TPC Journal-Vol 10- Issue 3-FULL ISSUE

358 The Professional Counselor | Volume 10, Issue 3 Table 4 Bivariate Correlations Between Predictor Variables 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 1. Nonsuicidal SIB - 2. Ambivalent SIB .35 - 3. Suicidal SIB .16 .46* - 4. Problem-Solving .35 .44* .19 - 5. Roles .21 -.01 .27 .39* - 6. Affect. Resp. .25 .36 .28 .68*** .42* - 7. Affect. Involv. .22 .13 .27 .52** .43* .72*** - 8. Conflict .38* .35 .26 .66*** .14 .45* .15 - 9. Comm. (AD) .47* .56** .61*** .57** .44* .51** .50** .43* - 10. Comm. (C) .09 .11 .04 .42* .28 .29 .29 .27 .17 - 11. Beh. Cont. (AD) .27 .06 -.21 .54** .48** .51** .56** .34 .41* .25 - 12. Beh. Cont. (C) .27 .26 .13 .49** .53** .61*** .36 .38* .33 .40* .47* Note. SIB = self-injurious behavior; Affect. Resp. = Affective Responsiveness; Affect. Involv. = Affective Involvement; Comm. = Communication; (AD) = adolescent report; (C) = caregiver report; Beh. Cont. = Behavioral Control. * p < .05. ** p < .01. *** p < .001 . Next, we used SEM to predict SIB with our simplified model of family functioning. We tested three SIB outcomes separately because of concerns with sample size. For all models predicting SIB, we freed all FAD factors (problem-solving, roles, affective responsiveness, affective involvement, adolescent-reported communication and behavioral control, and caregiver-reported communication and behavioral control) to correlate because variables from the same measure are likely to be related. The model predicting nonsuicidal SIB had good absolute fit: χ 2 (7) = 4.28, p = .747, CFI = 1.00, RMSEA = 0.00, SRMR = 0.04. In all, family functioning explains 20% of the variance in nonsuicidal SIB. See Figure 2 for the standardized path coefficients between family functioning variables, the latent variable of family functioning, and nonsuicidal SIB. Notably, family functioning predicted nonsuicidal SIB: β = .44, B = 1.27, SE B = 0.62, p = .039. Based on effect sizes (see Figure 2), the strongest predictors were problem-solving (averaged; β = .79, B = 0.90, SE B = 0.03, p = .008, R ² = .62), communication (adolescent- reported; β = .55, B = 0.05, SE B = 0.03, p = .034, R ² = .31), and conflict (averaged; β = .84, R ² = .71; this was the constrained parameter used to identify the regression model). The model predicting ambivalent SIB had good absolute fit: χ²(7) = 5.69, p = .577, CFI = 1.00, RMSEA = 0.00, SRMR = 0.04. In all, family functioning explains 33% of the variance in ambivalent SIB. See Figure 3 for the standardized path coefficients between family functioning variables, the latent variable of family functioning, and ambivalent SIB. Notably, family functioning predicted ambivalent SIB: β = .58, B = 1.04, SE B = 0.46, p = .025. Based on effect sizes (see Figure 3), the strongest predictors were problem-solving (averaged; β = .94, B = 0.15, SE B = 0.07, p = .022, R ² = .89), communication (adolescent-reported; β = .83, B = 0.11, SE B = 0.05, p = .030, R ² = .69), and affective responsiveness (averaged; β = .69, B = 0.13, SE B = 0.07, p = .049, R ² = .47).