British Josrnal of Developntental Psychology (1998), 16, 539-55 1 Printed in Great Britain 539

0 1998 The British Psychological Society

Children’s rule discrimination within the context

of the school

Eithne Buchanan-Barrow” and Martyn Barrett

Department 4Psycbology, University of Surrty, Gtddfird, Surrey G U 2 S X H , UK

Following Turiel’s (1983) proposal of two distinct conceptual domains of social

thinking, the moral and the social-conventional, research has revealed that children
acquire competent rule-differentiation abilities at a very early age within the family
context (Smetana & Braeges, 1990;Smetana, Schlagman & Adams, 1993). This study
investigated children’s rule-understanding within the context of the school by
examining children’s judgements of a range of infringements by schoolchildren,
involving both moral and social-conventional rules. Children, aged 5-1 1 years, from
four primary schools, assessed six hypothetical rule-breaking scenarios, two moral and
four socio-conventional, according to the traditional criteria. However, in order to
probe the children’s differentiation skills beyond a simple dichotomy between moral
and socio-conventional, the four socio-conventional infringements were sub-divided
between two rules with discernible purpose and two rules where the function was
unclear or more arbitrary. There were two main findings: (1) the children displayed
competent and complex rule-discrimination abilities with minimal age differences,
not only differentiating between the moral and socio-conventional but also perceiving
significant differences between the two types of socio-conventional rules; (2) the
children also revealed a clear sense of the authority competencies of head teachers and
teachers, again largely unaffected by age. However, despite these overall levels of rule
understanding, there were some individual differences in the moral thinking of some
of the younger children, possibly as a consequence of aspects of the school system.

Turiel(1783) proposed the existence of two separate domains of social thinking, the moral
and the social-conventional. While all rule understanding is based ultimately on factual
knowledge about the social world, moral rule understanding does not depend on this
informational content alone, but derives its force from an important and highly-valued set
of principles, which are held to be universal, unalterable and independent of rule,
authority or context. By contrast conventional rules are held to be constructs of certain
social groupings and must be learned. They are also relative, being dependent on context,
rule or authority and can be changed. The ability to differentiate between these two rule
domains is believed to play an essential part in the overall comprehension of the social
environment.
Research suggests that children learn to make this important distinction at a very early
age, through social interactions within the family (Smetana & Braeges, 1770, Smetana,

* Requests for reprints.

540 Eithne Buchanan-Barrow and Martyn Barrett
Schlagman & Adams, 1993). Further studies with older children have examined the ways
in which they fine-tune their capacity to differentiate in middle childhood, both in the
family and in school (Blumenfeld, Pintrich & Hamilton, 1987; Davidson, Turiel & Black,
1983; Laupa and Turiel 1986, 1993; Laupa, 1991, 1995; Tisak & Turiel, 1984, 1988;
Weston & Turiel, 1980) as they extend their understanding to more complex situations,
giving consideration to such aspects as the type of rule, consequences of action, status of
authority and social-organizational function.
The specific role of context in children’s rule-understanding development, however,
has been relatively under-researched. Socio-conventional rule understanding is evidently
dependent on the society to which it applies, and in each new context, children must
internalize a fresh set of rules. However, even moral rule understanding, despite the belief
that it is context-free once established, is not acquired in a vacuum. Children first learn
about rules in a very specific context, that of the family, in interaction with their care
givers and their peers (Dunn & Munn, 1985). Therefore, it may well be that the child’s
entrance into a new social context, such as the school, affects many aspects of their rule
understanding.
Furthermore, after the family, the school is one of the most important contexts in
which children develop and yet the context of the school is not only an unfamiliar one, it
is also an extremely complex one. Children require time to make sense of it and to grasp
the interconnections between the various different aspects (Buchanan-Barrow & Barrett,
1996, 1998; Emler, Ohana & Moscovici, 1987). Authority is an important aspect of
school life and children have to grapple with the intricacies of the organizational and
power structure of the school, with a number of adults wielding considerable authority
over them. Previous research has revealed how children’s thinking about authority is
influenced by social organizational concepts (Laupa, 1991; Laupa & Turiel, 1986; 1993)
becoming increasingly differentiated with respect to authority status and function with
age (Laupa, 1995).In addition, there is also a direct connection linking those in authority,
the head teacher and other teachers, with the school-rules; the exercise of power, which
helps to define those in authority, is largely achieved through the existence of the rules
which those in authority seek to enforce. The comprehension of rules is therefore
inextricably enmeshed with the understanding of power or authority.
The complexity and variety of school rules, however, must present children with a
considerable task of comprehension. Some school rules do have clear moral import, such
as those concerned with protecting others from harm or with stealing, thus echoing rules
which will have been internalized by the children prior to school entry, and consequently
be relatively easy to grasp. However, the socio-conventional domain includes a large
diversity of rules designed for particular contexts and, as their wrongness is less evident
and less likely to be elucidated, they make greater demands on children’s comprehension
than moral rules (Smetana, 1993). In addition, while the children’s rule understanding
would be facilitated by knowledge of the purpose of the rules, the considerable number
and the variety of school rules often means that they are either inadequately explained to
the children or that the children’s understanding may become overloaded, particularly in
the case of the youngest children. Some socio-conventional rules may have a function
discernible by children, such as rules to prevent running in school for fear of injury to self
or others or rules which govern academic matters, and such evident purpose should
facilitate their internalization. Many school rules, however, have obscure or arbitrary
 Children’s rule discrimination in school 541
function, such as rules structuring interactions between teachers and children, and thus
may strain the child’s comprehension still further.
In view of this difficult interpretative task confronting the young school pupil, this
study set out to examine whether children discriminated between socio-conventional
rules with discernible purpose and those which have unclear or obscure function.
Therefore, in addition to moral rules, the children’s perceptions of two types of socio-
conventional rules was examined. Thus, the children assessed three types of
transgressions: (1) moral rules; (2) socio-conventional rules, where the purpose and
function was more discernible; and (3) socio-conventional rules of a much more arbitrary
nature. As previous research in various aspects of children’s rule discrimination has
demonstrated their increasing capacity to differentiate to a wider range of events and to
balance greater numbers of factors, it was predicted that the children would display an
ability to distinguish between the three types of rules developing with age. Additionally,
the older children’s rule-differentiation skills should reflect their greater experience of the
school context. It was also predicted that the socio-conventional rules with the more
discernible purpose would be perceived as having greater importance. Thus, they would
be rated as more generalizable, more serious etc. than the more arbitrary rules on the
traditional dimensions; however, they would be perceived as having less importance than
the moral rules.
The children’s rule differentiation within the school context was investigated by
focusing upon their assessments of incidents of school rule-breaking. Their reasoning
about the moral and the two social-conventional rule types was examined on the
traditional dimensions, such as generalizability, rule- and authority-independence,
changeability and seriousness, but with the school as the vital context for these
judgments. The children’s readiness to report transgressions was also examined.

