آزمون فرضیه رفتار ضداجتماعی آیزنک در به کارگیری دانش آموزان 11-15 ساله دو بخشی برای PEN و L

Abstract This study examined the strongest form of Eysenck’s antisocial behavior hypothesis. The hypothesis predicts that high scores on the P, E, and N traits combined with a low score on the L Scale from Eysenck’s personality instrument is a profile that holds the greatest risk for the development of antisocial behavior. Ninety-four 11–15-year-old participants were selected from a pool of 763 potential participants. Participants in the study were divided into two groups of 47 matched by age and gender. One group consisted of participants who met the at-risk profile and the second group had the opposite profile. All participants were administered a self-report scale that assessed externalizing conduct behavior. An ANOVA was used to test the dichotomous categorization of the participants. All differences between the groups on the traits comprising the independent variable were highly significant. An independent t-test was used to test the difference between the groups on the dependent variable. This difference was highly significant in the predicted direction and yielded a very large effect size. A regression analysis determined that three of the four scores used for categorizing participants best predicted the dependent variable. Further, the analysis indicated a significant interaction between P and E. The results are discussed as well as suggested directions for future research.

Introduction The problem of children and adolescents exhibiting antisocial behavior (ASB) in public school programs has been widely discussed (Kamps & Tankersley, 1996; Maag & Howell, 1991; Nelson, Center, Rutherford, & Walker, 1991; Nelson, Rutherford, Center, & Walker, 1991; Skiba, Peterson, & Williams, 1997; Sprague & Walker, 2000; Vance, Fernandez, & Biber, 1998). There are many factors that contribute to the development of conduct problems (McMahon & Wells, 1998; Sprague & Walker, 2000), including a number of biological factors (Chess & Thomas, 1987; Niehoff, 1999). One clearly articulated theory that addresses biological factors in ASB is the biosocial theory of Eysenck (1997a). Eysenck’s temperament-based theory is sometimes referred to as a three-factor model of personality in which the three factors are Extraversion (E), Neuroticism (N), and Psychoticism (P). The Extraversion (E) trait is represented by a bipolar scale that is anchored at one end by sociability and stimulation seeking and at the other end by social reticence and stimulation avoidance. Extraversion is hypothesized to be dependent upon the baseline arousal level in an individual’s neocortex and mediated through the ascending reticular activating system (ARAS) (Eysenck, 1967, Eysenck, 1977 and Eysenck, 1997a). The difference in basal arousal between introverts and extraverts is evident in research on their differential response to drugs. Claridge (1995) reviewed drug response studies that demonstrated introverts require more of a sedative drug than do extraverts to reach a specified level of sedation. This finding is explained by the higher basal level of cortical arousal in introverts. The Neuroticism (N) trait is anchored at one end by emotional instability and spontaneity and by reflection and deliberateness at the other end. This trait’s name is based on the susceptibility of individuals high on the N trait to anxiety-based problems. Neuroticism is hypothesized to be dependent upon an individual’s emotional arousability due to differences in ease of visceral brain activation, which is mediated by the hypothalamus and limbic system (Eysenck, 1977 and Eysenck, 1997a). The Psychoticism (P) trait is anchored at one end by aggressiveness and divergent thinking and at the other end by empathy and caution. The label for this trait is based on the susceptibility of a significant sub-group of individuals high on the P trait to psychotic disorders (Eysenck & Eysenck, 1976). Psychoticism is hypothesized to be a polygenic trait (Eysenck, 1997a). Polygenic refers to a large number of genes each of whose individual effect is small. Each of these “small effect” genes is additive, so that the total number inherited determines the degree of the P trait in the personality. When the small effect genes are found in combination with large effect genes associated with susceptibility to psychotic disorders, there is a significant increase in the risk for developing a psychosis. Eysenck’s theory predicts that individuals high on the P trait will be predisposed to developing antisocial behavior (Eysenck, 1997a). Further, an individual high on both the P and E traits will be predisposed to developing antisocial, aggressive behavior. Aggressive behavior is associated with low cortical arousal (high E) because a person with a relatively under reactive nervous system does not learn restraints on behavior or rule-governed behavior as readily as do individuals with a higher basal level of cortical arousal. Further, when such an individual is high on the N trait as well, this will add an emotional and irrational character to behavior under some circumstances. Finally, antisocial individuals typically score lower than others on the Lie (L) Scale in Eysenck’s personality questionnaire. The L Scale is a measure of the degree to which one is disposed to give socially expected responses to certain types of questions. A high score on this scale suggests that the respondent is engaging in impression management. A low score suggests indifference to social expectations and is usually interpreted as an indication of weak socialization. The strongest form of Eysenck’s antisocial behavior (ASB) hypothesis would be high P, E, and N with low L. Center and Kemp (2002) conducted a meta-analysis of Eysenckian studies examining personality differences in children with and without behavior problems. This analysis found a large effect size (ES) (Cohen, 1988) for P (ES=0.89), a low intermediate effect size for N (ES=0.43), a small effect size for E (ES=0.20) and an intermediate effect size for L (ES=−0.51). In this meta-analysis it became clear that what the prior studies had failed to do was examine the interactive effects of the traits in Eysenck’s ASB hypothesis. The reason for this was that the prior studies employed Eysenck’s personality traits as dependent variables and behavior as the independent variable. Eysenck clearly views his personality traits as a psychometric approach to assessing antecedents (Eysenck, 1997b). His traits are psychometric constructs that provide a conceptual interface between the distal and proximal antecedents for behavior and outcomes or proximal and distal consequences. Conduct behavior should be classified as a distal consequence of various antecedents including temperament-based personality traits. The PEN traits serve as a conceptual interface between antecedents and consequences. Thus, they should be used as independent variables rather than distal consequences as has been the case in most studies. It is only by using the PEN traits as independent variables that the interactive effects can be explored. There presently appear to be only two studies that have used Eysenck’s traits as independent variables in a test of one of his hypotheses (Jackson & Center, 2002; Jackson, 2003). Eysenck (1976) proposed a hypothesis concerning good behavior and the development of conscience, which will hereafter be referred to as the morality hypothesis. In the morality hypothesis, Eysenck (1976) proposed that good conduct could be the result of socialization that establishes a system of conditioned inhibitions on behavior. Eysenck’s hypothesis was that individual differences in susceptibility to conditioning result from the interaction of two temperament traits: Extraversion (E) and Neuroticism (N). Persons high on E are less responsive than persons low on E to the conditioning of operant and respondent responses. High N adds an emotional character to behavior, which often leads to an over reaction. Eysenck hypothesized that individuals who are low to average on both the E and N traits will be more likely to acquire an effective system of inhibitions or conscience. An effective system of behavioral inhibitions would, in turn, lead to a pattern of behavior likely to be characterized as good conduct. Jackson and Center (2002) conducted an initial test of the morality hypothesis by organizing a sample of students into three groups according to their Eysenckian trait scores. One group was above the mean on both E and N; a second group was below the mean on both E and N, and the third group was mixed. All of the students had also completed the Externalizing Scale from the Youth Self-report Scale (YSR) (Achenbach, 1991). A significant difference (p<0.05) was found between the groups defined as being either above the mean or below the mean on both the E and N traits and in the predicted direction. Jackson (2003) improved on the earlier study by using two groups of equal size and matched by age and gender of which one group was above the mean on both the E and N traits and the other group was below the mean on both traits. In addition, the effects of the P trait were controlled for by requiring that each participant’s P score had to be within the range of normal variation. The combined effect of E and N were assessed for moral reasoning as well as externalizing behavior. Moral reasoning was assessed using the Defining Issues Test ( Rest, 1986) and externalizing behavior was assessed using the YSR. Statistically significant differences were found for both dependent variables at the 0.01 level and in the expected directions. The current study examined the strongest form of Eysenck’s ASB hypothesis. This hypothesis proposes that the interaction of P, E and N when all are high and combined with low L will create the greatest susceptibility to the development of antisocial behavior. To test the hypothesis participants who were low on P, E and N and high on L served as the contrast group.

