Experimental stress induction in children and adolescents with the Trier Social Stress Test (TSST): A systematic review and meta-analysis

Abstract

The Trier Social Stress Test (TSST) and its variations for children and adolescents are among the most frequently used laboratory stressors. A previous meta-analysis revealed moderate but robust psychobiological stress responses in children and adolescents. Here, we focused on non-clinical samples and performed both a systematic review as well as a meta-analysis for the following: (a) to provide an overview of study characteristics in children and adolescents, (b) to update the overview of the reported effect sizes, and (c) to identify established and new potential moderators. Reviewing 143 studies, we identified three major variants in practical use: the TSST-C, the TSST-M, and the original TSST. Although the studies were inconsistent in reporting methodological characteristics, the variants did not differ in exclusion criteria, main procedures (administration and preparation time, duration), and composition of the panel. Based on the 92 studies (N = 8291 participants) included in the meta-analysis, we identified an overall publication bias-corrected effect on salivary cortisol stress responses of Hedge’s g = 0.56, p < .0001. Moderation analysis revealed the proportion of girls, number of judges, and total duration of the stress test as significant moderators, all positive. Explorative analysis showed that higher baseline cortisol levels predicted lower baseline-peak reactivity. Overall, variations of the TSST for children and adolescents are suitable for inducing acute psychobiological stress responses in these samples. To promote open and reproducible (meta) science, raising the reporting standards for methods and results in future studies is desirable.