Method
Participants
A total of 144 children from years one to six in four primary schools in one London borough took part. The
children were selected randomly by the teachers, who were asked to present a cross-section of children in
respect of academic ability, but equal numbers of each year and sex. The children were grouped into three
age groups for the purpose of analysis: young, comprising 48 children from years 1 and 2 (24 girls, mean
age = 6.8 years, range = 5.67.7 years; 24 boys, mean age = 6.9 years, range 5.9-7.7 years); middle
comprising 48 children from years 3 and 4 (24 girls, mean age = 8.8 years, range = 7.8-9.7 years; 24
boys, mean age = 8.7 years, range 7.8-9.4 years) and old comprising 48 children from years 5 and 6 (24
girls, mean age = 10.8 years, range = 9.8-11.7; 24 boys, mean age = 10.7 years, range 9.8-1 1.6 years).
There were thus 3 (age) X 2 (gender) X 4 (school) independent groups; the 24 subjects in each subgroup
were made up of three children, of the appropriate sex, drawn from each of the relevant years in each of the
four schools.

Procedure
The interview schedule (see Appendix A) was successfully piloted in two schools on different children from
those used in the main study. The children were all interviewed singly, told there were no right or wrong
answers and assured of anonymity. The interviewer also checked specifically that each child understood
what a rule was; all the children were able to respond by furnishing one or more examples of school rules.
They were told that they would be questioned about school rules which were broken. They were then asked
about six hypothetical scenarios, two for each rule type, in which specific school rules were broken by a
542 Eitbne Bucbanan-Barrow and Martyn Barrett
child: (1)moral-two moral rule situations: (a) pushing a child off a climbing frame; and (b) taking a child’s
book; (2) sociokonventional-two socio-conventional rule scenarios with more discernible purpose: (c)
running in the corridor, where such an activity might be adjudged likely to cause an accident; and (d)
leaving school library books scattered on the floor, where again either injury to person or to property might
be seen as possible; and (3) sociokonventional-two socio-conventionalscenarios representing those school
rules which may appear more arbitrary to children, because their purpose is either not known or harder to
surmise in the absence of an explanation: (e) crossing a ‘forbidden’line in the playground; and (f) calling
a teacher by her first name. Thus, there were two rules in each of the three categories, in order to ensure that
differences between the categories were not solely due to specific features of the particular scenarios
presented for each category.
In all six cases, the scenario described the school rule and the transgression very briefly with little
indication as to the purpose of the action. The two examples of each rule type were differentiated by having
a boy v. a girl as the principal protagonist. The scenarios were presented in a randomized order. Once the
scenario had been carefully read to the child, questions 1-6 followed in a differently randomized order for
each scenario, testing the usual dimensions: (1)parental authority over action; (2) teacher’s authority over
action; (3) head teacher’s authority over action; (4) reporting of transgression; (5) seriousness of
transgression; and (6) changeability of rule. However, questions 7 and 8, investigating rule dependence
and generalizability, were always presented at the end, to avoid confusing the child by introducing such
complications early in the schedule. The method of scoring the children’s responses is given at the end of
Appendix A; this method involves scoring responses as binary 0s and 1s.

