LD, Interpersonal Understanding and Social Behavior in the Classroom

Abstract

This study used Baron and Kennys (1986) criteria for mediation to investigate the extent to which interpersonal understanding mediates the relation between learning disabilities (LD) and social adaptation in the classroom. Twenty-two children with and 22 children without a diagnosis of LD completed a semistructured developmental clinical interview measure of interpersonal understanding. They were also rated by their fourth- and fifth-grade teachers on a measure of social adaptation in the classroom. Interpersonal understanding and social adaptation in the classroom were found to be positively correlated. Children with LD exhibited less interpersonal understanding and social adaptation. Although this group difference on social adaptation was greatly reduced when interpersonal understanding was statistically controlled, it remained statistically significant. These results suggest that reduced social adaptation in the classroom and lower interpersonal understanding are both associated with a diagnosis of LD. However, they do not conclusively support the claim that interpersonal understanding mediates the relation between LD and social adaptation. Thus, whether the social difficulties of people with LD stem from the same complex phenomena that produce these people’s learning problems remains an open question.

Past investigations of children with learning disabilities (LD) focused mainly on these children’s cognitive and learning problems and on interventions designed to ameliorate these problems (Bryan, 1974; Toro, Weissberg, Guare, & Liebenstein, 1990). However, during the last decade, the awareness of these children’s behavioral, social, and emotional problems has considerably increased. This awareness was demonstrated by Bender and Smith’s (1990) meta-analysis of studies of the relation between LD and behavioral problems and Kavale and Forness’s (1996) meta- analysis of investigations of the relation between LD and social skill deficits. Both meta-analyses provided convincing evidence that children and adolescents with LD experience social problems (e.g., low self-esteem), emotional difficulties (e.g., depression), and conduct problems (e.g., aggression) and suggested that these problems and difficulties are, at least partly, due to deficits in social skills. This study examines the contribution of deficient interpersonal understanding to the social and behavioral problems that children with LD exhibit in the classroom.

Reports of the various behavioral problems of children with LD seem to indicate that related social and emotional inadequacies are core features of these problems (Americaner & Omizo, 1984; Bender, 1985; Bryan, 1974, 1977; Bryan & Pearl, 197 9; Gresham & Reschly, 1987; Kershner, 1990; Pearl, Donahue, & Bryan, 1986; Stone & La Greca, 1990; Vaughn, 1985). Acting-out misconduct in the classroom is often mentioned when such behavioral problems are discussed. Cornwall and Bawden (1992) reviewed a large number of studies of the relation between LD and behavioral difficulties and concluded that learning disabilities are associated with aggression and misconduct in the classroom. Although these personal and interpersonal difficulties are usually identified during the first years of formal education, when the child with LD first begins to experience difficulties in the classroom, adolescents and adults with LD have also been found to exhibit a high incidence of maladjustment forms such as delinquency and marital and vocational problems (Hoffmann, Minskoff, & Echols, 1987; Vogel & Forness, 1992; for a review, see Spafford & Grosser, 1993). These difficulties could be a consequence of such interpersonal processes as low selfesteem and feelings of inferiority. Alternatively, they could be a consequence of the same hypothetical neurological dysfunction that produces the learning problems. The former explanation is generally referred to as the secondary cause hypothesis, whereas the latter has been termed the primary cause hypothesis (Vogel & Forness, 1992).

Much of the research that has investigated whether the social difficulties of persons with LD is mediated by the impairment of interpersonal understanding or other higher order cognitive processes seems to-have been motivated by the primary cause hypothesis. Thus, Conte and Andrews (1993) and Spafford and Grosser (1993) claimed that the same neurological defect that creates difficulty in comprehending written language could also underlie problems in understanding nonverbal communication. Research based on such claims has revealed that children and adolescents with learning disabilities are less sensitive to the social meaning of gesture and facial expression and have more difficulty discriminating vocal tones than children and adolescents without learning disabilities (Holder & Kirkpatrick, 1991; Sisterhen & Gerber, 1989). This lack of sensitivity could seriously undermine the social interaction of individuals with LD.

