TPC Journal V8, Issue 2 - FULL ISSUE

The Professional Counselor | Volume 8, Issue 2 179 Instrumentation and Procedures The authors followed the instrument development guidelines discussed by experts in psychometrics and questionnaire design (DeVellis, 2016; Fowler, 2014). An initial set of 18 items was created on a Likert-type scale, ranging from 1 ( strongly disagree ) to 5 ( strongly agree ). As discussed above, the original theoretical framework of SDT (Ryan & Deci, 2000) and its contemporary extensions (Adams et al., 2017) guided the development of item content. Item content was also derived from major themes identified in the literature review (comfort, stigma, referrals, prevalence, and complexity), particularly those related to student and faculty members’ connection to college student mental health support (Bishop, 2016; Eisenberg et al., 2011; Lee et al., 2017). The following CMHPCS items, for example, reflect SDT (the positive association between one’s sense of competency and action) and the research findings that one’s sense of comfort with mental health disorders is associated with increased referrals to resources for mental health disorders: “I am comfortable talking to students about mental health”; “I am comfortable referring college students with mental health issues to the health center on campus”; “I am aware of the university resources for mental health”; and “Mental health issues are increasing among college students.” Negatively worded items were recoded so that higher scores would indicate higher perceived competence. To obtain background information on the respondents, 11 demographic items were added to the questionnaire. These were developed in light of previous college counseling research that showed group differences (e.g., gender, ethnicity, previous attendance in counseling) on various mental health–related variables (Eisenberg et al., 2016; Kalkbrenner & Hernández, 2017). Sample items included the following: (1) Please select your gender; (2) Please specify your age (in years); and (3) Indicate your ethnic identity. The initial item pool was subjected to expert review and pilot testing to establish content validity. The items were sent to three expert reviewers with advanced training in clinical psychology, mental health counseling, and psychometrics. Their recommendations informed slight modifications to 15 items, improving their clarity and readability. A few additional items and formatting revisions were made based on pertinent feedback from pilot study participants (22 graduate students). For example, we clarified the meaning of “referred another student to counseling services” to “referred (recommended) that another student seek counseling services.” Statistical Analyses A series of statistical analyses were computed to answer the research questions, including exploratory factor analysis (EFA), confirmatory factor analysis (CFA), and hierarchical logistic regression (HLR). During phase 1 of the study using the faculty sample, a principal factor analysis (PFA) was conducted to determine the underlying latent factor structure of the CMHPCS (Mvududu & Sink, 2013). Given that the constructs related to SDT are generally correlated (Adams et al., 2017), the researchers used an oblique rotation (direct oblimin, ∆ = 0). The Kaiser criterion (eigenvalues [Λ] > 1), meaningful variance accounted for by each factor (≥ 5%), a review of the scree plot, and parallel analysis results guided the factor extraction process. Factor retention criteria were used based on the recommendations of Mvududu and Sink (2013): factor loadings > .40, commonalities ( h 2 ) > .30, and cross-loadings < .30. The content of items that loaded on each factor were reviewed for redundancy, as it is an accepted practice to remove an item that is highly correlated and conceptually similar to at least one other item (Byrne, 2016). To cross-validate these initial factor analytic results, a CFA using a maximum likelihood estimation method was conducted to test the validity of the factor solution that emerged in the EFA with a sample of undergraduate students (research question 2). Using the recommendations of Byrne (2016), the