Results
The primary data were binary in form and ANOVA was used to analyse these data.
Although ANOVA is not normally used to analyse such data, it is well established that
ANOVA produces accurate results when used to analyse binary data that have been scored
as 0s and 1s (Cochran, 1950; Cox, 1970). The children’s categorical responses to the
criteria judgements on the rule scenarios were initially analysed by a mixed 3 (rule type)
X 2 (gender of protagonist in story) X 3 (age) X 4 (school) X 2 (sex of participant)
ANOVA, with repeated measures on the first two factors and independent groups on the
other factors. However, there were no main effects for either school or sex of participant,
and only a few interactions involving these factors. Therefore, these factors were deleted
from the ANOVA, and it was reanalysed as just a 3 (rule type) X 2 (sex of protagonist in
story) X 3 (age) ANOVA, with repeated measures on the first two factors.

Analysis of rule-type pairs

The mean scores which the children obtained on each pair of rule scenarios (moral, socio-
conventional ( 1 ) (less arbitrary) and socio-conventional(2) (more arbitrary) are shown in
Table 1 ,broken down by age. The ANOVAS revealed a significant main effect of rule type
on all the eight criteria. Planned comparisons (using the planned contrast subcommand
under MANOVA in SPSS,p < .05) were used to establish where the main differences fell
and, on almost all the criteria, significant differences were perceived between the three
types: on seriousness [F(2,142) = 40.50, p < .005]; on generalizability [F(2,141) =
92.01,p < .005]; on rule-independence [F(2,141) = 86.46,p c .0051; on changeability
[F(2,140) = 21.82,p < .005]; on likelihood ofreport [F(2,140) = 28.50,p < .005]; and
on authority independence [head teacher: F(2,142) = 122.17, p < .005; teacher:
F(2,142) = l06.06,p < .005]. Only in the case of the authority ofparents was there any
variation in the pattern of the children’s assessments; while the more arbitrary socio-
conventional transgression type was duly rated as significantly more likely to be within
 Children’s rule discrimination in school 543
the ambit of parents’ authority, there was no significant difference between the ratings for
the moral or the less arbitrary socio-conventional type [F(2,142) = 11.61, p < .005],
which were therefore assessed as the same.

Table 1. Children’s mean responses to criterion judgements by age (standard deviations

in parentheses)
Age 1 Age 2 Age 3 Total
Parent’s Moral“ .oo (.OO) .02 (.lo) .03 (.12) .02 (.lo)
authority SocioIcon ( 1)’ .02 (.lo) .03 (.16) .02 (.lo) .02 (.12)
Sociolcon (2)’ .17 (.28) .07 (.21) .08 (.21) .ll (.24)
Teacher’s Moral .17 (.32) .04 (.14) .09 (.20) .10 (.23)
authority Sociolcon ( 1 ) .32 (.38) .48 (.42) .33 (.38) .38 (.40)
Sociolcon ( 2 ) .51 (.39) .53 (.33) .57 ( 2 9 ) 3 4 (.34)
Head’s Moral .19 (.35) .17 (.32) .06 (.17) .14 (.29)
authority Sociolcon ( 1 ) .44 (.43) S O (.41) .40 (.42) .44 (.42)
Sociolcon (2) .64 ( 3 9 ) .66 ( 3 4 ) .70 (.30) .66 (.34)
Report Moral .97 (.16) .97 (.12) .93 (.21) .95 (.16)
Sociolcon ( 1 ) .95 (.16) .88 (.28) .68 (.41) .83 (.32)
Sociolcon ( 2 ) .86 (.25) .78 (.31) .59 (.38) .75 (.34)
Seriousness Moral 1.00 (.69) $2 (.21) .99 (.07) .96 (.42)
Sociolcon ( 1 ) .80 (.30) .74 (.34) .64 (.42) .73 (.36)
Sociolcon ( 2 ) .69 (.37) .55 (.38) .46 (.35) .57 (.38)
Change rule Moral .39 (.40) .39 (.44) .13 (.30) .30 (.40)
Sociolcon ( 1 ) .43 (.43) .46 (.44) .23 (.37) .37 (.42)
Sociolcon ( 2 ) .53 (.43) .64 ( 3 8 ) .46 (.40) .54 (.41)
No school Moral .19 (.37) .15 (.29) .03 (.12) .12 (.29)
rule Sociolcon ( 1 ) .33 (.40) .20 (.34) .21 (.34) .25 (.36)
Sociolcon ( 2 ) .54 (.45) .55 (.43) .62 (.38) .57 (.42)
Elsewhere Moral .07 (.21) .02 (.lo) .02 (.lo) .04 (.15)
Sociolcon ( 1 ) .29 (.38) .27 (.34) .18 (.24) .25 (.33)
Sociolcon ( 2 ) .51 (.38) .42 (.36) .49 (.38) .47 (.37)

‘Moral = composite means for A and B.

