TPC-Journal-V2-Issue1

The Professional Counselor \Volume 2, Issue 1 71 two (nulls 2 and 3: sample 1 mean less than or equal to means of samples one and two, respectively). Nulls 2 and 3 are rejected on the basis of a 1-way ANOVA F (2, 593) = 20.02, p = .000, η 2 = .064 with an independent variable of 3 groups and dependent variable of student PTGI scores; and Bonferroni post-hoc t-tests of student PTGI scores: sample one and sample 1, t (467) = 6.44, p = 0.000 < α = .0167 (2-tail), d = .73; sample two and sample 1, t (237) = 4.70, p = 0.000 < α = .0167 (2-tail), d = .61; sample one and sample two, t (462) = .79, p = .430 > α = .0167 (2-tail), d = .08. As expected, the sample 1 mean of 69.09 is greater than the means of both sample one and sample two. Comparison three. Visual inspection of the subscale per-item mean scores of sample one (Anderson & Lopez- Baez, 2008) and sample two (Anderson & Lopez-Baez, 2011) found that four of the per-item means in each sample are approximately equal and that all four are greater than the corresponding mean Spiritual Change score by 1.50 to 2.00 scale divisions. Therefore, we expected the subscale scores of our sample 1 to exhibit the same pattern. Visual inspection (Table 1) finds that they do. Thus, as expected, the five subscale per-item mean scores of each of our three samples are characterized by four comparatively high and approximately equal subscale mean scores and a comparatively low mean Spiritual Change score. Comparison four. Members of all three samples completed the same 3-credit course (during different semesters) designed to facilitate personal growth. Members of sample two and sample 1 were asked for attributions of cause for the course (Anderson & Lopez-Baez, 2011). The resulting attributions of cause to a course designed to facilitate growth are themselves evidence of internal validity. We expected members of sample two to attribute a larger percentage of total growth for one semester to the class than members of sample 1 attributed for the longer period of several semesters (directional null 4: sample two mean attributions of cause to course less than or equal to corresponding mean of sample 1). Null 4 is rejected on the basis of an independent t (237) = 12.18, p = .000 < α = .05 (1-tail), d = 1.56. As expected, the sample two mean attribution of cause to the course of 25.28 (14.28)% is greater than the corresponding sample 1 mean attribution of 7.94 (6.46)%. Comparison five. Concurrent validity is the extent to which different measures of the same construct agree. We formulated Comparison five only after we saw an opportunity to investigate concurrent validity among the data of sample 1 by measuring the correspondence between high and low Spiritual Change scores in sample 1 and the presence or absence of corresponding written attributions of cause, respectively (null 5: no agreement). We defined high participant Spiritual Change scores as per-item mean scores of at least 3.50 scale divisions (at least 1 division above scale midpoint 2.50). Sample 1 contains 11 high scorers (2 students scored 5.00, 1 scored 4.50, 3 scored 4.00, 5 scored 3.50). We defined low Spiritual Change scores as per-item mean scores no larger than 1.50 scale divisions. Sample 1 contains 64 low scorers (14 students scored .00, 17 scored .50, 20 scored 1.00, and 13 scored 1.50). We added 19 questionnaires selected at random from those of the 64 low scorers to provide a sample of 30 questionnaires for analysis. We examined each sample member’s written attributions of cause for evidence of spiritual growth. We defined evidence of spiritual growth narrowly by the wording of the Spiritual Change items (PTGI items 5 and 18). Thus, we defined evidence of spiritual growth as student descriptions of one or more attributions of cause with at least one explicit reference to religion or spirituality including references to worship, God (or other deity or deities), prayer, or religious writings; but not references to ethics, morality, or meaning of life in the absence of the required explicit references. (Example attributions interpreted as evidence include: my faith deepened, came to accept God’s will, learned more about Bible; but not: grew in commitment to boyfriend, learned importance of moral behavior, decided on career choice, or obtained new understanding of life). We easily reached consensus on all identifications because of the simple and specific definition of evidence adopted beforehand. Attributions by 6 of the 11 high scorers and 2 of the 19 low scores contained at least one reference to spiritual growth. Half of the 8 attributions with references to spiritual growth referred to religious studies classes. High and low Spiritual Change scores corresponded to the presence or absence, respectively, of attributions of cause for 24 of 30 sample 1 members, percentage agreement = 80.00%. Spearman r is a measure of agreement between two series of nominal data that does not take into account the probability of chance matches. The probability of chance matches is high among data of interest because the sample is characterized by many low scores and many non-positive attributions of cause. Cohen’s κ (1960) is a measure of agreement between two series of nominal data that accounts for the probabilities of chance matches. Cohen’s κ was originally developed to assess inter-rater agreement and is widely used