TPC-Journal-V2-Issue1

72 The Professional Counselor \Volume 2, Issue 1 for that purpose. We used it instead to measure agreement after accounting for chance between two series of nominal data rated jointly by consensus. Kappa values have a range 0 to 1 and are interpreted like positive correlation coefficients. Null 5 is rejected on the basis of a Spearman r = .51, p = .005 (1-tail) and κ = .44, approximate SE = .19, approximate p = .013. Discussion Measures This study is based on a mixed-methods design with two measures for data collection. The first is the PTGI, a self- report instrument developed from standard psychometric techniques. Posttraumatic researchers use the PTGI to measure personal growth attributed to a trauma of interest to researchers. We use the PTGI as described by the instrument developers (Tedeschi & Calhoun, 1996) to measure total personal growth attributed to the sum of experiences during a time period of interest. Our second measure is an open table for obtaining participant attributions of cause for total personal growth. Participants complete the table with written descriptions of personal experiences and estimates of corresponding percentage contributions. The table design is adapted from one introduced in a previous study (Anderson & Lopez-Baez, 2011). Validity The purposes of the current study require a degree of veridicality (truthfulness), a form of internal validity. Researchers have reported little evidence of any kind for the validity of subscale scores beyond the results of exploratory factor analysis (see review in Anderson & Lopez-Baez, 2008). This is perhaps the reason why researchers have drawn few conclusions about growth from subscale scores in their results. Our comparisons one to three in the current study demonstrate the existence of predicted relationships among total PTGI scores of three samples and therefore a degree of internal validity for the collection of individual items and total PTGI scores. Comparison four demonstrates the existence of a predicted relationship between the attributions of cause from two studies and therefore a degree of internal validity for data obtained with the table-based approach. Comparison five demonstrates a degree of concurrent validity for both the subscale of Spiritual Change and the table-based approach. Posttraumatic Growth Inventory as a Measure of Personal Growth The PTGI was developed as a measure of posttraumatic growth. Tedeschi and Calhoun (2004) described their instrument as reflecting the core of personal growth. The results of the current study offer strong support for Tedeschi and Calhoun’s description because the results report high levels of personal growth for students without regard to prior experiences of trauma. We believe that if the PTGI can be used to measure the personal growth of samples of populations as dissimilar as college students and trauma survivors, then the PTGI can probably be used as a measure of personal growth under other circumstances in future studies. College-Student Growth and Posttraumatic Growth Magnitude. Theorists have identified the college-student undergraduate years as a time of personal growth in response to both academic and non-academic experiences (Chickering, 1969; Chickering & Reisser, 1993; Perry, 1970) in terms that suggest the definition of personal growth embodied in the items of the PTGI. Researchers have generally confirmed theorist predictions of student growth (c.f., Hassan, 2008; Higgins et al., 2009), but not with measures that allow for comparison of personal growth in response to trauma. Therefore, we are not surprised that the results of the current study reflect student growth. We are surprised, however, by the magnitude of that growth, m = 69.09 ( SD = 11.52) because it is so near the maximum of the range reported in previous studies of posttraumatic growth (Linley & Joseph, 2004). We believe that this comparison helps readers appreciate the magnitude of growth described by each population. Factor structure and subscales. The developers of the PTGI reported a 5-factor structure for the items of their instrument on the basis of an exploratory factor analysis (Tedeschi & Calhoun, 1996) and developed five corresponding subscales of unequal length. Subsequent posttraumatic studies have reported 2- and 3-factor structures (see review in Anderson & Lopez-Baez, 2008). Taken together, we interpret these EFA results as evidence that the factor structure of posttraumatic growth is not highly differentiated or stable across different samples. The results of several EFA described in our first study (Anderson & Lopez-Baez, 2008) demonstrated that the factor structure of a sample of student scores also lacked differentiation and stability, and therefore resembled that of posttraumatic growth.