TPC Journal V8, Issue 2 - FULL ISSUE

166 The Professional Counselor | Volume 8, Issue 2 Autocorrelation. Autocorrelation was measured and evaluated at the .05 significance level. There was significant autocorrelation for Rose for the Academic Competence factor ( p = 0) in the treatment phase. There was no significant correlation for Janelle. There was significant autocorrelation in several areas for Kara: college knowledge ( p = 0.003), positive personal characteristics (p = 0.001), and academic competence ( p = 0) in the treatment phase. No transformationswere applied to correct for autocorrelation because of lack of independence between data points, the small sample size, and the significant trends in some of the phases; therefore, non-parametric data analyses were used. Regression. Regression was measured for each participant, factor, and phase to determine if there is a trend in each phase of the study. All three participants exhibited unique trend patterns for each of the factors. Rose exhibited a significant trend for academic competence in the treatment phase. The strong positive slope (R 2 = 0.7399, Slope = 0.16084, p = .000332) suggested a steady increase during the treatment phase. Janelle exhibited negative treatment phase trends for positive personal characteristics (R 2 = 0.3392, Slope = -1.094, p = 0.049), academic competence (R 2 = 0.411, Slope = -0.9650, p = 0.0247), and potential to achieve future goals (R 2 = 0.411, Slope = -0.6434, p = 0.0247). The negative slopes suggest a decrease in self- efficacy across all factors except college knowledge. Lastly, Kara exhibited significant positive trends for college knowledge (R 2 = 0.7142, Slope = 0.5594, p = 0.000538), positive personal characteristics (R 2 = 0.6138, Slope = 0.22378, p = 0.00257), and academic competence (R 2 = 0.6823, Slope = 0.24825, p = 0.00093) in the treatment phase, suggesting a steady increase in these factors. The overall findings indicated that further parametric data analyses (e.g., ANOVAs) would not be appropriate because of the significant trends in various factors. RCDC. The autocorrelation indications, regression trends, and additional complexity of outlier scores indicated that the RCDC (Borckardt, 2008), a robust non-parametric method, should be used rather than the Conservative Dual-Criteria method (Fisher, Kelley, & Lomas, 2003; Swoboda, Kratochwill, & Levin, 2010) and parametric methods such as student’s t -test and ANOVA. The RCDC significance threshold is based on the mean and regression lines and the number of comparisons in the comparison phase. Datum that fall above or below the desired zone, as determined by the mean and regression lines, are considered significant. The sign of the slope determines the direction of the difference. For Rose, there were significant increases in the academic competence scores in the treatment phase. Enhancing her academic competence was one of the customized goals set at the beginning of the treatment phase. For Janelle, college knowledge and academic competence scores improved significantly in the treatment phase. These were the two customized goal categories for Janelle. For Kara, positive personal characteristics and academic competence scores improved significantly. Enhancing academic competence was one of the customizing goals set for Kara. Visual Analyses The graphic data are presented in Figure 1. The baseline, treatment, and withdrawal phase CCRSI factor scores for each participant are presented visually. The visual analysis confirmed the findings of the RCDC analyses. Effect Sizes Cohen’s (1988) G-index effect size findings varied across the three participants, indicating that the interventions had differential treatment effects. For Rose, there was a large effect size (1.00) for academic competence from baseline to end of treatment, with a medium negative effect size from end