Sociokon (1) = composite means for C and D.
‘Sociokon (2) = composite means for E and F.

Analysing their responses for pairs of rule types, using the planned comparisons, the
children ranged the transgressions along a spectrum of gravity, placing the moral and the
more arbitrary socio-conventional rules at the most serious and least serious ends
respectively, but, more crucially, they also distinguished the less arbitrary socio-
544 Eithne Buchanan-Barrow and Martyn Barrett
conventional as significantly different from both the other rule types, and placed this pair
of transgressions between them. Their composite mean ratings for the moral transgres-
sions (a) and (b) showed that they perceived them as significantly more serious, more
generalizable, less rule-dependent, less authority-dependent (teacher and head-teacher),
less changeable and more likely to be reported than the socio-conventional transgressions.
Similarly, there were also significant differences between the mean scores for the pair of
less arbitrary socio-conventional transgressions (c) and (d) and the pair of more arbitrary
socio-conventional transgressions (e) and (f) for seven out of the eight criteria; only on the
likelihood of reporting the transgression did the children fail to perceive a significant
difference between the two types of socio-conventional rules. Thus, the children duly
differentiated between the three types of rules, and in the predicted order for almost all of
the criteria.
However, unexpectedly, there was also a significant main effect of the sex of the
protagonist on five out of the eight criteria, again on composite mean ratings: on authority
independence {head teacher: F(1,140) = 11.35,p < .005; teacher: F(1,141) = 53.71,p
< .005); on changeability {F(1,139) = 4.67, p < .05); on the likelihood of report
{F(1,138) = 32.30, p < .005); and on seriousness {F(1,141) = 5.40, p < .05). There
were also some interactions with rule type: on authority independence, both head teacher
and teacher; on the likelihood of report; on generalitability; and on seriousness. As
seriousness is arguably the most essential criterion for assessing the transgressions,
planned comparisons (again using the planned comparison subcommand under MAN-
OVA in SPSS,p < .05) were used to probe this effect. They revealed that the differences
were in two of the paired rule types, the moral transgressions and the more arbitrary socio-
conventional; in both cases, the transgression which was attributed to a boy offender was
viewed as significantly more serious than the action attributed to the girl (thus, a was seen
as more serious than b, and e as more serious than f). These distinctions were also still
apparent in a comparison of the ratings made by boy and girl participants separately. Thus
while the children ordered the three pairs of transgressions on the dimension of
seriousness in the predicted fashion, they all also perceived some significant differences
between the pairs of rules. Post-hoc testing of a different group of 72 5 to 11-year-old
children on the same six scenarios, but with the sex of the protagonists reversed, found no
significant difference between the seriousness as a function of protagonist sex {F(1/214) =
0.42, non significant) suggesting that the differences stemmed from the actions them-
selves rather than from the sex of the protagonists.
There were three main effects of age (seriousness, changeability and likelihood of
report); some interactions with age (with rule type, on authority independence (teacher),
likelihood of report, and on generalizability); and some three-way interactions (sex of
protagonist and rule type, on authority independence for both head teacher and teacher).
Therefore, given this additional evidence of the variations perceived by the children
between the rule types and in order to explore these interactions in more detail, the
children’s responses to each of the six transgressions were also examined separately, using
log-linear analyses to examine these differences more closely. As in the ANOVAS, there
were only very few significant associations with either school or sex of participant out of
the 72 individual variables, probably within that expected by chance. It would appear
therefore that neither of these factors are of major importance in children’s school rule
understanding and therefore these associations are not reported.
 Children's rule discrimination in school 545
Analysis of separate transgressions
With both the moral rule scenarios, whereas there were no significant variations with age
in the assessments of either seriousness or generalizability, there were differences on other
criteria, which were examined by post-hoc X'or Fisher's exact probability tests, as
appropriate. On rule-independence, the young group on (a), pushing child off the
climbing frame, and the young and middle group on (b), taking the child's book, were
significantly more likely to say that the actions would be acceptable in the absence of a
rule [(a) young vs. old: ~ ' ( 1 =) 6 . 6 , ~c .05); (b) young vs. old: ~ ' ( 1 =) 2.8,p < .05;
middle vs. old: ~ ' ( 1 =) 4.7, p < .05}. On changeability, the young and middle groups
in their judgements of both (a) and (b) showed a significant tendency over the old group
to say that the rule could be changed [(a) young vs. old: x'( 1) = 4.1, p c .05; middle vs.
old: ~ ' ( 1 =) 5 . 7 , ~c .05: (b) young vs. old: ~ ' ( 1 =) 11.3,p < ,0005; middle vs. old:
~'(1= ) 8.9, p < .005]. Finally, with the head teacher's authority in scenario (a), the
young and middle groups were significantly more inclined to reply that the transgression
would be acceptable if the head teacher gave permission (young vs. old: Fisher's z = 2.1,
p < .05; middle vs. old: Fisher's z = 1.6,p < .05). The children's responses to the four
socio-conventional rules displayed some age differences on many of the judgments, with
the responses of the old group generally varying from the two other age groups. On
seriousness, on (c), running in the corridor, the old group were significantly less likely to
rate the action as 'very bad' [old vs. young: ~ ' ( 1 =) 9.3,p c .005; old VS. middle: ~ ' ( 1 )
= 6 . 1 , ~< .05] or as 'quite bad' [old vs. young: ~ ' ( 1 =) 3.5,p c .057; on (e), crossing
the playground line, the old group were also less likely to say that the action was 'very bad'
