charman, Ted Ruffman and Wendy Clements
© Blackwell Publishers Ltd 2002 Social Development, 11, 1, 2002
Dataset 2
The number and percentage of girls and boys passing the location change false belief
task at each age quartile are shown in Table 1. Again, separate sequential logistic
regressions were conducted on the sample divided by cumulative age quartiles,
with age entered into the prediction equation first. In the 1st
quartile neither age
nor gender predicted false belief performance, although both showed a nonsignificant
trend towards doing so (X2
(1, N = 242) = 2.94, p = .09; X2
(1, N = 242)
= 2.98, p = .08, respectively). In the 1st
and 2nd
quartiles combined both age and
gender predicted false belief performance (X2
(1, N = 508) = 27.7, p < .001; X2
(1,
N = 508) = 4.71, p < .05, respectively). In the 1st
, 2nd
and 3rd
quartiles combined and
in the whole sample age but not gender significantly predicted false belief performance
(X2
(1, N = 807) = 118.8, p < .001; X2
(1, N = 807) = 2.03, p > .10; and X2
(1,
N = 1093) = 288.8, p < .001; X2
(1, N = 1093) = .43, p > .10, respectively). Odds
ratios and other relevant statistics are shown in Table 2. In the 1st
and 2nd
age quartiles
combined where the effect of both age and gender significantly predicted false
belief performance, the relative increase in likelihood of a subject passing the false
belief task was approximately 12 times higher for each year of age than for being a
girl rather than a boy.
For a subset of this dataset we also had data on girls’ and boys’ performance on the
three memory questions and the BPVS. These effects were explored for the older and
Table 2. Regression Coefficients, Odds Ratios and 95% Confidence Intervals for
Odds Ratios for Age in Years and Gender in Each Dataset Broken by Cumulative
Age Quartiles
Age quartiles Predictor B Odds ratio 95% CI
Dataset 1
1st
quartile Age -.13 .88 .14 to 5.61
Gender -.02 .98 .33 to 2.89
1st
+ 2nd
quartile Age 1.50 4.50 2.05 to 9.90
Gender -.60 .55 .29 to 1.05
1st
+ 2nd
+ 3rd
quartile Age 1.51 4.51 2.70 to 7.52
Gender -.65 .52 .31 to .88
1st
+ 2nd
+ 3rd
+ 4th
quartile Age 1.28 3.58 2.58 to 4.96
Gender -.43 .65 .41 to 1.02
Dataset 2
1st
quartile Age 1.68 5.37 .83 to 34.9
Gender -.68 .51 .23 to 1.11
1st
+ 2nd
quartile Age 2.00 7.31 3.32 to 16.1
Gender -.48 .62 .40 to .96
1st
+ 2nd
+ 3rd
quartile Age 1.96 7.07 4.80 to 10.4
Gender -.23 .79 .58 to 1.09
1st
+ 2nd
+ 3rd
+ 4th
quartile Age 1.60 4.96 3.97 to 6.19
Gender -.09 .91 .69 to 1.20
Gender Effects in False Belief 7
younger half of the sample, split by median age. On the memory questions there was
no difference between girls’ and boys’ performance in either the younger (ANOVA;
F(1, 298) = .06, p = n.s.; girls: M = 2.33, SD = .89; boys: M = 2.30, SD = .92) or the
older half of the sample (ANOVA; F(1, 379) = .009, p = n.s.; girls: M = 2.36, SD =
.62; boys: M = 2.35, SD = .67). Similarly, BPVS scores showed no significant advantage
for girls over boys in either the younger (ANOVA; F(1, 164) = .15, p = n.s.; girls:
M = 41.5 months, SD = 10.0 months; boys: M = 40.8 months, SD = 11.1 months) or
the older half of the sample (ANOVA; F(1, 351) = 1.44, p = n.s.; girls: M = 57.4
months, SD = 15.5 months; boys: M = 55.6 months, SD = 13.6 months).
Discussion
A significant but weak female advantage was found for false belief task performance
in the 1st
, 2nd
and 3rd
cumulative age quartiles in Dataset 1 and a similar significant
but weak effect was found in the 1st
plus 2nd
age quartiles combined in Dataset 2. Over
both datasets, although girls did better than boys, the gender advantage was weak and
substantially weaker than the effect for age.
