Moral judgment in high-functioning pervasive developmental disorders


Toshinobu Takeda, MD, PhD, Department of Neuropsychiatry, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan. Email:


Abstract  Individuals with pervasive developmental disorder (PDD) are characterized by deficits in socialization. To date, moral judgment, which may have a considerable influence on socialization, has not been fully investigated in high-functioning PDD (HFPDD), particularly from a viewpoint of practical adjustment with peers. Human External Action and its internal Reasoning Type (HEART), a standardized test for evaluating moral judgment in school children developed in Japan, was used to compare various aspects of moral judgment between 23 students with HFPDD (6–14 years old) and 23 students with typical development matched for age, intelligence, and socioeconomic status. Students with HFPDD scored significantly lower on Internal Moral Reasoning than control students. As for the level of Internal Moral Reasoning, while both groups reached a conventional (third) level in almost all items, fewer students with HFPDD achieved an autonomous (fourth) level and more students with HFPDD remained at a heteronomous (second) level than did control students. In the HFPDD group there were significant positive correlations between some items of Internal Moral Reasoning and verbal ability-related items of Wechsler Intelligence Scale for Children-III. A comparatively lower score in students with HFPDD may relate to difficulty in socialization.


Increasing attention has been paid to high-functioning pervasive developmental disorders (HFPDD) in the last 20 years. Among pervasive developmental disorders (PDD), HFPDD is defined as IQ ≥70 or IQ ≥ 85 by a narrower definition. Chakrabarti and Fombonne reported a 0.626% prevalence of PDD,1 74.2% of which are HFPDD in a broad sense.2

While intellectual ability in HFPDD is within or near normal levels, scholastic or social maladjustments in HFPDD are common.3 Serious deficits in socialization have often been suggested as the main reasons for these maladjustments.4

One of the factors that has an influence on socialization is moral judgment. Moral judgment consists of three facets. Moral judgment has an emotional component, because powerful feelings cause us to empathize with another's distress or feel guilty when we are the cause of distress. Moral judgment also has an important cognitive component. Children's developing social understanding allows them to make more profound judgments about actions they believe to be right or wrong. Finally, moral judgment has a vital behavioral component because experiencing morally relevant thoughts and feelings only increases the likelihood, but does not guarantee, that people will act in accord with them.5

Piaget identified two broad stages of moral understanding.6 Five- to 10-year-old children are in the first stage, called the heteronomous stage. In this stage children view rules as handed down by authority (God, parents, teachers) as having a permanent existence, unchangeable, and requiring strict obedience. The second stage is called the autonomous stage. Childrenaged from approximately 10 years are in this stage. They no longer view rules as fixed but see them as flexible, socially agreed-on principles that can be revised to suit the will of the majority.

Following on from Piaget's theory, Kohlberg developed a six-stage sequence of moral understanding.7 Kohlberg organized his six stages into three general levels and made stronger claims than Piaget about a fixed order of moral change. At the first level, the preconventional level, moral judgment is externally controlled. As in Piaget's heteronomous stage, children accept the rules of authority figures, and actions are judged by their consequences. Behaviors that result in punishment are viewed as bad, and those that lead to rewards are seen as good. At the second level, the conventional level, individuals continue to regard conformation to social rules as important, but not for reasons of self-interest. Rather, they believe that actively maintaining the current social system ensures positive human relationships and societal order. At the final level, the post-conventional level, individuals move beyond unquestioning support for the rules and laws of their own society. They define moral judgment in terms of abstract principles and values that apply to all situations and societies.

Of the factors that can influence development in moral judgment, several have been noted as being impaired in PDD. Empathy is essential for recognizing and understanding the distress of others, but some studies have found that individuals with PDD have difficulties in empathizing.8 Similarly, some studies suggested that theory of mind (ToM), which is necessary for speculating upon another's mind state, is impaired in PDD.9 Furthermore, social referencing, which is involved in checking whether one's own behaviors are acceptable or not by examining the behavior of others, has also been found to be lacking in PDD.10

To date there has been no detailed investigation about moral judgment in narrowly defined HFPDD. Blair investigated the responsiveness to the distress of others of 20 children with autism (age range, 8–17 years; mean verbal IQ [VIQ], approx. 75) through their responses to the moral/conventional distinction measure.11 According to Smetana, the judgment between moral transgression (e.g. hitting another, damaging another's property) and conventional transgression (e.g. talking in class, dressing in opposite-sex clothes) is made from the age of 39 months and the presence of victims distinguishes moral and conventional transgressions.12 The author found that the children with autism successfully made the aforementioned distinction between moral and conventional transgression in their judgments.12

Grant et al. studied the ability of 19 children with autism (mean age, 12 years; mean VIQ, 74) to make moral judgments.13 In their study subjects were presented with pairs of vignettes in which actions were either deliberate or accidental and caused either injury to a person or damage to property, and were then asked to judge which protagonist was the naughtier. Results showed that the children with autism were as likely as controls to judge culpability on the basis of motive, and to judge injury to persons as more culpable than damage to property.

