Sample size for LMM using Design Effect

This fact sheet provides practical guidance on calculating sample size for linear mixed models using the Design Effect (Deff) approach, focusing on random-intercept models commonly used for repeated measures and clustered data. It explains how non-independence arising from shared context affects statistical power, and how the intraclass correlation coefficient (ICC) and number of observations per unit should be incorporated into planning decisions. The document distinguishes between within-unit (Level-1) and between-unit (Level-2) effects and outlines a clear two-step workflow for translating independence-based sample size calculations to multilevel designs. An accompanying spreadsheet-based calculator is provided to support transparent calculation, sensitivity analysis, and exploration of plausible design scenarios during the planning stage.