METHOD Research Subjects The research subjects were 267 grade 9 students from three secondary schools located in an affluent suburban community. One of the schools was a vocational school serving students from the entire district and included basic level students only. The other two schools (academic) had students at general and advanced levels and were considered to have students representative of those within the district. If students were taking two or more academic courses at the general level they were classified as general. Measuring Instruments The Nowicki-Strickland Locus of Control Scale for Children (NS) was used to measure generalized expectancies and the IARQ was used to measure specific expectancies (I+ and I-). Because of limited testing time a 21-item version of the NS (Nowicki & Strickland, 1973) for grades 7-12 and a 20-item version of the IARQ (Crandall, 1968) for grades 6-12 were used. Crandall (1968) reported correlations of .89 for I+ and .88 for I- between the short form (grades 6-12) and the long form of the IARQ. Data Collection The measuring instruments were administered to students in grade 9
English classes by their teachers during a regular class period. Prior
to test administration, a researcher met with the teachers or
vice-principal to review administration procedures. Testing at two
schools (the vocational school and one of the academic schools) was
completed in December 1985. Data were collected from the other
academic school in April 1986. RESULTS Reliabilities of the three measures were estimated using the * coefficient. They were .70 for the NS, .43 for I+, and .46 for I-. The reliabilities for I+ and I- were quite low; this could have increased the probability of type II error when data were analyzed. Prior to testing the hypotheses, analyses of variance were used to test for differences between the two academic schools. No significant differences were found between schools. To determine whether differences were as hypothesized, analyses of variance were used with level and sex as independent variables and NS, I+ and I- scores as dependent variables. Results are shown in Table 1. Since sample sizes were not equal an unweighted means analysis was used. Significant differences were found only for level on both the Nowicki-Strickland and I-. Post hoc, the Tukey-Kramer modification of Tukey's WSD procedure was used to determine which groups differed significantly. For the NS, advanced level students were significantly more internal than either general or basic level students (See Table 2 for means). These results were as expected. For I-, advanced level students were significantly more internal than general level students; basic level students did not differ significantly from either advanced or general level students. Sex differences were not found for any of the dependent variables. It had been predicted that females would have I- scores higher than those of males. Although not significant, the opposite was found. Males had an I- mean higher than that of females. Their mean for I+ minus I-, was used as a dependent variable. A significant sex difference was found. For males, the difference between I+ and I- scores was less (M = .32) than that for females (M = .88). To further explore, correlated t tests were used to determine whether males or females scored significantly higher on I+ as compared to I-. Females did score significantly higher on I+ than on I- (t = 3.35); there was no significant difference for males (t = 1.68). The results for sex were not consistent with what had been suggested in the literature. In examining the means for sex by groups it would appear that the sex differences are attributable to basic and general level students. At the advanced level the mean differences between I+ and I- were .46 for males and .31 for females, while at the general level differences were .26 for males and 1.43 for females; at the basic level males differed by .14 and females by 1.27. Thus for basic and general levels, females scored substantially more internal for academic success than for academic failure. It would be of interest to explore these relationships in future studies. |