These two studies are important in that they indicated that children with PDD did not necessarily have greater difficulty making moral judgments than children with typical development; however, when social adjustment in HFPDD is considered, the clinically important point is to compare the level of moral function of children with HFPDD to that of their peers.

There are two main advantages in conducting studies of moral judgment in Japan. First, in response to educational need there are moral judgment tests available that have been standardized with large populations. Second, because Japan is generally less heterogeneous than many other countries, factors such as nationality and religion,5,14,15 which can have a profound influence on moral judgment, are relatively even in Japan.

In the present study a comparison of performance on a detailed test of moral judgment was conducted between school children with narrowly defined HFPDD (i.e. full-scale IQ [FSIQ]≥85), and children with typical development matched for age, sex, FSIQ, VIQ, performance IQ (PIQ) and socioeconomic state (SES). The aim of the study was to investigate differences in moral judgment between the two groups and to explore factors related to moral judgment, through which clues useful to diagnosis or therapeutic intervention can be found.


This study was approved by the Medical Ethics Committee at the University of Tokyo, and parents gave written informed consent prior to participation.


The 23 subjects with HFPDD were recruited from school students attending either the outpatient clinic of the University of Tokyo Hospital or two clinics in and near the Tokyo region specializing in developmental disorders. In each facility, resident clinical teams of experienced professionals participated in the evaluation of the children and then made diagnoses of PDD and its subcategories by consensus according to DSM-IV16 and ICD-10.17

All of the students with HFPDD met the following four conditions: (i) diagnosis of PDD by experienced psychiatrists; (ii) FSIQ, VIQ, and PIQ all ≥85; (iii) absence of comorbid DSM-IV axis-I disorders; and (iv) regular attendance of school classes for children with typical development. The second condition was imposed in the process of matching intellectual ability between students with HFPDD and students with typical development. As for the fourth conditions, students with HFPDD attending classes other than those for children with typical development are less likely to have the morals class usually given once a week in regular schools.

The following DSM-IV-defined PDD subtypes were present in the sample: autistic disorder (AD; n = 5; 3 male, 2 female), PDD not otherwise specified (PDDNOS; n = 10; 10 male), and Asperger's disorder (AS; n = 8; 7 male, 1 female). The mean score of the Childhood Autism Rating Scale-Tokyo Version18 was 28.6 ± 2.3 (range, 19.5–32.5).

Twenty-three control subjects were drawn from classes for children with typical development in Tokyo and its vicinity to match the same number of children with HFPDD for age, gender, FSIQ, VIQ, and PIQ of Wechsler Intelligence Scale for Children-III (Japanese version) and parental SES under a group-matching procedure.19 (Table 1).

Table 1.  Participant characteristics
 Control (n = 23)HFPDD (n = 23)P
  1. SES scores range from 1 to 5, one being the highest SES.

  2. HFPDD, high-functioning pervasive developmental disorders; PIQ, performance IQ; SES, socioeconomic status; VIQ, verbal IQ; WISC-III, Wechsler Intelligence Scale for Children-III.

Gender (male/female) 20/3   20/3  1.000
Age (months)117.222.375–156112.625.876–1700.516
IQ (WISC-III)109.411.492–134106.112.188–1370.358

All of the control students passed the ToM task (Sally Ann test), while six students with HFPDD failed it. All of the parents of the students in both groups were Japanese.


Moral judgment test

HEART evaluates decisions reflecting internal moral reasoning by using cartoon-like illustrated questions and evaluates decisions reflecting external moral actions by use of short-sentence questions (Appendix I).20 A separate total score is generated for both Internal Moral Reasoning and External Moral Actions. In addition three subscores are extracted by a factor analysis; Empathy (kindness, sympathy, friendship, impartiality; relationship with others), Self-Help (rationality, originality, ambition, independence, autonomy, justice, responsibility), and Norm (keeping things neat and tidy, courtesy, well-regulated life, observing the rules, public spirit). Because the present study focused on moral judgment, hereafter we concentrate chiefly on Internal Moral Reasoning.

