Modeling Intervention Effects on Mean and Intra-individual Variability: A Pretest–Posttest Mixed-Effects Location–Scale Framework

Summarise with VIRAT
Authors: YuelingHe, Yue Liu
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Randomized controlled trials with pretest–posttest repeated measurements (RCT-RM), in which identical assessments are administered repeatedly before and after an intervention, provide abundant within-person information for evaluating intervention effects. However, commonly used analytic approaches primarily focus on mean-level changes and typically assume homogeneous residual variances, limiting their ability to capture heterogeneity in intra-individual variability (IIV), and to evaluate intervention effects reflected in variability-level changes. To address these limitations, the present study extends the mixed-effects location–scale model (MELSM) to RCT-RM designs and introduces a pretest–posttest parameterization (PP-MELSM). The proposed framework jointly models intervention effects on both mean outcomes and IIV within a unified multilevel structure, explicitly accounts for heterogeneous residual variances, and allows individual-level covariates to moderate intervention effects on both components. A motivating empirical example illustrates how PP-MELSM provides additional insights beyond conventional mean-based analyses. Two Monte Carlo simulation studies systematically evaluated statistical power and estimation accuracy across a wide range of design conditions. Results showed that intervention effect size, sample size at both levels, intraclass correlation, and variance of random effects in the pretest scale model substantially influence statistical power and estimation accuracy. Detecting intervention effects on IIV and moderation effects consistently required larger Level 2 sample sizes than detecting mean-level effects. Overall, PP-MELSM provides a flexible methodological framework for modeling intervention effects on both mean outcomes and IIV in RCT-RM designs and offers simulation-based guidance for sample size planning in intervention studies.

Fields of Study & Subjects

Statistical Methods and Bayesian Inference Statistical Methods in Clinical Trials Meta-analysis and systematic reviews

Publication details

Journal / source
OSF Preprints (OSF Preprints)
Publication date
19 March 2026
Language
EN
Document type
Preprint