TPC Journal V8, Issue 2 - FULL ISSUE

The Professional Counselor | Volume 8, Issue 2 151 the course section in which they participated. These identifiers were kept in an Excel file that was password protected and was kept away from the rest of the data in order to reduce bias. Data were analyzed in two phases to identify central themes associated with the participants’ experiences and perceptions with DSLP. First, data corresponding to each of the four foci of the research question were grouped into the following a priori categories: (a) understanding of course content, (b) understanding of human development in non-dominant groups, (c) perception of self, and (d) a social justice perspective in counseling. During the second phase of analysis, data within each category were coded by meaning units, which was defined as a collection of words, sentences, or paragraphs that referred to a discrete idea. Closely related codes were collapsed into themes. Researchers used NVivo 10 (QSR International, 2012) for the coding process and to calculate interrater reliability statistics. Trustworthiness During the study, the researchers engaged in several strategies to ensure the study’s trustworthiness. The research team consisted of the course instructor and a graduate student research assistant who was trained in the research procedures. Prior to the study’s design and again before data analysis, researchers examined their potential biases. As recommended by Rossman and Rallis (2003), researchers engaged in reflexivity through writing, discussing, and revising researcher-as-instrument statements throughout the process. This process was done to bracket the researchers’ beliefs and opinions to ensure that the participants’ voices could be heard fairly and clearly. Data were collected from documents that participants completed at two different points during the academic term (midterm and end of term), providing the basis of a longitudinal analysis. At the beginning of data analysis, researchers spent several hours coding data together to support shared meaning of codes and ensure credibility of the analysis. Additionally, researchers engaged in peer debriefing of codes and the coding process at weekly research meetings. Within each phase of coding, the researchers calculated interrater reliability statistics in NVivo 10 (QSR International, 2012) to determine the credibility of the analysis. After each coding session, researchers documented their reflections, questions, and ideas in a reflexive journal designed to document decision making related to the analysis. An audit trail was kept ensuring confirmability of the study’s findings. Interrater Reliability During each phase of coding, researchers conducted interrater reliability testing using NVivo 10 (QSR International, 2012) to ensure credibility of the coding process. In the first phase of grouping data into four a priori categories for further coding, an interrater reliability test resulted in a kappa coefficient of .68. This outcome is considered a “substantial” benchmark for kappa coefficients by Landis and Koch (1977). During the second phase of coding into emergent categories, the kappa coefficient for data that was coded by both researchers was .96. This is an “almost perfect” benchmark for kappa coefficients (Landis & Koch, 1977). These results demonstrated that raters consistently coded the data in a similar matter and increased the data’s credibility. Results The study’s results indicated the level of impact the DSLP experience had on participants’ understanding of course content, understanding of people in non-dominant groups, perceptions of themselves, and what social justice in a counseling context meant to them. For participants, the DSLP experience became a lens to look at the world in a different way and was a primary frame of reference for