[old vs. young: ~ ' ( 1 =) 6 . 7 , ~c .005; old vs. middle: ~ ' ( 1 =) 3.3,p c .05l and more
likely to rate it as 'a little badhot bad' [old vs. young: ~ ' ( 1 =) 4.8, p c .05];and on (0,
addressing the teacher, the old group were much less likely to rate it as 'very bad' [old VS.
young: ~ ' ( 1 =) 8.9,p c .005; old vs. middle: ~ ' ( 1 =) 3.7,p < .05]. With respect to
changeability, on (d), leaving books on the floor, the old group were significantly less
likely to say that the rule could be changed [old vs. young: ~ ' ( 1 =) 7.7,p c .05; old VS.
middle: ~ ' ( 1 )= 5.3, p c .05]. On the likelihood of report, there was a significantly
greater reluctance in the old group to tell a teacher: on (d), leaving library books on the
floor [old vs. young: ~ ' ( 1 =) 13.7,p c .005; old vs. middle: ~ ' ( 1 =) 6.1,p < .051; on
(e), crossing the playground line [old vs. young: ~ ' ( 1 )= 5.9, p < .05; old vs. middle:
~ ' ( 1 =) 3.0,p < ,051; and (0,addressing the teacher [old vs. young: ~ ' ( 1 =) 12.4,p c
.0005; old vs. middle: ~ ' ( 1 =) 5.0,p c .05]. With generalizability, the old group were
significantly less likely to see the action in (d), leaving library books on the floor, as
acceptable outside school [old vs. young: ~ ' ( 1 =) 8.9,p c .05; old vs. middle: Fisher's z
= 2 . 1 , ~c .05}. There were also some age differences in authority dependence: on (e),
crossing the playground line, the old group were significantly more likely to believe that
authority could be given for the action: by head teacher [old vs. young: ~ ' ( 1 =) 7 . 6 , ~c
.005; old vs. middle: ~ ' ( 1 =) 2.7,p < .05] and by teacher [old vs. young: ~ ' ( 1 =) 5.7,
p < .05];on (d), leaving books on the floor, in a more complex finding, the middle group
were significantly more likely than both the young and old groups to state that permission
could be granted for this action by a teacher [middle vs. old: ~ ' ( 1 =) 5.3,p c .05; middle
vs. young: ~ ' ( 1 =) 5.3,p c .05}.
546 Eithne Buchanan-Bawow and Martyn Bawett
Discussion
Children’s general rule-understanding
As predicted, the children’s responses to the rule-infringement scenarios were indicative
of an overall grasp of the variations between the three types of rules described, but there
were fewer differences with age than expected. Most of the children were apparently not
only alert to the traditional distinction between the moral and socio-conventional
domains of rule understanding, but they also seemed to be able to perceive differences
between the socio-conventional violations, which might have been due to the variations
in function. The breaches of the rules against running in the corridor and leaving books
on the library floor were viewed differently and generally more seriously by the children
than the other conventional rules (not crossing the playground line and addressing a
teacher correctly) which may have been seen as more arbitrary and less momentous in
effect. This is support for the view that, while the socio-conventional domain is more
varied, complex and extensive than the moral one, even young children are alert to the
differences in socio-conventional rules, actively perceiving and responding according to
variations in situations. Overall, therefore, the children displayed the competent social
skills of rule discrimination of earlier findings but, in addition, they were able to
differentiate between the two types of socio-conventional rules.
Furthermore, the pattern of the children’s differentiation was consistent; the children
ranged the three rule-types on all the criteria in the same order. The children’s skills may
reflect the complexity of the social environment, even for the young; social rule
comprehension is too multi-faceted to be subscribed by only two rule-domains or
categories; this interpretation was given added emphasis by the children’s additional
discrimination within the paired rule type examples. Other studies have found that
children may indeed be actively perceiving and assessing variations in scenarios beyond
simple rule categorization. For example, Tisak and Turiel(1988) expressed the view that
moral violations could be subdivided according to their consequences, with physical
injury assessed as a more major consequence than stealing. Even relatively simple
scenarios, such as those presented in this study, may prompt both conscious and
unconscious differentiations on other criteria, beyond those examined here. In short,
children’s rule-discrimination ability may be more complex and accomplished than
previously thought.
There were also indications, again with relatively few age differences, that children
have a good grasp of the school-as-context, because they recognized which adults have
authority over rules within the school. They generally acknowledged that the head
teacher, and to a slightly lesser extent the teachers, could have jurisdiction over all the
socio-conventional rules, while denying such authority to parents. This would indicate
that children have a clear understanding that power in school rests with those adults in
authority within the school, and that generally there are no external authority sources,
even in the case of such an important and concerned group of adults as the parents.
However, almost all the children denied all adults, teachers and parents any jurisdiction
over moral rules.
The children’s responses to only three of the criteria showed some differences with age.
On the assessments of seriousness, while there was no difference for the moral rules, the
oldest children tended to rate the socio-conventional infringements as significantly less
 Children’s rule discrimination in school 547
grave than the youngest children. However, these lowered ratings of seriousness did not
affect the oldest children’s judgments on other criteria, such as generalizability, rule- or
authority-independence, where there were no age variations. This age variation may have
stemmed from the oldest children’s greater experience of school; they had probably
viewed many such transgressions during their time in school and this familiarity had
rendered the events less remarkable, though the rule and authority aspects were judged
the same. The age differences in the likelihood of reporting and in changeability were
more complex and are dealt with in discussion of the children’s social-conventional
reasoning.

