Concealed correlations meta-analysis: A new method for synthesizing standardized regression coefficients.

摘要: The synthesis of standardized regression coefficients is still a controversial issue in the field meta-analysis. difficulty lies fact that belonging to models include different sets covariates do not represent same parameter, and thus their direct combination meaningless. In present study, new approach called concealed correlations meta-analysis proposed allows for using common information from contain improve precision combined focal coefficient estimate. performance this was compared with two other approaches: (1) carrying out separate meta-analyses studies used model, (2) performing meta-regression on while including an indicator variable as moderator indicating model which each belongs. comparison done through simulation study. results showed that, expected, led more accurate estimates under both random- fixed-effect models.