TPC Journal-Vol 10- Issue 3-FULL ISSUE

384 The Professional Counselor | Volume 10, Issue 3 the error variances to correlate. In all but one case, the suggested correlated item error variances were adjacent to each other on the survey (i.e., item 2 with 3, 6 with 7, 11 with 12, 17 with 18, 20 with 21, 22 with 23, and 27 with 28). Given the proximity of the items, it was plausible that some systematic error variance between items would correlate. The only pair of items that were not adjacent were item 5 and item 19. Both of these items referred to the school’s teacher assistance team. For all the models, the path coefficients were statistically significant ( p < .001). Results of the Satorra-Bentler scaled chi-square difference test suggested that the four-factor model was a better model than the one-factor model ( p < .001), and there was no statistically significant difference between the four-factor model and the second-order four-factor model. Because of the high intercorrelations among the factors (ranging from .81 to .92), the second-order four-factor model was tested using the second subset from the sample. Confirmatory Phase The holdout sample of 2,033 participants was used to verify the second-order four-factor model. The initial model (see bottom of Table 3) with no modifications suggested the model marginally fit the data (chi-square = 5,424.82, RMSEA = .066, CFI = .88, SRMR = .058). After modifying the model by allowing for the eight correlated error variances, which were the same eight correlated error variances identified in the exploratory stage, as expected there was an improvement in the model fit (chi-square = 4,468.62, RMSEA = .060, CFI = .90, SRMR = .051). The observed item loading coefficients and standard errors are reported in Table 4. All coefficients are statistically significant and all above 0.50, suggesting stable item alignment to the factor being measured. Coefficient alpha values were 0.95 for total score with all items, 0.88 for Factor 1 (Individualized Supports and Practices), 0.86 for Factor 2 (Schoolwide Supports and Practices), 0.78 for Factor 3 (Targeted Supports and Practices), and 0.65 for Factor 4 (Collaborative Supports and Practices). The results provide evidence that the SCKSS has potential to provide inferences about counselors’ knowledge and skills for MTSS. Discussion and Implications The SCKSS was based on the TKSS, which measured teachers’ knowledge and skills related to PBS. After adapting the survey to align to MTSS and the role of school counselors, this study aimed to examine the latent structure of the SCKSS for examining MTSS knowledge and skills. Using both exploratory and confirmatory factor analyses, results suggest that a second-order four-factor model had the best fit. The findings indicate that the SCKSS has high internal consistency with Cronbach’s alpha for the total score at 0.95, and a range between 0.65 and 0.88 for each of the four factors. The first factor, Individualized Supports and Practices, contains 14 items; the second factor, Schoolwide Supports and Practices, contains 10 items; the third factor, Targeted Supports and Practices, contains six items; and the fourth factor, Collaborative Supports and Practices, is composed of three items. These findings confirm that the SCKSS yields valid and reliable inferences about school counselors’ knowledge and skills for MTSS. Previous measures that were specific to school counselors focused on confidence and beliefs in implementing response to intervention (RtI; Ockerman et al., 2015; Patrikakou et al., 2016). Although these studies contribute to the literature by aligning RtI with the ASCA National Model, they did not focus on the specific knowledge and skills related to MTSS.