Children’s moral rule-understanding

In the case of the moral transgressions, the overall pattern of responses indicated that
children duly acknowledge the crucial importance of moral rules. Their judgments on all
the criteria generally supported the view that an understanding of moral rules is
established prior to school entry (Smetana & Braeges, 1990; Smetana, Schlagman &
Adams, 1993). Nevertheless, the few age differences which were apparent indicated some
important variations in the moral understanding of some of the younger children.
Whereas there was substantial agreement across the age groups about the considerable
seriousness of the moral breaches and with almost all of the children expressing the view
that such actions would be just as wrong in other places, nevertheless, despite the strength
of these judgments, some of the younger children were prepared to temper their views on
two of the other dimensions, namely authority-independence (the head teacher) and rule-
independence. With regard to pushing a child off the climbing frame (a), whereas the
large majority of all children (90 per cent) said that the head teacher could not authorize
this action, the young and middle group children were significantly more likely to say
that this violation would then be acceptable. In addition, some of the children’s moral
understanding displayed further ambivalence when they reassessed the violations in the
absence of rules. With both moral infringements some of the young and middle groups
were significantly more prepared than the old group to say that the contraventions would
be acceptable if there was no rule. These findings suggest an interesting contradiction.
While all the children appeared to be cognisant of the distinctive gravity and general-
izability attached to moral transgressions, which understanding they would have acquired
well before their entry into school (Smetana, Schlagman & Adams, 1993), some of the
younger children were apparently nevertheless influenced by other considerations to
judge these violations in somewhat contradictory fashion on the other criteria of authority
dependence and rule dependence.
With regard to rule dependence, the responses of some of the younger children may
reflect their tendency to see school rules generally as defining what is good and bad
behaviour (Buchanan-Barrow & Barrett, 1996,1998), rather than having higher purpose,
such as the avoidance of injury or hurt and that therefore an action in the absence of a rule
might appear to become acceptable. While moral understanding is held to be rule
independent, it would appear that some younger children’s thinking may become
confused when confronted with the extensive school rule system. According to Corsaro
(1990) rule understanding within a context develops in a two-stage process; while
children are still struggling with the first stage, that of simply internalizing the extensive
548 Eithne Buchanan-Barrow and Martyn Barrett
rule system of school, it may be that the presence or absence of rules would appear to be
the most salient feature when assessing them.
The variations in responses to the role of authority may also reflect the younger
children’s more immature understanding of school rules, but is probably more likely to be
a consequence of their preoccupation with the power of the head in particular, and the
teachers to a lesser extent (Buchanan-Barrow & Barrett, 1996, in press). Authority is a
central component of the school system and these younger children, still adapting to the
complex life of the school, may be over-conscious of the power of the head and teachers.
It may be that, given the apparent conflict between those in authority and the nature of
the action, the children simply make the wrong choice, either because the question of
authority is more salient or because the children hold an immature belief that they should
always comply with a teacher’s order without question. There were also some children, in
all three age groups, who believed that the moral rules could be changed, but these
responses must be viewed with some caution, particularly in the older children, as they
appear to counter their otherwise competent assessments of the moral transgressions. The
question may have been ambiguous, and as the children were not asked to justify their
responses, it is not possible to explain these findings.
Overall, these results indicate that children’s moral rule understanding may require
some re-adjustment to the new context of the school. More generally, they are also
consistent with the view that the context in which children are developing their thinking,
even their moral reasoning, may be a factor in that comprehension (Brownell, 1989).This
is despite the belief that such knowledge should be context free or the view that children’s
moral rule knowledge is well established prior to school entry and should therefore
transfer to a new context (Siegal & Storey, 1985). Given the centrality not only of rules,
but also of the closely-linked concepts of authority and power in the children’s under-
standing of the school system, the apparent contradictions in the responses of some of the
younger children to the moral scenarios in this study can be seen as indicative of their
attempts to balance the competing claims of moral rule-understanding with their
perceptions of the school structure. However, as the responses of the older children would
suggest, children eventually re-establish their rule-discrimination abilities, and success-
fully adapt them to the new situation. The study by Laupa and Turiel(1986) supports the
view that this is a short-lived inconsistency in the youngest children’s rule understanding;
in hypothetical situations in which adults in authority ordered children to harm others,
they found that almost all the children (aged 7-1 1 years) refused to comply.