Moreover, Oliva and La Greca (1988) suggested that the interpersonal problems of people with LD may be viewed as the consequence of an impaired ability to attain and apply metacognitive rules and strategies. This suggestion was based on their survey of research, which indicated that such an impairment may account for some of the problems that people with LD have with learning tasks. Oliva and La Greca (1988), along with other theoreticians and researchers (e.g., Bruck & Herbert, 1982), related this impairment to behavioral problems, because appropriate social interaction is assumed to require the application of such higher order cognitive processes as figure-ground discrimination and perspective taking.

Parril-Burnstein (1981) found that children with LD tended to produce less varied and more rigid coping strategies. These children were also unable to adapt appropriate cognitive strategies to different social situations (Gerber, 1983) or to organize spontaneously efficient strategies directed to the achievement of social goals (Oliva & La Greca, 1988).

However, the argument that deficient interpersonal understanding mediates the social and behavioral problems of people with LD is also consistent with the secondary cause hypothesis. Gresham and Elliott (1989), who also attributed the behavioral problems of children with learning disabilities to impaired social cognition, claimed that these deficiencies evolve from the lack of accumulated social experiences rather than from the same neurological cause as the LD. Gresham and Elliot (1989) pointed to research that showed children with LD to achieve less peer acceptance than children without LD (Gresham & Reschly, 1987; Haager & Vaughn, 1995; Kistner & Gatlin, 1989). Therefore, children with LD may have fewer opportunities to engage in social interaction and to accumulate the social experience that is the basis for interpersonal understanding.

Because the assumption that impaired interpersonal understanding mediates the social difficulties of people with LD is consistent both with the primary and the secondary cause hypothesis, our study only constitutes a test of the more general interpersonal understanding mediation hypothesis. Past studies of variations on this hypothesis have limited their assumptions to the examination of the relation of social cognition and interpersonal understanding to learning disabilities and to the social difficulties of people with LD (Bruck & Hebert, 1982; Stiliadis & Wiener, 1989; Toro et al., 1990).

However, evidence in support of these relations alone is not sufficient to confirm the claim that social cognition and interpersonal understanding are mediating variables for LD. According to Baron and Kenny’s (1986) classic distinction, a moderator variable is a variable that influences the direction or strength of the relation between an independent variable and a dependent variable, whereas a mediating variable is a variable that is assumed to account for a significant part of the relation between an independent variable and a dependent variable. In keeping with this distinction, the evidence that variables reflecting social perception, social cognition, or interpersonal understanding mediate the relation between LD and social functioning should consist of a specific set of hypothesized correlations, regressions, or differences between means. Four hypotheses can be inferred from Baron and Kenny’s approach to assessments of mediating variables. The first three hypotheses relate to (a) the relation between the independent variable (i.e., the presence or absence of a learning disability) and the dependent variable (i.e., behavioral problems in the classroom), (b) the relation between the independent variable and the mediating variable (i.e., interpersonal under-, standing), and (c) the relation between the mediating variable and the dependent variable. However, the final test of the mediating role of a variable is the hypothesized reduction of the relation between the independent variable and the dependent variable when the mediating variable is statistically controlled for.

The previously cited studies did not make use of Baron and Kenny’s strict criteria when they evaluated the relation between interpersonal understanding and the social difficulties of people with LD. Only Stiliadis & Wiener (1989) examined certain of these correlations and regressions to assess the contribution of social perception to the lower peer status of children with LD. However, even they did not test explicitly and formally the mediating role of their two measures of social perception. Our research investigated all the hypothesized relations that are logically related to the assumption that a variable is a mediator. If interpersonal understanding mediates the relation between LD and difficulties with social interaction, the following relations should be found:

1. LD should be positively related to difficulties with social behavior.
2. LD should be positively related to impaired interpersonal understanding.
3. Impaired interpersonal understanding should be positively related to difficulties with social behavior.
4. When impaired interpersonal understanding is controlled for, the relation between LD and difficulties with social behavior should be significantly reduced.