Two potentially complimentary explanations for this weak, age-specific theory of
mind advantage are apparent. First, previously identified gender differences in the
amount of supportive and emotion talk that mothers enter into with their offspring
(Cervantes & Callanan, 1998; Dunn et al., 1987; Leaper et al., 1998), and older siblings
with younger siblings (Brown & Dunn, 1996; Brown et al., 1996), both favour
girls over boys. This difference in social milieu and social experience may ‘boost’ girls
to acquire theory of mind understanding, on average, slightly earlier than boys. One
unknown is exactly how facilitation of false belief takes place. It is possible that even
if mothers do enhance false belief understanding, they do so without directly instructing
about beliefs. Perhaps the increased emotion talk with girls in non-disciplinary
situations (Brown et al., 1996; Cervantes & Callanan, 1998; Dunn et al., 1987;
Hughes et al., 1999), and in disciplinary situations (Ruffman et al., 1999), is sufficient
in getting children to think more carefully about social relations in general.
When measured by a single yardstick, as in this study, the advantage is merely temporary
and boys eventually acquire a sufficient quality and quantity of social experience
in order to understand first order false beliefs so that the gender difference
disappears for older children.
The second explanation is that there is an innate female advantage in mindreading
abilities (Baron-Cohen & Hammer, 1997; Skuse, 1998). These accounts draw on data
from both typically-developing adults and children, and from clinical groups who are
considered to show an ‘extreme’ male cognitive style—individuals with autism and
individuals with Turner’s syndrome who are monosomic with a maternally-acquired
X chromosome. Recent research has found a high heritability in theory of mind in
typically-developing pre-school, same-sex twin pairs, independent of verbal ability
(Hughes & Cutting, 1999).
The socialisation and nativist accounts may indeed be complementary in that a
genetic predisposition to engage in or understand certain types of material is likely to
influence a child’s environment (e.g. how much their mother or older siblings talks to
them about such matters) in ways that may increase the innately-specified advantage.
Such gene-environment interactions are becoming better understood in child development
(Plomin & Rutter, 1998), and gender differences in the development of social
and emotional understanding and any possible links to neuropsychiatric disorders will
© Blackwell Publishers Ltd 2002 Social Development, 11, 1, 2002
8 Tony Charman, Ted Ruffman and Wendy Clements
© Blackwell Publishers Ltd 2002 Social Development, 11, 1, 2002
be an important area for future study. However, the present findings cannot inform
this debate in terms of the relative contribution of, and role of gene-environment
interaction in, genetic and environmental influences on gender differences in
mental state understanding (see Hughes & Cutting, 1999, for a discussion of these
issues).
Another possible explanation for the gender effect is that the age-specific false belief
advantage for girls is mediated by general language ability. For instance, it has been
shown that language ability is related to theory of mind ability (Astington & Jenkins,
1999; de Villiers, 2000; Happé, 1995; Jenkins & Astington, 1996). In a subsample of
Dataset 2 we found that girls were not better on either the BPVS, which measures
receptive vocabulary, or on the memory questions that are of a similar verbal complexity
to the false belief test question. However, a recent view is that syntax is most
closely associated with theory of mind ability in both typical (de Villiers, 2000) and
atypical groups (Tager-Flusberg, 2000) and no direct measure of syntactic competence
was available in the present datasets. A variant on the language hypothesis is that mediation
is more evident in the early phase of language acquisition in comparison to the
latter phase (e.g. Bornstein & Haynes, 1998; Huttenlocher et al., 1991; Hyde & Linn,
1988). We cannot rule out this explanation with our data but future studies could
examine this by including more extensive tests of language abilities that are related
to belief understanding (Astington & Jenkins, 1999; de Villiers, 2000).
The strength of this study is in the very large numbers of children tested, overcoming
the insufficient statistical power in previous studies to identify a weak effect
that is apparent in a limited time window using a single task. With such large numbers
and with similar patterns across two independent laboratories we can be relatively
certain that our effects are representative of genuine gender differences. Yet the necessity
to test such large numbers to get the effect, and the fact that the effect for gender
is weak is equally important. In this study the analysis of gender was conducted posthoc
on existing datasets. Future research into gender effects on theory of mind development
should adopt prospective designs to test in more detail the a priori predictions
that follow from the psychocultural (Maccoby, 2000), psychosocial (Hughes & Dunn,
1998) and nativist (Baron-Cohen & Hammer, 1997) accounts. In addition this research
will need to take account of the potentially mediating role of factors such as language
development and parenting style.
