Exploring Consequences of Simulation Design for Apparent Performance of Statistical Methods. 2: Results from simulations with normally and uniformly distributed sample sizes

TL;DR

This paper examines how different simulation design choices, specifically the distribution of sample sizes, influence the apparent performance of statistical methods in meta-analysis of log-odds-ratios, extending prior work on constant sample sizes.

Contribution

It investigates the impact of normally and uniformly distributed sample sizes on simulation outcomes, providing new insights into simulation design effects in statistical method evaluation.

Findings

Sample size distribution affects method performance conclusions.

Results differ between normal and uniform sample size simulations.

Implications for designing simulation studies in meta-analysis.

Abstract

This report continues our investigation of effects a simulation design may have on the conclusions on performance of statistical methods. In the context of meta-analysis of log-odds-ratios, we consider five generation mechanisms for control probabilities and log-odds-ratios. Our first report (Kulinskaya et al. 2020) considered constant sample sizes. Here we report on the results for normally and uniformly distributed sample sizes.