For the purpose of this study, teachers’ ratings of classroom behaviors that appear to reflect social functioning served as a measure of social behavior in the classroom. Gresham (1984) found such measures to be highly reliable. Furthermore, as pointed out above, a large number of studies have shown that learning disabilities are associated with aggression and misconduct in the classroom (Bender & Smith, 1990; Cornwall & Bawden, 1992). Finally, classroom behavior represents a broad range of social functioning in a significant area of life.

In keeping with this study’s focus on social behavior in the classroom, the classification of LD was determined by formal. placement committees set up by Israel’s Ministry of Education. These committees consist of professionals such as educational psychologists, special education teachers, and social workers. They diagnose and allocate resources for children with learning disabilities within the public school system. These committees’ evaluations are based on academic performance and standardized psychological tests.

The mediating hypothesis investigated by our study is based on the assumption that deficits in the central processing of social information are a correlate of LD. A major conceptual and methodological problem of this assumption is the variety of terms and measures that have been used with regard to this process in investigations of the social and behavioral problems associated with LD. This process has been referred to as social perception, social problem solving, and social skills, and measures have been selected to match the different terms. Tasks and behavior rating scales that assess social perception seem to tap the individual’s ability to interpret social cues (e.g., Stiliadis & Wiener, 1989). Operationalizations of social problem solving take the form of tasks that require generating alternative solutions in response to a specific social problem presented in a brief story. This operationalization suggests that the process of social problem solving involves cognitive flexibility in social situations (e.g., Toro et al., 1990). Social skills is a generic term that may include both social perception and social problem solving but is generally considered to include such social functions as the ability to initiate a conversation with a classmate or ask directions from a stranger (e.g., Gresham & Elliot, 1989).

In this study, social information processing was conceptualized and measured in the context of Selman’s (1980) theory of the development of interpersonal understanding. According to this theory, a central aspect of mature interpersonal understanding is the development of an individual’s ability to differentiate between his or her own viewpoint and the viewpoint of others and to compare these viewpoints. This aspect of interpersonal understanding is often referred to as perspective or role taking (Flavell, 1977). Selman has provided a clinical interview method to assess the stage of interpersonal development that an individual has reached. This method consists of the content analysis of responses to short, provocative vignettes. One of the deficits in social information processing that has been associated with LD is the reduced tendency to take the thoughts and feelings of others into consideration when making social decisions (e.g., Clements, 1987). Previous studies that compared the role and perspective taking ability of children with and without LD generally found the role taking ability of children with LD to be lower than that of children without LD (Bruck & Hebert, 1982; Dickstein & Warren, 1980; Horowitz, 1981; justice & Beard, 1982; Wong & Wong, 1980). In keeping with the primary cause hypothesis, Bruck and Hebert (1982) showed that role taking was consistently related to measures of hyperactivity. Consequently, we adopted Selmans (1980) conceptual and theoretical framework, with its emphasis on the development of role and perspective taking, to investigate the interpersonal understanding of children with learning disabilities.

In keeping with this conceptual and empirical framework, our hypotheses were

1. The classroom behavior of children diagnosed with LD will be rated by teachers as less socially adaptive than the classroom behavior of nondisabled children.
2. Children diagnosed with LD will be rated lower on Selmans (1980) measure of interpersonal understanding than nondisabled children.
3. Ratings of the interpersonal understanding of children with and without LD should correlate positively with teachers’ ratings of these children’s socially adaptive behavior in the classroom.
4. When the ratings of interpersonal understanding are controlled for, the difference between the classroom behavior of the children with LD and that of the nondisabled children should be significantly reduced.

