An Effective Approach to Violence Prevention: Traditional Martial Arts in Middle School

Adolescence, Fall, 2001 by Gail Zivin, Nimr R. Hassan, Geraldine F. DePaula, Daniel A. Monti, Carmen Harlan, Kashfia D. Hossain, Ksai Patterson

Teacher ratings at end of Semester 1: controls vs. students. Figure 2 shows that, according to teacher ratings on the Sutter-Eyberg Student Behavior Inventory, controls (Group B at the end of Semester 1) increased over baseline in terms of violence and inappropriate social behavior in the classroom. Students (Group A at the end of Semester 1) consistently decreased across all four types of troublesome behavior. The results of t tests for uncorrelated samples were not significant.

Self-report at end of Semester 1: controls vs. students. On the happiness and schoolwork dimensions of the Piers-Harris Self-Concept Scale, Figure 3 shows consistent improvement over baseline self-reports for students (Group A at the end of the Semester 1). Simultaneously, there was definite deterioration from baseline for controls (Group B at the end of Semester 1). The t tests (one-tail, alpha = .05, Bonferroni alpha = .012) for uncorrelated samples indicated significant differences on how happy the boys had become and how good they reported their schoolwork to be. On Happy, t(50) = 1.83, p = .04; for controls (n = 22), M = - .05 (SD = .20), and for students (n = 30), M = .04 (SD = .14). On Work, t(50) = 3.00, p = .002; for controls, M = .06 (SD = .11), and for students, M = .03 (SD .11).

Attention at end of Semester 1: controls vs. students. Figure 4 shows consistent improvement from baseline for students (Group A at the end of Semester 1) on three of four measures of the IVA computerized test of attentional self-control. Meanwhile, controls (Group B at the end of Semester 1) deteriorated on three of the four measures. The t test (one-tail, alpha = .05, Bonferroni alpha = .013) for uncorrelated samples indicated a significant difference on Auditory Vigilance: t(46) = 1.75, p = .04; for controls (n = 22), M = -1.7 (SD = .15), and for students (n = 26), M = 7.7 (SD = .21).

Semesters 1 and 2: teacher ratings, pooled groups. Both groups were pooled and the improvements in teacher-rated behaviors were further examined with a larger sample size (n = 48), providing greater statistical power. Comparing both groups' baseline scores with their postcourse scores, using correlated t tests, revealed that the boys' scores improved (i.e., decreased) significantly (p < .05) on three of the four scales: Resists Rules, Impulsive, and Inappropriate Social Behavior (see Table 1).

Semesters 1 and 2: teacher ratings compared across groups. The postcourse raw scores (not difference scores) for the two groups were notably similar. Three of the four scale means differed by only about 0.1; the fourth differed by 0.25, which was not statistically significant (see Table 1).

Semesters 1 and 2: follow-up for Group A on teacher ratings. Teacher ratings at the end of Semester 2 were available for 19 of the original 32 boys who took the course during the first semester. These ratings constituted a follow-up measure a semester after having taken the course (see Figure 1). At 4-month follow-up, the boys in Group A were rated even better on three of the four scales of troublesome behavior (Violent, Resists Rules, and Impulsive). On Inappropriate Social Behavior, their mean rating remained better than baseline, but their improvement declined from that immediately following the course (see Figure 2).