Poster #2 - Assessment of Using Equal Cluster Size Assumption to Calculate Sampling Variance in Meta-Analysis

• In Event: 47.099 - MTCC Poster Session 11
  In Poster Session: 47.099-1 - Division D Section 2 Statistical Theory and Quantitative Methodologies

Abstract

The standardized mean difference is commonly used as the effect size quantifying treatment effects on continuous outcomes. Many meta-analysts use inverse-variance weighting to pool effect sizes that needs each cluster size information. While most primary studies reported total sample size, some primary studies did not report cluster sizes separately resulting those studies were excluded. Therefore, we conducted real data analysis and simulation study evaluating performance of alternative approach (equal cluster size assumption) under RVE and three-level model. Results indicate that use of list-wise deletion showed low power whereas use of equal sample size assumption with RVE provided accurate and unbiased results even when true cluster sizes were unbalanced. Results are discussed and guidelines for applied meta-analysts are provided.

Author

• Sunyoung Park, California Lutheran University