Method

Participants

Twenty-two children with learning disabilities and 22 children without learning disabilities, all boys, participated in this study. These children studied in the fourth and fifth grades of two general education elementary schools located in middle class neighborhoods in Jerusalem. Although the children with LD participated with the other children in classroom activities such as natural science, English as a second language, gymnastics, and arts and crafts, they also attended additional special education classes. All the children with LD were placed in these classes after being diagnosed with LD by formal assessment committees set up by the Ministry of Education. These committees’ diagnoses were based on academic performance and standardized psychological tests such as the WISC-R (Wechsler, 1974), the Bender-Gestalt test (Koppitz, 1975), Figure Drawings (Koppitz, 1968), and achievement tests. Although the criteria for the diagnosis of LD used by these committees are relatively general and flexible, two specific inclusion criteria are (a) achievement test scores at least 2 years below grade level and (b) average or above-average intelligence with deficits in specific cognitive areas such as language use and arithmetic. Specific exclusion criteria are extreme behavioral and attentional difficulties and the presence of clearcut neurological problems. Furthermore, due to possible language difficulties, a final diagnosis is given only to children who have resided in Israel for’ at least 4 years (Ministry of Education, Culture, and Sports, 1996).

Measures

Interpersonal understanding

Interpersonal understanding was assessed by Selman’s (Selman, 1980; Selman & Byrne, 1973, 1974) semistructured clinical interview. This consists of reading a short description of a social problem to the participant. The participant is asked a number of open questions about how the story’s protagonists might respond to the problem. These questions require the participant to explain his or her answer. The participant’s answers are rated on a 6-point scale. This scale represents the level of interpersonal understanding and requires that the raters evaluate the extent to which and the manner in which the participant takes into consideration the multiple perspectives that bear upon the problem.

The following social problem, which was translated and adapted from Selman and Byrne (1974), was read to each of the participants:

Assume that your best friend John is the best tree climber in the class. Whenever the children compete among themselves as to who can climb the highest and fastest, he always wins. One day, a group of children were playing near John’s house. Suddenly, the ball the children were playing with landed on a tree next to the house. John began to climb quickly. Without warning, the branch broke and he fell. His father saw all of this from the window, and when John returned home his father told him that he no longer was permitted to climb trees. John promised that he would not.

A few days later, on his way back from school, John heard children calling him to come over. When John came over, a little girl approached him with tears in her eyes and asked him to take down her favorite cat from a tree. The cat sat on the tree and was afraid to come down, and none of the children knew how to climb. John asked himself what he should do.

After the story was read, the participants were asked 11 open questions about it. The following are examples of these questions:

“Why, in your opinion, did John wonder what he should do?”

“Suppose that John decided to help the girl and climb the tree, and his father found out about it. What, in your opinion, would his father think about what John had done?” “Suppose that John decided not to climb the tree. What, in your opinion, would the children and the girl who asked for help think about what he had done?”

The participant’s oral answers were recorded. These answers were rated by three independent raters-two of the researchers and a special education teacher-in accordance with the six stages of Selman’s (Selman & Byrne, 1974; Selman, 1980) model of the development of interpersonal understanding. The essential attributes of these stages are as follows:

Stage I (rating = 0).

At this stage, no distinction is made between the participant’s own point of view and the points of view of the protagonists in the vignette. Although the participant is aware that the protagonists have a point of view, he or she identifies their point of view with his or her own point of view (“John wondered what to do because he fell off the tree”).

Stage 2 (rating = 1).

At this stage, the participant is aware that the protagonists have different points of view, which are distinguishable from each other. There is awareness that the protagonist in the vignette can think about others and about him- or herself, but there is no indication that the participant is aware that the protagonist takes this awareness into consideration (“If John decided to help the girl and to climb the tree, John’s father would think that he did something bad”).

Stage 3 (rating = 2).

At this stage, the participant is aware that the protagonists partially base their points of view on the points of view of others (“If John told them about his promise, the girl would be less angry if he didn’t climb the tree”).

Stage 4 (rating = 3).

