A meta-analytic review of factors that influence the effectiveness of training in organizations

Abstract

This dissertation examined the training literature focusing on potential factors that could influence the effectiveness of training interventions, and used meta-analysis to quantify the impact of these factors. A total of 177 articles and 466 independent data points were used in this study. Across all 466 data points, training was found to be more effective than expected. Using meta-analytic procedures, a "population" estimate of the effectiveness of training (b = .75 1) was calculated. In addition, there was sufficient evidence (SD =.540) to support a search for hypothesized moderators of training effectiveness. The hypothesized moderators included the implementation quality of the program, the criteria used to evaluate the training, the methods used for training, and the skill and task characteristics to be trained. In addition, results from this study were expected to contribute to the resolution of issues associated with the impact of aptitude-treatment interactions and the existence of positive-findings bias associated with the methodological rigor or quality of the evaluation study. Implementation quality was found to be a significant moderator of training effectiveness. Although the majority of studies (93 %) did not report any needs assessment activities, those that did report these activities were found to be markedly more effective than those that did not. Results related to the criteria used to evaluate training were not as expected. The effect size for training did not systematically vary as a function of the "level" of criteria and the observed effect size for results criteria was substantially larger than expected. Different training methods were found to be effective for different skills and tasks, and therefore functioned as moderators of training effectiveness. Quantitative indicators of the relative overall effectiveness of a variety of training methods, the overall effectiveness of training for the various skills and tasks to be trained, and the relative effectiveness of the methods for training specific skills and tasks were calculated. Finally, no empirical evidence supporting the existence of aptitude treatment interactions or the presence of positive-findings bias, in studies of training effectiveness was found.