The Professional Counselor | Volume 10, Issue 4 477 male. Four student participants were in the 25 to 30-year-old range, and four were in the 31 to 35-yearold range. The instructor was in the 50 to 55-year-old range, who identified as a White cisgender male. None of the student participants reported having previous teaching experience. Data Collection After obtaining IRB permission, the first author collected the initial consent, demographic, Q-sort, and post–Q-sort written data from the students and instructor using a semi-structured questionnaire. The nine participants (n = 8 students; n = 1 instructor) were each asked to rank-order the 48 items in the Q sample along a forced choice grid frommost agree (+4) to most disagree (-4). The conditions of instruction used for the students’ and instructor’s Q-sorts stemmed directly from the research question. After completing this Q-sort, participants were asked by the first author to provide written responses, using a semi-structured questionnaire, for the top three items with which they most (+4) and least (-4) agreed and were asked to comment on any other items of significance. The first author asked the course instructor to respond in writing to three questions, in addition to those prompts contained in the semi-structured questionnaire. This was done to add nuance and context to the results. The additional questions and highlights from the instructor’s responses are shown in Table 2. Data Analysis Nine Q-sorts completed by participants were each entered into the PQMethod software program V. 2.35 (Schmolck, 2014). A correlation matrix was then generated reflecting the “nature and extent of relationships” among all the participants’ Q-sorts in the data set (Watts & Stenner, 2012, p. 111). The correlation matrix served as the basis for factor analysis, which was completed using the centroid method (Brown, 1980). Essentially, factor analysis allows researchers to examine the correlation matrix for patterns of similarity among the participants’ Q-sorts. In the current study, we were interested in similar and divergent patterns among the instructor’s and students’ Q-sorts on essential doctoral CETI course components. In other words, data analysis in Q studies is possible because all participants rank-order a Q sample of similar items, which allows researchers to inter-correlate those Q-sorts for subsequent factor analysis. Given the low number of participants, we initially extracted five factors from the correlation matrix, which yielded fewer significant factor loadings (i.e., a correlation coefficient reflecting the degree to which a participant’s Q-sort correlates with the factor). Therefore, we extracted three factors, which yielded a higher number of factor loadings. The three factors were rotated using the varimax method, which we selected because (a) we had no preconceived theoretical notions regarding the findings, (b) we were blind to participant identifying information in the data, and (c) we intended to obtain dominant views among participants within the same course (Watts & Stenner, 2012). The varimax factor rotation method helps researchers to identify individual factor loadings “whose positions closely approximate those of the factor” (Watts & Stenner, 2012, p. 142). In Q methodology, a factor is a composite or ideal Q-sort to which individual participants correlate (Watts & Stenner, 2012). Overall, data analysis steps yielded a 3-factor solution containing at least two significant factor loadings on each factor, which is the minimum suggested number of factor loadings for a factor to hold significance (Brown, 1980). Notably, the final 3-factor solution contained significant factor loadings for all nine of the study participants, which suggests the rigor of the collective viewpoints (i.e., factors) discussed in the results.
RkJQdWJsaXNoZXIy NDU5MTM1