At this stage, the participant is aware that the protagonists can simultaneously have a number of points of view, which they can integrate in a relatively objective manner (“If John told them about his promise, they would be a bit disappointed because of the promise but, on the other hand, they would think he is right”).

Stage 5 (rating = 4).

At this stage, the participant is aware that the protagonists can have relatively objective points of view that do not necessarily lead to consensus because communication in a group is regulated by social agreements (“If John was aware that the other children knew about his promise but still ask him to climb the tree, he wouldn’t climb. He would think that they wanted to test him, and he would be irritated”).

Stage 6 (rating = 5).

At this stage, the participant is aware that social viewpoints are relative rather than absolute and that the protagonists examine events according to their personal systems of analysis, which are influenced by such factors as their life histories, the social setting, and their present emotional status (“If, instead of the girl’s cat, John’s bird, which he had been taking care of for many months and which he loved very much, was on the tree, John would climb the tree if he were selfish. If he wasn’t selfish, he would help both the bird and the cat”).

Each participant’s answers were rated by three raters. The answers were rated blindly; the raters were not informed which answers were produced by which children. In keeping with Selman’s recommendation, each participant was given the highest rating that he achieved over all answers. Interrater reliability was assessed by Pearson product correlations between each pair of raters and ranged from .79 to .84. Each participant’s interpersonal understanding score was the mean rating of his answers over the three raters.

Classroom Behavior. Schaefer and Edgerton’s (McKinney & Forman, 1982; Schaefer & Edgerton, 1978; Schaefer, Edgerton, & Hunter, 1983) classroom behavior inventory was used to measure the classroom behavior of the children in the LD and the control group. Margalit and Shulman (1986) prepared a 42-itern Hebrew translation and adaptation of this in inventory. The items in this measure depict typical classroom behavior. Teachers were requested to indicate how frequently each behavior occurred on a response scale from 1 (never) to 5 (very often). This scale provides scores in the following ten areas: personality, verbal intelligence, creativity, curiosity, consideration, hostility, introversion extroversion, goal orientation, distractibility, and independence-dependence. For the overall Hebrew translation and adaptation of the scale, Cronbach’s alpha was .90. A factor analysis of the Hebrew version of the scale produced the same three factors that Shaefer & Edgerton (1978) revealed for the original English version, namely

1. academic aptitude, composed of verbal intelligence, creativity, and curiosity (Cronbach’s alpha .96);
2. social adaptation, composed of consideration versus hostility (Cronbach’s alpha = .95); and
3. extroversion, composed of extroversion versus introversion (Cronbach’s alpha = .80).

In this study, the social adaptation factor was used. Each classroom teacher was asked to rate the children in both groups only on the items of this factor. Cronbach’s alpha was .94 over all these items for both teachers.

Intelligence. Raven’s Progressive Matrices (Raven, Court, & Raven, 1983) were used to assess the intellectual level of the participants. The percentage of correct responses constitutes the participant’s score X = 77.1, SD = 5.4, for children without LD; 9 = 81.2, SD = 6.1 for children with LD; t(42) = 1.73, p < .05). As this test is a nonverbal measure of intelligence (Brothers, 1965), it is especially appropriate for the investigation of children with learning disabilities.

Procedure

Each participant was interviewed and tested individually. The Raven Standard Progressive Matrices were administered first, followed by Selman’s semistructured clinical interview. The classroom teachers participating in the study independently filled out the classroom behavior inventory.

Results

To test the hypothesis that the classroom behavior of children diagnosed with LD would be rated by teachers as less socially adaptive than the classroom behavior of nondisabled children, a t test was carried out on the mean ratings of social adaptation for the children with and without LD. This test indicated that the children without LD received a statistically significant higher rating on social adaptation than the children with LD (nondisabled children: X = 4.23, SD = 0.69; children with disabilities; 9 = 2.98, SD = 0.89), t(40) = 5.18, p < .001).