When calculating the score for and evaluating the level of internal moral reasoning skills with HEART, four levels of moral judgment were adopted. These levels may be described as follows.

Level I (amoral–egocentric)

Thinking mainly of satisfying one's own needs or pursuing one's own interest and rarely thinking of others. Those who remain at this level act temperamentally and impulsively. Consequently their decision to behave morally or not depends primarily on their interest.

Level II (heteronomous–preconventional)

Those who are in this level basically act to follow authorities such as parents, teachers or friends lest their behaviors result in disapproval or punishment. Hence, in the presence of their authorities or friends, they act morally, in their absence they may act immorally.

Level III (conventional)

Those who are at this level establish their own internal standard of right–wrong or good–evil to some extent. They continue to regard conformation to social rules as important and try to follow rules while on some occasions they apply rules too rigidly.

Level IV (autonomous–altruistic)

Those who reach this level move beyond unquestioning support for the rules and therefore apply rules flexibly. They can take not only their own happiness but also others' happiness into consideration.

Each illustrated question on HEART has four choices, which correspond to each level of internal moral reasoning. The student's answers are evaluated based on the moral level corresponding to the answer they selected. Based on scores from all questions, a final score and subsequently a total and three subscale z-score are calculated. The total final scores and three subscales scores can range from 1 to 4. Scores between 1.00 and 1.49, 1.50 and 2.49, 2.50 and 3.49, and 3.50 and 4.00 correspond to moral levels I, II, III, and IV, respectively.

HEART consists of 40 short-sentence and 14 illustrated questions. It takes around 30 min. It was standardized in 1991 using data on 23 255 students who attended classes for children with typical development in Japan. As for its internal consistency reliability, Cronbach alphas were ≥0.8 in all the items except Empathy. Its test–retest reliability with a 5-week interval was generally satisfactory. Furthermore, it was confirmed that each student's score of HEART was consistent with the daily observation by his or her homeroom teacher.

Data analysis

The results of the Internal Moral Reasoning were compared between the two groups by t-test. Correlations of the results of the Internal Moral Reasoning with each item of the WISC-III scores were also computed in the HFPDD and control group, respectively. Furthermore, in the case of significant correlation between certain items in either group, comparison of two regression slopes between the two groups was done. When necessary, χ2 test or analysis of covariance (ancova) was applied. All statistical analyses were performed with SPSS 11.0 J for Windows (SPSS Japan, Tokyo, Japan) with the significance level set at P < 0.05 (two-tailed).


Comparison of the results of Internal Moral Reasoning between HFPDD and control groups

For Internal Moral Reasoning the control group recorded a significantly higher score than the HFPDD group for the Total score as well as for the subscores of Empathy, and Norm. ancova with age, IQ, VIQ, PIQ, or SES as covariates did not change the statistical conclusion (Table 2).

Table 2.  Difference in z-score for Internal Moral Reasoning
  Control (n = 23)HFPDD (n = 23)P
  1. HFPDD, high-functioning pervasive developmental disorders.

Internal MoralEmpathy0.2570.834−1.13–1.67−0.4570.939−2.37–1.080.014

Level of Internal Moral Reasoning

While both groups reached level III in almost all items, fewer students with HFPDD achieved level IV for the Total score and more students with HFPDD remained at level II for Empathy than did control students (Table 3). However, in chi-square test there were not significant proportional differences between the two groups in any items. In neither group were there any significant correlations between the level of any item on Internal Moral Reasoning and age.

Table 3.  Morality level
Morality levelTotalEmpathySelf-HelpNorm
n = 23
n = 23
n = 23
n = 23
n = 23
n = 23
n = 23
n = 23
  1. HFPDD, high-functioning pervasive developmental disorders.

I (Amoral–egocentric)00000000
II (Heteronomous–preconventional)11270011
III (Conventional)162120161314913
IV (Autonomous–altruistic)6110109139

Correlation between Internal Moral Reasoning and WISC-III

In the control group there were significant negative correlations between Norms and VIQ. There were significant positive correlations between Empathy and Object Assembly, between Self-Help and Digit Symbol, and between Norm and Digit Symbol.

