In this study we investigated the influence of data nonnormality in the primary studies on meta-analysis of the standardized mean difference (SMD) for a two-independent-group design. The bias, mean squared error, and confidence interval coverage probability of the mean effect sizes under different types of population distributions were compared. Also, the performance of the Q test was examined. The results showed that oppositely skewed distributions (i.e., distributions skewed in different directions) showed poor performance for point and interval estimates of mean effect sizes in meta-analysis, especially when the tails were pointing toward each other. The previously found adverse impacts due to nonnormality in primary studies do not disappear when primary studies with nonnormal data are meta-analyzed, even when the average sample size and number of studies are large. The results also showed that, when the tails were pointing toward each other, the Type I error rates of the Q test were inflated. We suggest that the impact of violating the assumption of normality should not be ignored in meta-analysis.

More and more researchers use meta-analysis to conduct multivariate analysis to summarize previous findings. In the correlation-based meta-analytic structural equation modeling (cMASEM), the average sample correlation matrix is used to estimate the average population model. Using a simple mediation model, we illustrated that random effects covariation in population parameters can theoretically bias the path coefficient estimates and lead to nonnormal random effects distribution of the correlations. We developed an R function for researchers to examine by simulation the impact of random effects in other models. We then reanalyzed two real data sets and conducted a simulation study to examine the magnitude of the impact on realistic situations. Simulation results suggest parameter bias is typically negligible (less than .02), parameter bias and root mean square error do not differ across methods, 95% confident intervals are sometimes more accurate for the two-stage structural equation modeling approach with a diagonal random effects model, and power is sometimes higher for the traditional Viswesvaran-Ones approach. Given the increasing popularity of cMASEM in organizational research, these simulation results form the basis for us to make several recommendations on its application.

A heightened interest in online gaming has emerged during COVID-19, and people have become increasingly vulnerable to internet gaming disorder (IGD). However, playing video games can also have a positive effect; gaming has been recognized as an efficient coping strategy. Currently, relatively little is understood about how online gaming can turn from an efficient coping strategy into an addiction disorder. This study investigated the mediating roles of social cynicism, escape and coping motives on the association between daily disruption during COVID-19 and IGD, seeking to reveal the underlying mechanism that influences the effects of gaming. A total of 203 participants in Hong Kong who reported having played electronic games during COVID-19 were surveyed. We conducted three hierarchical multiple regressions, then tested a serial mediation model using path analysis with structural equation modeling. The results revealed that escape motives significantly mediated the relationship between daily disruption related to COVID-19 and IGD, but no such effect was found for coping motives. Social cynicism alone was not a significant mediator, but social cynicism and escape motives in series mediated the relationship between daily disruption and IGD. These difference outcomes suggested different underlying mechanisms of escape and coping motives.

Previous procedures for meta-analyzing dependent correlations have been found to overestimate or underestimate the true variation in effect sizes. Samplewise-adjusted procedures have been shown to perform better than simple within-study means when meta-analyzing dependent correlations. However, such procedures cannot be applied when correction for artifacts such as unreliability is desired. In the present study, we extended the procedures to correct for attenuation due to artifacts when meta-analyzing dependent correlations. Monte Carlo simulation was conducted in order to examine conditions with various degrees of dependence, degrees of heterogeneity, sample sizes, and numbers of studies, among other factors. The previous procedures, including the samplewise-adjusted procedures without correction, yielded biased point estimates and confidence intervals with low coverage probabilities of the population mean correlation and degree of heterogeneity. More importantly, the bias and undercoverage of the confidence interval increased with the mean sample size and number of studies in many conditions. The new samplewise-adjusted procedures with correction for attenuation yielded negligible biases when estimating the mean population correlation, even in the presence of dependent correlations. Given that the need for correction for attenuation due to artifacts is becoming more recognized in meta-analysis, our findings highlight the importance of such considerations when meta-analyzing dependent correlations. Conditions under which these procedures can be further improved are also discussed.

No abstract