To test the hypothesis that children diagnosed with LD would be rated lower on Selman’s measure of interpersonal understanding that nondisabled children, a t test was carried out on the mean ratings of interpersonal understanding of both groups of children. According to this analysis, the responses of nondisabled children were created as expressing statistically significant more interpersonal understanding than those of children with LD (nondisabled children: 9 = 2.10, SD = 0.19; children with LD: 9 = 1 .10, SD = 1.05, t(40) = 3.32, p < .01.

To test the hypothesis that the ratings of interpersonal understanding of the children with and without learning disabilities would correlate positively with teachers’ ratings of these children’s social adaptation in the classroom, a Pearson product-moment correlation between these two measures was calculated, yielding a statistically significant correlation, r(40) = .70, P < .01.

An analysis of covariance was carried out to determine the extent to which interpersonal understanding mediated the difference between the social adaptation of the children with and without LD. For this analysis, group (LD versus control) was the independent variable, interpersonal understanding the covariant, and teacher rating of social adaptation the dependent variable.

Original mean social adaptation ratings were 2.98 and 4.23 for the children with LD and the non-LD children, respectively. After statistical adjustment, the means were 3.22 and 3.90, respectively. Thus, the initially lower mean social adaptation rating for the children with LD was adjusted upward, whereas the initially higher mean social adaptation rating for the children without LD was adjusted downward. If the covariate interpersonal understanding is highly related to the dependent variable social adaptation and if this relation accounts for all or part of the between- group difference on the dependent variable, these directions of change in the adjusted means are to be expected (Kerlinger & Pedhazur, 1973). As social adaptation appears to some extent to be a function of interpersonal understanding, removing the influence of interpersonal understanding from the difference in mean ratings on social adaptation should decrease the mean for children who are high on interpersonal understanding and social adaptation and increase the mean for children who are low on interpersonal understanding and social adaptation. Thus, adjusting for interpersonal understanding greatly reduced the difference in social adaptation between the two groups of children. However, despite this reduction in the difference between the mean ratings on social adaptation, the difference remained statistically significant, F(1, 41) = 12.30, p < .01.

Although the groups with and without LD were matched on intelligence on the basis of their performance on Raven’s Progressive Matrices (Raven et al., 1983), the correlations between intelligence as measured by the Raven’s Matrices and socially adaptive classroom behavior and interpersonal understanding were examined. The correlations between these variables (r = .29 and r = .28, respectively) were found not to be statistically significant (p > .05). Thus, the results appear to be independent of the participants’ general intelligence.

Discussion

The results of this study seem to confirm only the first three of its four hypotheses. With regard to the fourth hypothesis, the results are inconclusive. Although controlling for interpersonal understanding considerably reduced the difference in social adaptation scores between children with and without LD, the difference between the two groups remained statistically significant. Consequently, this study provides only limited support for the assumption that the problems in social functioning often exhibited by children with learning disabilities may be due in part to these children’s relative lack of interpersonal understanding. Teacher ratings showed that the classroom social behavior of children with LD was less adaptive than that of children without LD. This finding is in keeping with our first hypothesis and with the results of other studies of the interpersonal and social behavior of children with LD.

Furthermore, the ratings of these children’s responses to an interpersonal developmental task adapted from Selman (Selman, 1980; Selman & Byrne, 1973) showed that children with LD exhibit less interpersonal understanding than children without LD. The mean interpersonal understanding rating for the children with LD (X = 1.10) places them approximately at the second of Selman’s six stages of the development of interpersonal understanding, whereas the children without LD received a mean interpersonal understanding rating of 2.10, which places them approximately at the third stage of interpersonal development. According to Selman (1980), children between the ages of 5 and 9 years produce responses at the second level of interpersonal understanding development, whereas children between the ages of 10 and 12 pro-, duce responses at the third level of interpersonal development. The ages of the children who participated in this study ranged from 9.5 to 11.5 years. Thus, although the ratings of interpersonal understanding of the children without LD seemed to fall within the expected age range, the ratings for the children with LD were below those expected for children in their age range. In terms of the psychological process that theoretically emerges at each stage of development, the interpersonal understanding of children with LD is more egocentric and, therefore, less influenced by their awareness of the points of view of others, even though they do demonstrate an awareness of these points of view. Thus, their acting-out behavior could easily be interpreted by their peers, and by authority figures such as parents and teachers, as willful disobedience, although it may stem in part from their lack of appreciation of the significance of others’ points of view.