In the HFPDD group there were significant positive correlations between Self-Help and Verbal Comprehension, between Total score and Information, between Self-Help and Information, between Total score and Comprehension, and between Empathy and Comprehension.

The comparison of regression slopes between the two groups found significant differences in the slopes for correlations between Total score and Information, between Total score and Comprehension, between Empathy and Comprehension, and between Empathy and Object Assembly (Table 4).

Table 4.  Correlations between Internal Moral Reasoning and items in WISC-III
WISC-III variablesInternal moral reasoning
  • *

     P < 0.05,

  • **

     P < 0.001. A/B, correlation coefficient in the control group (A)/correlation coefficient in the HFPDD group (B); Bold, significant difference in the slopes for correlations between two groups.

  • HFPDD, high-functioning pervasive developmental disorders; PIQ, performance IQ; VIQ, verbal IQ; WISC-III, Wechsler Intelligence Scale for Children-III.

Verbal Comprehension0.056/0.3850.285/0.4090.015/0.431*−0.195/0.200
Perceptual Organization0.164/0.0390.399/0.1740.109/−0.010−0.157/−0.038
Freedom from Distractibility−0.041/0.158−0.068/0.229−0.015/0.1950.034/0.080
Processing Speed0.152/0.184−0.104/0.0520.271/0.2190.333/0.193
Digit Span0.320/0.1320.222/0.2110.370/0.1370.265/0.095
Picture Completion0.114/0.0780.306/0.2330.093/0.042−0.068/−0.019
Picture Arrangement0.172/−0.0400.210/−0.0040.129/−0.0300.059//−0.002
Block Design−0.094/−0.0280.070/0.019−0.090/−0.079−0.059/−0.066
Object Assembly0.246/0.2340.443*/0.2150.172/0.299−0.088/0.133
Symbol Search0.311/0.020−0.072/−0.0690.431*/0.1020.501*/0.023


This study is the first attempt at a detailed investigation of moral judgment in students with HFPDD compared with closely FSIQ-, VIQ-, PIQ-, and SES-matched students with typical development. In Internal Moral Reasoning the HFPDD group scored significantly lower than the control group for Total score, Empathy, and Norm. Significant differences between the two groups in each of three items remained after ancova with age as covariate and with IQ, VIQ, PIQ, and SES controlled for. As for the level of Internal Moral Reasoning, independent of age, both groups reached level III on almost all of the items. In the HFPDD group there were significant positive correlations between some items of Internal Moral Reasoning and Verbal Comprehension, Information, and Comprehension. Furthermore, the slopes for these correlations significantly differed between the two groups and there was also significant difference in correlations between these combinations.

According to the model of the relationship between moral judgment and moral action proposed by Kohlberg and Candee, between the process from moral judgment to moral action, there are non-moral skills such as IQ, attention, and delay of gratification that are influential on the final action.21 From the viewpoint of this model, the HFPDD group is in a more disadvantaged position with respect to moral action because children with HFPDD were frequently found to have deficits of attention or self-control. However, in contrast it is often noted that once children with PDD have learned behavioral rules, they apply them firmly and rigidly. At any rate, results in the present study do not necessarily imply that children with HFPDD act less morally than children with typical development.

As for the level of Internal Moral Reasoning, Furuhata has noted that on average students with typical development reached level III, relatively independent of both grade and scholastic ability.20 In the present study, independent of age, both the HFPDD and control groups generally reached level III on almost all of the items, but in both groups a small number of students reached level IV. These results are consistent with the Colby et al. findings in their 20-year longitudinal study of moral reasoning that few people move beyond conventional moral judgment to post-conventional moral judgment.22

In Empathy of Internal Moral Reasoning, the HFPDD group scored significantly lower than the control group. This result cannot be explained by the group difference in FSIQ or VIQ scores. Moreover, more students with HFPDD remained at level II in Empathy than did control students. Two plausible explanations can be provided to account for this difficulty in Empathy of students with HFPDD. First, questions regarding empathy, even if they are only written on a paper in an experimental setting, do demand more empathic ability than questions in the other aspects. Second, rules related to the Self-Help and Norm measures may be more easily indicated explicitly than those in Empathy. Therefore, these rules are easily understandable to students with HFPDD and it is easier for parents or teachers to give concrete advice to students with HFPDD in these aspects.