Children’s socio-conventionalunderstanding
With the socio-conventional transgressions, there were some predictable patterns: on
seriousness the ratings decreased both along the rule-spectrum and with age; on the
likelihood of report the older children were considerably less likely to tell a teacher,
probably more as a consequence of growing peer solidarity than of the decreasing
importance of the rule. But if children are indeed conscious of the lesser gravity of socio-
conventional transgressions, there were also indications of a developing awareness of the
community value of rules, which may have produced some more contradictory trends. For
example, the old group of children were almost unanimous in saying that leaving library
books on the floor (d) would not be acceptable elsewhere (generalizability), while the
 Children’s rule discrimination in school 549
other groups were significantly more likely to approve it, and the old group were also the
least likely to say that the socio-conventional rules might be changed, although the only
significant difference was again with (d). The complex patterns in the assessments of
authority also indicated a growing sense of the need for order in school. While there was
increasing understanding that heads and teachers could authorise these actions, within
the general trend the age groups reported different positions: for example, with (d), the
middle group were the most likely to say this action could be authorised, with perhaps the
older children again influenced by the possible serious consequences. Overall, there were
indications of the varying emphasis and developing perceptions that the children brought
to their judgements of the rules which support the view that the socio-conventional
domain is a very complex one.

Conclusion
Finally, there are three main conclusions to be drawn. First, there are indications that
children are able to differentiate at least three rule types, not only making the core
distinction between the moral and socio-conventional domains, but also perceiving clear
differences between the two types of socio-conventional rules postulated in this study.
The transgressions relating to the socio-conventional rules which had less evident
function were perceived as different from the other socio-conventional ones. This ability
to distinguish and effectively range the rule types along a continuum suggests that
children possess a considerable capacity for dealing with the complexity of the system of
school rules and furthermore, this hardly varied with age. Even the youngest children,
apart from a few minor instances, ordered the rules as the oldest children did; whatever
criteria or dimensions were most salient or whatever differences were perceived by each
age-group, the end result was the same and the children’s judgments varied with the
nature or type of rule and not with age. Secondly, almost all the children were clear who
had authority over rules in school, namely the head teacher and the teachers, and were
almost unanimous in rejecting a role for parents. In support of previous research (Laupa,
1991, 1995) children are conscious of the system of the school from an early age and
understand that authority within it is wielded by those holding social organizational roles
in school. Thirdly, there is some evidence here that children’s moral rule understanding
may be affected by their entry into a new social environment, with the youngest children’s
moral reasoning displaying some inconsistencies. This suggests that children’s social
environment is an important influence upon their moral cognition.

Acknowledgements
This research was supported by a Research Studentship from the Economic and Social Research
Council.