As predicted, a high, statistically significant positive correlation was revealed between adaptive social behavior in the classroom and interpersonal understanding; children who were found to possess less interpersonal understanding tended to be rated as displaying less adaptive social behavior. We also found that the mean difference between the adaptive social behavior of the two groups was reduced when the influence of interpersonal understanding was statistically held constant. Both these findings support the claim that the lower ratings of social adaptation of the children with LD is due at least in part to these children’s deficient interpersonal understanding. Deficient interpersonal understanding might account for some of the social difficulties often experienced by children with learning disabilities.

However, this claim is considerably weakened by our finding that the difference in classroom social adaptation between children with and without LD remained statistically significant after interpersonal understanding was statistically controlled for. As we pointed out in the introduction, the social difficulties of people with LD may stem from the direct impact of the neuropsychological condition that produces the LD, from the low self-esteem that is one of the consequences of the LD, from the indirect effect of the social stigma associated with LD, which limits opportunities to exercise interpersonal understanding, or from a combination of these factors. Thus, the significant difference in classroom social behavior that remained after the measure of this behavior was adjusted by statistically removing the influence of interpersonal understanding may be due to the children’s reaction to the emotional and social consequences of their poor academic performance.

According to these considerations, interpersonal understanding, as measured by Selman’s perspective taking task, mediates at most some of the social difficulties that a child with LD experiences in the classroom. This study used Baron and Kenny’s (1986) formal criteria for mediation to test its version of the interpersonal mediation hypothesis explicitly. The implications of this study depend to a great extent on our theoretical and empirical understanding of the results of this test and of the construct that assumes the mediating role. Because the difference in classroom behavior between the children with and without LD remained statistically significant after interpersonal understanding was controlled for, this study has at most shown that part of the apparent social misconduct of children with LD in the classroom may be a result of their relative lack of interpersonal understanding. However, both this social misconduct and the deficient interpersonal understanding may simply be correlates of other consequences (e.g., low self-esteem, social isolation) of LD.

In either case, the specific implications of our findings for understanding how aspects of LD influence social behavior depend to a great extent on the overall validity of Selman’s developmental theory and measure of interpersonal understanding. The phenomena with which this theory and measure are concerned are complex, involving various social-cognitive processes (Cole & Cole, 1989; Flavell, 1977). Similarly, the phenomena that are the referents of the diagnosis of LD are also quite complex. As this study has demonstrated, no single field or correlational study can provide definitive answers to questions concerning the etiological link between these phenomena and the social behavior of persons with LD. Therefore, further research should focus on revealing the extent to which and the manner in which aspects of LD overlap with aspects of interpersonal understanding and social conduct. This research should be carried out with larger samples consisting of girls as well as boys and include measures of verbal ability to control for the possibility that the lower scores on measures of interpersonal understanding may be due to the impaired linguistic skills of children with LD.

About the authors

Shlomo Kravetz, PhD, is an associate professor of psychology at Bar-Ilan University, Israel. His current interests include psychiatric rehabilitation, learning disabilities, and the neuropsychology of language. Miriam Faust, PhD, is a lecturer in psychology at Bar-Ilan University. Her current interests include psycholinguistics, neuropsychology of language, and learning disabilities. Shahar Liphshitz, BA, is currently a graduate student of law at BarIlan University. Shlomo Shalhav, BA, is currently a graduate student in clinical child psychology at Bar-Ilan University. Address: Shlomo Kravetz, Department of Psychology, Bar-Ilan University, Ramat-Gan, Israel 52900.