In the control group it is generally difficult to interpret the results of the correlation analyses. In contrast, in the HFPDD group there were significant positive correlations between some of the verbal ability-related items and Total score, Empathy, and Self-Help. These correlations may partly reveal the close relationship between verbal ability and internal moral reasoning in students with HFPDD. Of the subscales in WISC-III, Comprehension is considered to bear the closest resemblance to questions on Internal Moral Reasoning in HEART; thus, it is natural that there are correlations to a certain extent between Comprehension and items of Internal Moral Reasoning in HEART. However, this correlation can be regarded as unique to students with HFPDD because there was a strong positive correlation at 1% significance between Comprehension and Empathy in the HFPDD group, while there are no such significant correlations in their counterparts. It has been reported that children with HFPDD appeared to be able to compensate for some of their difficulties in social–emotional understanding by making use of their intellectual ability and that such associations are rarely found in normal development.23 Similarly, the results of the present study suggest a plausible hypothesis that students with HFPDD may compensate for their difficulties in internal moral reasoning by making full use of their verbal abilities.

Grant et al. indicated that children with autism did not differ from children with either moderate learning difficulties or children with typical development in judging damage to a person as being more culpable than damage to property.13 Their simplistic interpretation of the observation is that the children with autism in their study have been explicitly taught that damage to people is more culpable than damage to objects or property. Furthermore, Blair found that the level of ability on the ToM task is not associated with the tendency to distinguish moral and conventional transgressions; even the least able group of children with autism recognized the moral/conventional distinction.11 One of Blair's interpretations of this result is that the children with autism may not process as moral those transgressions (e.g. stealing) that require an individual to represent the mental state of another in order to realize that the other is a victim. Combining these interpretations with a relatively strong relationship between verbal ability and internal moral reasoning as suggested by the present study, the following conclusion can be drawn: students with HFPDD exclusively try to understand the question in HEART and then choose the answer that is closest to what they have been taught (by their parents or teachers) in previous similar conditions, and that they do this by executing their verbal ability without representing the mental state of the characters in the questions in HEART.

Some limitations of the present study should be addressed. First, some studies suggest that children with HFPDD have a difficulty in reasoning.24 Thus significant group differences in the mean score in Internal Moral Reasoning would result from a poorer representation of children's reasoning. However, in any case, it can be safely said that children with HFPDD could have much more difficulty in moral reasoning in everyday situations than hypothetical situations, given that everyday situations are far more complicated than hypothetical situations; for instance, they are full of subtle non-verbal cues that children with HFPDD have difficulty picking up. Second, the fact that internal consistency in Empathy is not satisfactory could be a problem, which may compromise the present results. Third, comparisons among subgroups (according to diagnosis or ToM) were not done, mainly due to the small number of subjects, which forced us to put PDD subgroups together as PDD. Fourth, the relationship between aging and development in moral judgment should be essentially investigated by following up the moral development of each individual. Finally, many factors are related to moral understanding, including the child's personality and a wide range of social experiences such as peer interaction, sibling interaction, child-rearing practice, schooling, and aspects of culture.5 However, some of these factors were not taken into account in the present study. Further studies based on a larger pool of subjects and using more rigorous methodology should be carried out.

In conclusion, children with HFPDD were associated with lower scores of Internal Moral Reasoning than matched healthy children, and these scores were partially predicted by level of verbal ability. These findings provide clinically useful implications for assessing sociality and educational intervention in individuals with PDD.


This work was supported in part by a grant-in-aid from the Ministry of Health, Labor and Welfare, Japan (to KK and NK). The authors thank Dr Mark A. Rogers and Mr. Mark Bogenschultz for helpful comments on the manuscript.


Appendix I

Example question in HEART (Internal Moral Reasoning – Empathy)

There is a school rule that states that if students forget to do their homework they must stay after school alone and finish it. Hanako who has not done her homework is staying back in the classroom to finish it. Her friend, Yoko, cannot decide whether to leave school but finally decides to go home.

Illustration with speech balloons: Hanako asks her classmates “Please wait for me. After I finish this assignment, let's go home together”. A couple of classmates are leaving the classroom. One says “Bye-bye”. The other says “Yoko, let's go”. Yoko looks perplexed beside Hanako.

Why does Yoko decide to go home?

  • 1Because waiting for Hanako is boring. (Level I)
  • 2Because if Yuko waits for Hanako, it won't give Hanako a chance to reflect on her punishment. (Level III)
  • 3Simply because it's the school rule to stay after school alone and finish the homework. (Level III)
  • 4Because the other classmates are already leaving school. (Level II)