References
Blumenfeld, P.C., Pintrich, P.R. & Hamilton, V.L. (1987). Teacher talk and students’ reasoning about
morals, conventions and achievement. Child Development, 58, 1389-1401.
Brownell, C.A. (1989). Socially-shared cognition: the role of social context in the construction of
550 Eithne Budanan-Baww and Martyn Bawett
knowledge. In L.T. Winegar (Ed.), Social Interaction and the Development of Children’s Unrlerstanding.
Norwood, NJ: Ablex.
Buchanan-Barrow, E. & Barrett, M. (1996). Primary school children’s understanding of the school. British
Journal of Educational Psychology, 66, 33-46
Buchanan-Barrett, E. and Barrett, M. (1998). Individual differences in children’s understanding of the
school. Social Development, 7, 250-268.
Cochran, W.G. (1950). The comparison of percentages in matched samples. Biometrika, 37,256-266.
Corsaro, W.A. (1990). The underlife of the nursery school: young children’s social representations of adult
rules. In G. Duveen & B. Lloyd (Eds), Social Representations and the Development ofKnowledge. Cambridge:
Cambridge University Press.
Cox, D.R. (1970) Analysis ofBinary Data. London: Methuen.
Davidson, P., Turiel, E. and Black, A. (1983). The effect of stimulus familiarity on the use of criteria and
justihcations in children’s social reasoning. BritishJournal of Developmental Psychology, 1 , 47-65.
Dunn, J. and Munn, P. (1985). Becoming a family member: family conflict and the development of social
understanding in the second year. Child Development, 56,480-492.
Emler, N., Ohana, J. & Moscovici, S. (1987). Children’s beliefs about institutional roles: a cross-national
study of representations of the teacher’s role. British Journal ofEducation Psychology, 57, 26-37.
Laupa, M. (1991). Children’s reasoning about three authority attributes: Adult status, knowledge and
social position. Developmental Psychology, 27, 321-329.
Laupa, M. (1995). Children’s reasoning about authority in home and school contexts. Social Development, 4,
5-16.
Laupa, M. & Turiel, E. (1986). Children’s conceptions of adult and peer authority. Child Development, 57,
404-412.
Laupa, M. and Turiel, E. (1993). Authority reasoning and social context.Journal ofEducational Psychology,
85,191-197.
Siegal, M. & Storey, R.M. (1985). Day care and children’s conceptions of moral and social rules. Child
Development, 56,1001-1008.
Smetana, J. (1993). Understanding of social rules. In M.Bennett (Ed.), The Child as Psychologist. Hemel
Hempstead: Harvester Wheatsheaf.
Smetana, J.G. & Braeges, J.L. (1990). The development of toddlers’ moral and conventional judgements.
Mewill-Palmer Quarterly, 36, 329-346.
Smetana, J.G., Schlagman, N. & Adams, P. (1993). Preschoolers’ judgements about hypothetical and
actual transgressions. Child Development, 64, 202-214.
Tisak, M. & Turiel, E. (1984). Children’s conceptions of moral and prudential rules. Child Development. 55,
1030-1039.
Tisak, M.S. & Turiel, E. (1988). Variation in seriousness of transgressions and children’s moral and
conventional concepts. Deve[opmental Psychology, 24, 352-357.
Turiel, E. (1983). The Development ofsocial Knowledge: Morality and Convention. New York, Ny: Cambridge
University Press.
Weston, D.R. & Turiel, E. (1980). Act-rules relations: children’s concepts of social rules. Development
Psychology, 16,417-424.

Received 18 November 1996; revised version received 23 January 1998.

Appendix: Children’s interview schedule

(A) Moral
In David’s school, there is a rule that you must never push anyone off the climbing frame. One day at
school, when David is playing on the climbing frame, he pushes Stephen so hard that he falls to the
ground.

(1)Would it be all right to push someone off the school climbing frame if his mother or father told David
he could do that? YES..........NO..........
 Children’s rule discrimination in school 551
(2) Would it be OK for David to push someone off the school climbing frame ifa teacher told him he could
do that? YES..........NO ..........

( 3 ) Supposing the Head told David he could push someone off the school climbing frame, would it be all
right then? YES..........N O ..........

(4)Supposing the teacher didn’t see what happened but another child did, do you think that child should
tell a teacher? YES..........N O ..........

( 5 ) Do you think what David did was bad or not bad? BAD ..........NOT BAD ..........
If BAD, then - VERY BAD ..........QUITE BAD ..........JUST A LITTLE BAD..........

(6) Do you think chis rule could be changed? YES..........N O ..........

(7) Supposing there wasn’t a school rule about pushing people off the climbing frame, then do you think
David would be right or wrong? RIGHT ..........W R O N G ..........

(8)Supposing David pushed someone offanother climbing frame, not at school this time, do you think that
would be right or wrong? RIGHT ..........W R O N G ..........

( B ) Moral
In Lucy’s school there is a rule that you must not take other children’s things. One day, Lucy takes
Andrew’s book when he’s not looking and puts it in her bag.

( C ) Sociolconventional ( I ) (less arbitrary)

In William’s school there is a rule that you must never run anywhere inside school. One day, when morning
lessons are over, William runs very fast along the corridor because he wants to be first in the
playground.

(0)Sociolconventional(1) (less arbitrary)

In Rachel’s school there is a rule that you must always put books back tidily on the library shelves when
you have finished them. One day, after Rachel has been looking at several books, she simply leaves them
all on the library floor when she goes off to lunch.

( E ) Sociolconventional (2) (more arbitrary)

In Richard’s school there is a rule that you must never cross the double yellow line at one end of the
playground. One day during morning break, Richard kicks his ball over the yellow line. H e then crosses
the line to pick up his ball.

( F ) Sociolconventional(2) (more arbitrary)

In Sarah’s school there is a rule that you must always call the teacher by their full name. Sarah’s class teacher
is Mrs Brown and the children call her by that name when they need her help. Sarah hears another teacher
call Mrs Brown by her first name, Helen. She thinks it would be fun to use the teacher’s first name and in
the next lesson she calls out ‘Helen’ when she needs some help.
Analogous questions were asked about each scenario to those that were asked about scenario (A).
Questions 1, 2, 3 , 4 , and 6 were coded as follows: yes = 1; no = 0. Questions 7 and 8 were coded as
follows: right = 1; wrong = 0. Question 5 was coded as follows: very badlquite bad = 1; just a little bad/
not bad = 0.