What should one make of a relatively weak female advantage on theory of mind
tasks? On the positive side we note that the advantage, although weak, seems to persist
throughout much of development in that researchers have obtained such findings with
preschool children, pre-adolescent children, and young adults. Second, the advantage
has been linked to female advantages in social competence in pre-adolescents and adolescents
(Bosacki & Astington, 1999; Mathews & Keating, 1995), and to the largely
male condition of autism (Baron-Cohen & Hammer, 1997; Skuse, 2000). Thus, the
advantage, even if weak, might be of some consequence for both males and females.
Yet on the negative side, we think that many and possibly even most individuals will
not show a gender-typed pattern of performance. Gender effects are weak and might
perhaps be best construed as general tendencies with exceptions being the norm. Other
psychosocial factors such as family environment (Perner et al., 1994), parenting style
(Hughes et al., 1999; Ruffman et al., 1999) and interaction with siblings, other family
members and peers (Hughes & Dunn, 1998; Lewis et al., 1996) may be more important
in determining the rate and range of mentalising ability that individual children
acquire.
Gender Effects in False Belief 9
References
Astington, J. W. & Jenkins, J. (1999). A longitudinal study of the relation between language
and theory of mind development. Developmental Psychology, 35, 1311–1320.
Banerjee, M. (1997). Hidden emotions: pre-schoolers’ knowledge of appearance-reality and
emotion display rules. Social Cognition, 15, 107–132.
Baron-Cohen, S. & Hammer, J. (1997). Is autism an extreme form of the ‘male brain’?
Advances in Infancy Research, 11, 193–217.
Baron-Cohen, S., Jolliffe, T., Mortimore, C., & Robertson, M. (1997). Another advanced test
of theory of mind: Evidence from very high functioning adults with Autism or Asperger
Syndrome. Journal of Child Psychology and Psychiatry, 38, 813–822.
Bornstein, M. H. & Haynes, O. M. (1998). Vocabulary competence in early childhood: measurement,
latent construct and predictive validity. Child Development, 69, 654–671.
Bosacki, S. & Astington, J. W. (1999). Theory of mind in preadolescence: Relations between
social understanding and social competence. Social Development, 8, 237–255.
Brown, J. R. & Dunn, J. (1996). Continuities in emotion understanding from three to six years.
Child Development, 67, 789–802.
Brown, J. R., Donelan-McCall, N., & Dunn, J. (1996) Why talk about mental states? The significance
of children’s conversations with friends, siblings, and mothers. Child Development,
67, 836–849.
Cervantes, C. A. & Callanan, M. A. (1998). Labels and explanations in mother-child emotion
talk: age and gender differentiation. Developmental Psychology, 34, 88–98.
Cohen, D. (1988). Statistical power analysis for the behavioural sciences. Hillsdale, NJ: LEA.
Cutting, A. L. & Dunn, J. (1999). Theory of mind, emotion understanding, language, and family
background: Individual differences and interrelations. Child Development, 70, 853–865.
De Villiers, J. (2000). Language and theory of mind: What are the developmental relationships?
In S. Baron-Cohen, H. Tager-Flusberg, & D. Cohen (Eds.) Understanding other minds: Perspectives
from autism and developmental cognitive neuroscience (2nd Edn.), pp. 83–123.
Oxford: Oxford University Press.
Dunn, J., Brown, J., Slomkowski, C., Tesla, C., & Youngblade, L. (1991). Young children’s
understanding of other people’s feelings and beliefs: individual differences and their
antecedents. Child Development, 62, 1352–1366.
Dunn, J., Bretherton, I., & Munn, P. (1987). Conversations about feeling states between mothers
and their young children. Developmental Psychology, 23, 132–139.
Dunn, L. M., Dunn, L. M., Whetton, C., & Pintillie, D. (1982). British PictureVocabulary Scale.
London: NFER-Nelson.
Fonagy, P., Redfern, S., & Charman, T. (1997). Individual differences in theory of mind acquisition:
the role of attachment security. British Journal of Developmental Psychology, 15,
51–61.
Happé, F. G. E. (1995). The role of age and verbal ability in the theory of mind task performance
of subjects with autism. Child Development, 66, 843–855.
Happé, F. G. E., Winner, E., & Brownell, H. (1998). Linking theory of mind and central coherence
bias in autism and in the general population. Developmental Psychology, 34, 358–362.
Holmes, H. A., Black, C., & Miller, S. A. (1996). A cross-task comparison of false belief understanding
in a Head Start population. Journal of Experimental Child Psychology, 63, 263–285.
Hsieh, F. Y., Bloch, D. A., & Larsen, M. D. (1998). A simple method of sample size calculation
for linear and logistic regression. Statistics in Medicine, 17, 1623–1634.
