A meta-analysis of the relation of trauma and somatization: an investigation of methodological factors related to effect size
The existing literature suggests that there is a significant and positive relation of trauma and somatization. However, the existing literature seems to give little consideration to methodological, participant, and study characteristics that might affect estimates of the magnitude of this relation. The present study is a meta-analysis of the existing literature on trauma and somatization. The present study intended to estimate an overall effect size for the relation of trauma and somatization, to empirically describe the methods used to study the relation of trauma and somatization, and to determine if methodological differences between studies contribute to differences in estimates of the strength of the relation of trauma and somatization. In a sample of 244 studies, there was a positive and significant mean weighted correlation between trauma and somatization (r = .161, p < .05). However, study differences often had a small but significant influence on study results and effect sizes. The existing literature was overwhelmingly retrospective, and both trauma and somatization were most frequently assessed by self-report alone. Much of the existing literature employed methods that allow for alternative explanations of the apparent relation of trauma and somatization – e.g., overestimation due to poor recall, a response bias, or poor operationalization of trauma and/or somatization. In general, less stringent methods and assessments of trauma and somatization were associated with larger effect sizes, whereas more stringent methods and assessments were associated with smaller effect sizes. Future studies of the relation of trauma and somatization should consider that methodological differences between studies may affect estimates of this relation.