Power Approximations for Meta-Analysis of Dependent Effect Sizes

Meta-analytic models for dependent effect sizes have grown increasingly sophisticated over the last few decades, which has created challenges for a priori power calculations. We introduce power approximations for tests of average effect sizes based upon the most common models for handling dependent effect sizes. In a Monte Carlo simulation, we show that the new power formulas can accurately approximate the true power of common meta-analytic models for dependent effect sizes. Lastly, we investigate the Type I error rate and power for several common models, finding that tests using robust variance estimation provide better Type I error calibration than tests with model-based variance estimation. We consider implications for practice with respect to selecting a working model and an inferential approach.

A Meta-Analysis of Longitudinal Peer Influence Effects in Childhood and Adolescence

For decades, psychological research has examined the extent to which children’s and adolescents’ behavior is influenced by the behavior of their peers (i.e., peer influence effects). This review provides a comprehensive synthesis and meta-analysis of this vast field of psychological science, with a goal to quantify the magnitude of peer influence effects across a broad array of behaviors (externalizing, internalizing, academic). To provide a rigorous test of peer influence effects, only studies that employed longitudinal designs, controlled for youths’ baseline behaviors, and used “external informants” (peers’ own reports or other external reporters) were included. These criteria yielded a total of 233 effect sizes from 60 independent studies across four different continents. A multilevel meta-analytic approach, allowing the inclusion of multiple dependent effect sizes from the same study, was used to estimate an average cross-lagged regression coefficient, indicating the extent to which peers’ behavior predicted changes in youths’ own behavior over time. Results revealed a peer influence effect that was small in magnitude (β ̅ = 0.08) but significant and robust. Peer influence effects did not vary as a function of the behavioral outcome, age, or peer relationship type (one close friend vs. multiple friends). Time lag and peer context emerged as significant moderators, suggesting stronger peer influence effects over shorter time periods, and when the assessment of peer relationships was not limited to the classroom context. Results provide the most thorough and comprehensive synthesis of childhood and adolescent peer influence to date, indicating that peer influence occurs similarly across a broad range of behaviors and attitudes.

Revisiting the effectiveness of study abroad language programs: A multi-level meta-analysis

This study set out to re-examine the effectiveness of study abroad programs in second language (L2) acquisition through a multi-level meta-analysis. Overall, 42 primary studies published between 1995 and 2019 were identified, and in total 283 effect sizes were meta-analysed. This study implemented a three-level random effects model to account for the clustered, mutually dependent effect sizes often nested in the primary studies of L2 study abroad research. The results indicated a medium-to-large effect ( g = 0.87) on study abroad language programs. Essentially, the featured moderators in general explained more heterogeneity variances at level 3 (i.e. the between-study level) than at level 2 (i.e. the within study level). For study abroad language learners, language acquisition is optimal when learners, in particular those of a lower proficiency level, take both formal and content-based language courses while living with host families. Learners’ age and pre-program training may not moderate the effectiveness of study abroad language programs. Importantly, this study further established that the length of study abroad programs are positively associated with learners’ language gains, but that an extended and prolonged domestic program does not necessarily lead to such gains. Research and pedagogical implications are further discussed based on the research findings.

The Efficacy of Psychological Intervention on Body Image in Breast Cancer Patients and Survivors: A Systematic-Review and Meta-Analysis

The experience of breast cancer and related treatments has notable effects on women's mental health. Among them, the subjective perception of the body or body image (BI) is altered. Such alterations deserve to be properly treated because they augment the risk for depression and mood disorders, and impair intimate relationships. A number of studies revealed that focused psychological interventions are effective in reducing BI issues related to breast cancer. However, findings are inconsistent regarding the dimension of such effects. This meta-analysis synthesizes and quantifies the efficacy of psychological interventions for BI in breast cancer patients and survivors. Additionally, since sexual functioning emerged as a relevant aspect in the BI distortions, we explored the efficacy of psychological interventions on sexual functioning related to BI in breast cancer patients and survivors. The literature search for relevant contributions was carried out in March 2020 through the following electronic databases: Scopus, PsycINFO, and ProQUEST. Only articles available in English and that featured psychological interventions for body image in breast cancer patients or survivors with controls were included. Seven articles with 17 dependent effect sizes were selected for this meta-analysis. Variables were grouped into: Body Image (six studies, nine dependent effect sizes) and Sexual Functioning Related to the Body Image in breast cancer patients and survivors (four studies, eight dependent effect sizes). The three-level meta-analysis showed a statistically significant effect for Body Image [g = 0.50; 95% CI (0.08; 0.93); p < 0.05] but no significant results for Sexual Functioning Related to Body Image [g = 0.33; 95% CI (−0.20; 0.85); p = 0.19]. These results suggest that psychological interventions are effective in reducing body image issues but not in reducing sexual functioning issues related to body image in breast cancer patients and survivors. Future review efforts may include gray literature and qualitative studies to better understand body image and sexual functioning issues in breast cancer patients. Also, high-quality studies are needed to inform future meta-analyses.

Evaluating Meta-Analytic Methods to Detect Selective Reporting in the Presence of Dependent Effect Sizes

Selective reporting of results based on their statistical significance threatens the validity of meta-analytic findings. A variety of techniques for detecting selective reporting, publication bias, or small-study effects are available and are routinely used in research syntheses. Most such techniques are univariate, in that they assume that each study contributes a single, independent effect size estimate to the meta-analysis. In practice, however, studies often contribute multiple, statistically dependent effect size estimates, such as for multiple measures of a common outcome construct. Many methods are available for meta-analyzing dependent effect sizes, but methods for investigating selective reporting while also handling effect size dependencies require further investigation. Using Monte Carlo simulations, we evaluate three available univariate tests for small-study effects or selective reporting, including the Trim & Fill test, Egger's regression test, and a likelihood ratio test from a three-parameter selection model (3PSM), when dependence is ignored or handled using ad hoc techniques. We also examine two variants of Egger’s regression test that incorporate robust variance estimation (RVE) or multi-level meta-analysis (MLMA) to handle dependence. Simulation results demonstrate that ignoring dependence inflates Type I error rates for all univariate tests. Variants of Egger's regression maintain Type I error rates when dependent effect sizes are sampled or handled using RVE or MLMA. The 3PSM likelihood ratio test does not fully control Type I error rates. With the exception of the 3PSM, all methods have limited power to detect selection bias except under strong selection for statistically significant effects.