Results Statistical analysis was done using the Statistical Package for the Social Sciences (Version 11.0) for the PC. The first analysis done was an ANOVA to determine, if the screening procedure had selected two groups that were significantly different from one another relative to their respective personality profiles. The differences between the groups were statistically significant for P (F=232.011 (1, 92), p<0.001), for E (F=132.5.75 (1, 92), p<0.001), for N (F=305.277 (1, 92), p<0.001), and for L (F=267.434 (1, 92), p<0.001). The test of the ASB hypothesis was performed using an independent samples t-test. The difference between the two groups on the Externalizing Scale of the YSR was statistically significant (t=−11.948, df=92, p<0.001). The mean and standard deviation for the low PEN and high L group on the YSR was M=8.87 and SD=5.45. The mean and standard deviation for the high PEN and low L group on the YSR was M=27.47 and SD=9.17 (see Table 1). The bias corrected effect size was 2.45, which is considered very large ( Cohen, 1988). Table 1. Means and standard deviations for each variable in each of the two sample groups Variable Group n Mean SD P Low 47 1.17 1.01 High 47 7.83 2.82 E Low 47 14.51 3.86 High 47 21.32 1.25 N Low 47 6.09 2.76 High 47 15.32 2.35 L Low 47 11.19 2.54 High 47 3.45 2.02 YSR Low 47 8.87 5.45 High 47 27.47 9.17 Table options The final analysis was a backward regression analysis of the four scale scores from the JEPQ on the dependent variable (YSR). This yielded a model that was highly significant (F=64.357 (3, 90), p<0.001). The adjusted R square for this model was 0.671. The three scales from the JEPQ retained in the model and their respective standardized Betas were P (0.367), E (0.166) and L (−0.368). The regression analysis also examined all possible combinations of PEN and L for interactions. The only interaction was between P and E and the model accounting for the most variance was PE with L (F=92.559 (2, 91), p<0.001). The adjusted R square for this model was 0.670. The standardized Betas were PE (0.435) and L (−0.424).