Abstract
Stressful life events (SLEs) may elicit positive psychosocial change among youth, referred to as post-traumatic growth (PTG). We assessed types of SLEs experienced, degree to which participants reported PTG, and variables predicting PTG across 24 months among a sample of high-risk, ethnically diverse, early emerging adults. Participants were recruited from alternative high schools (n = 564; mean age = 16.8; 65% Hispanic). Multi-level regression models were constructed to examine the impact of environmental (SLE quantity, severity) and personal factors (hedonic ability, perceived stress, developmental stage, future time orientation) on a composite score of PTG. The majority of participants reported that positive changes resulted from their most life-altering SLE of the past two years. Predictors of PTG included fewer SLEs, less general stress, having a future time perspective, and greater identification with the developmental stage of emerging adulthood. Findings suggest intervention targets to foster positive adaptation among early emerging adults who experience frequent SLEs.
Notes
1. Analysis of variance showed that the method of data collection (by phone, on paper, or via home visit) did not significantly impact PTG scores (p = 0.09).
2. Spearman correlation coefficients on severity ratings from the five students ranged from r = 0.53–0.73, indicating acceptable-to-high agreement. However, because correlation coefficients are limited by not being able to take into account the agreement in ordering of items (vs. agreement in scoring), Bland-Altman plots were constructed (Bland & Altman, Citation1986) to ascertain the range of agreement in rater coding, with agreement defined as a mean bias ±2 standard deviations (SDs), and whether or not coding for each rater would be retained or additional raters would be needed. No additional raters were needed as 95% limits of agreement between each of the five raters ranged between two SDs of the mean differences comparing each rater to the other four.
3. To address concerns on using a composite scale of PTG, in which response options were scaled from negative to positive, we conducted a sensitivity analysis, in which the responses for PTG items were recoded as binary responses (1 = positive changes, 0 = no changes), then summed the items to produce a PTG score. The pattern of results was the same as presented, although hedonic ability was more strongly associated (at p < 0.05 vs. p < 0.10), while the number of stressful life events was more weakly associated (at p < 0.05 vs. p < 0.001) with PTG.