TPC Journal-Vol 9- Issue 2-Full-Issue

116 The Professional Counselor | Volume 9, Issue 2 because doing so might identify students as participants and cause participant bias. For example, a small number of students in the courses identified as male, African American, or Asian American, and if we asked these students to report their demographic information in the study, this information may have unintentionally identified the participants. We can report, though, that the control group participants included first-year school, clinical mental health, couples and family, and addictions counseling students. The experimental group participants included second-year school counseling and school psychology students. The average number of video lectures reportedly viewed by the experimental group participants was 7.4 (out of eight). Video lectures were not a part of the non- flipped course (control group). Instrumentation We distributed the Classroom Engagement Inventory (CEI; Wang et al., 2014) to participants to measure student classroom engagement because it comprehensively measures affective, behavioral, and cognitive engagement. Moreover, it can be used to measure engagement specific to the classroom level, rather than overall school or program engagement (Wang et al., 2014). Although Wang and colleagues (2014) developed the instrument with students in grades 4 through 12, they found that its factor structure was invariant when used with participants of different ages and grade levels, suggesting its relevance in higher education settings. The CEI consists of five subscales. They are: Affective Engagement (positive emotions students could encounter in class, ω = .90), Behavioral Engagement–Compliance (students’ compliance with classroom norms, ω = .82), Behavioral Engagement–Effortful Class Participation (students’ self-directed classroom behaviors, ω = .82), Cognitive Engagement (mental effort expended, ω = .88), and Disengagement (cognitive and behavioral aspects of not engaging in class, ω = .82; Wang et al., 2014). Example items are: “I get really involved in class activities” (Behavioral Engagement–Effortful Class Participation), “I feel excited” (Affective Engagement), and “I go back over things when I don’t understand” (Cognitive Engagement; Wang et al., 2014, p. 5). The instrument has 21 items and a 5-point frequency Likert-type scale ranging from never to hardly ever , monthly , weekly , and each day of class . We adapted the scale to be a 4-point scale by removing the answer choice each day of class because both courses only met once per week, therefore each day of class was synonymous with weekly. Data Analysis Using SPSS, we first analyzed internal consistency using Cronbach’s alpha to ensure that reducing the 5-point scale to a 4-point scale did not weaken reliability to an unacceptable degree. Then we ran independent samples t -tests to test for statistical significance at p < .05 in order to determine if experimental and control group scores differed by chance. We also ran Cohen’s d in SPSS to measure effect size, which quantifies the extent that the control group and experimental group diverged in the study (Thompson, 2006). We followed Cohen’s (1969) interpretation guidelines of small (0.2), medium (0.5), and large (0.8) effect sizes. We tested for significance among items grouped by scale, as well as overall measure of classroom engagement. Results The internal consistency for our results was deemed acceptable ( α = .85). We then compared classroom engagement for students in the flipped counseling courses to students in the non-flipped counseling courses in six ways. Table 4 contains a summary of each of these comparisons.