TPC-Journal-V6-Issue2

The Professional Counselor /Volume 6, Issue 2 178 The condition of instruction for this study was, “Sort the following counselor advocacy behaviors according to how important or unimportant they are to your career counseling work.” The two poles of the distribution were most important and most unimportant . Poles range from most to most so that the ends of the distribution represent the areas that hold the greatest degree of psychological significance to the participant, and the middle of the distribution represents items that hold relatively little meaning or are more neutral in importance (Watts & Stenner, 2012). The Q sorts for this study were conducted both in person and via phone or video chat (i.e., Google Hangouts, Skype). Once informed consent was obtained, I facilitated the Q sorting procedure by reading the condition of instruction, observing the sorting process, and conducting the post-sort interview. Once each participant felt satisfied with his or her sort, the distribution of statements was recorded onto a response sheet for later data entry. Post-Sort Interview Immediately following the Q sort, I conducted a semistructured interview with each participant in order to gain a greater understanding of the meaning of the items and their placement, as well as his or her broader understanding of the topic at hand (Watts & Stenner, 2012). The information gathered during the interview is used when interpreting the final emergent factors. Items in the middle of the distribution are not neglected and are specifically asked about during the post-sort interview so that the researcher can gain an understanding of the entire Q sort for each participant. Although the interview data are crucial to a complete and rigorous factor interpretation and should be conducted with every participant in every Q study, the data analysis process is guided by the quantitative criteria for factor analysis and factor extraction. The qualitative interview data, as well as the demographic data, are meant to help the researcher better understand the results of the quantitative analysis. Data Analysis Data were entered into the PQMethod program (Schmolck, 2014) and Pearson product moment correlations were calculated for each set of Q sorts. Inspection of the correlation matrix revealed that all sorts (i.e., all participants) were positively correlated with one another, some of them significantly so. This indicated a high degree of consensus among the participants regarding the role of advocacy in career counseling, which was further explored through factor analysis. I used centroid factor analysis and Watts and Stenner’s (2012) recommendation of beginning by extracting one factor for every six Q sorts. Centroid factor analysis is the method of choice among Q methodologists because it allows for a fuller exploration of the data than a principal components analysis (McKeown & Thomas, 2013; Watts & Stenner, 2012). Next, I calculated the significance level at p < .01, which was .516 for this 25-item Q sample. The unrotated factor matrix revealed two factors with Eigenvalues near or above the commonly accepted cutoff of 1 according to the Kaiser-Guttman rule (Kaiser, 1970). Brown (1978) argued that although Eigenvalues often indicate factor strength or importance, they should not solely guide factor extraction in Q methodology since “the significance of Q factors is not defined objectively (i.e., statistically), but theoretically in terms of the social-psychological situation to which the emergent factors are functionally related” (p. 118). Since there currently is little empirical evidence of differing perspectives on advocacy among career counselors, two factors were retained for rotation. In order to gain another perspective on the data, I used the Varimax procedure. I flagged those sorts that loaded significantly (i.e., at or above 0.516) onto only one factor after rotation. Four