Hughes, C. & Dunn, J. (1998). Understanding mind and emotion: Longitudinal associations
with mental-state talk between young friends. Developmental Psychology, 34, 1026–1037.
Hughes, C., Deater-Deckard, K., & Cutting, A. L. (1999). ‘Speak roughly to your little boy’?
Sex differences in the relations between parenting and preschooler’s understanding of mind.
Social Development, 8, 143–160.
Hughes, C. & Cutting, A. L. (1999). Nature, nurture and individual differences in early understanding
of mind. Psychological Science, 10, 429–432.
Huttenlocher, J., Haight, W., Bryk, A., Seltzer, M., & Lyons, T. (1991) Early vocabulary growth:
relation to language input and gender. Developmental Psychology, 27, 236–248.
Hyde, J. S. & Linn, M. C. (1988) Gender differences in verbal ability: a meta-analysis.
Psychological Bulletin, 104, 53–69.
© Blackwell Publishers Ltd 2002 Social Development, 11, 1, 2002
10 Tony Charman, Ted Ruffman and Wendy Clements
© Blackwell Publishers Ltd 2002 Social Development, 11, 1, 2002
Jarrold, C., Butler, D. W., Cottington, E. M., & Jimenez, F. (2000). Linking theory of mind and
central coherence bias in autism and in the general population. Developmental Psychology,
36, 126–138.
Jenkins, J. M. & Astington, J. W. (1996). Cognitive factors and family structure associated with
theory of mind development in young children. Developmental Psychology, 32, 70–78.
Leaper, C., Anderson, K. J., & Sanders, P. (1998). Moderators of gender effects on parents’ talk
to their children: a meta-analysis. Developmental Psychology, 34, 3–27.
Lewis, C., Freeman, N. H., Kyriakidou, C., Maridaki Kassotaki, K., & Berridge, D. M. (1996).
Social influences on false belief access: specific sibling influences or general apprenticeship?
Child Development, 67, 2930–2947.
Maccoby, E. E. (2000). Parenting and its effects on children: On reading and misreading behavior
genetics. Annual Review of Psychology, 51, 1–27.
Mathews, D. & Keating, D. (1995). Domain specificity and habits of mind: An investigation of
patterns of high-level development. Journal of Early Adolescence, 15, 319–343.
Meins, E., Fernyhough, C., Russell, J., & Clark-Carter, D. (1998). Security of attachment as a
predictor of symbolic and mentalising abilities: a longitudinal study. Social Development, 7,
1–24.
Perner, J., Ruffman, T., & Leekam, S. R. (1994). Theory of mind is contagious: You catch it
from your sibs. Child Development, 65, 1228–1238.
Plomin, R. & Rutter, M. (1998). Child development, molecular genetics, and what to do with
genes once they are found. Child Development, 69, 1223–1242.
Ruffman, T., Perner, J., Naito, M., Parkin, L., & Clements, W. A. (1998). Older (but not younger)
siblings facilitate false belief understanding. Developmental Psychology, 34, 161–174.
Ruffman, T., Perner, J., & Parkin, L. (1999). How parenting style affects false belief development.
Social Development, 8, 395–411.
Skuse, D. H. (2000). Imprinting, the X-chromosome and the male brain: Explaining sex differences
in the liability to autism. Paediatric Research, 47, 9–16.
Tager-Flusberg, H. (2000). Language and understanding minds: Connections in autism. In
S. Baron-Cohen, H. Tager-Flusberg, & D. Cohen (Eds.) Understanding other minds: Perspectives
from autism and developmental cognitive neuroscience (2nd Edn.), pp. 124–149.
Oxford: Oxford University Press.
Wimmer, H. & Perner, J. (1983). Beliefs about beliefs: Representation and constraining function
of wrong beliefs in young children’s understanding of deception. Cognition, 13,
103–128.
Notes
1. Across the pooled dataset the contents and the container varied (a Smarties box containing a pencil,
an eggbox containing a tomato, a Bandaid box containing an eraser) but the procedure and the test questions
were invariant.
2. Across the pooled dataset the characters and the containers used, and the contents which are moved,
varied but all had an invariant structure in which one character placed something in one of the two containers
and was absent when the second story character moved the contents to the second container.
3. The proportion of subjects in the older and younger halves of the sample for whom memory
question responses were recorded varied, as did the proportion who completed the BPVS. The same
age quartile boundaries are used in order that children remain in the same age group across the 2 sets
of analyses.