Back to search

University of North Texas

Performance of Indirect Effect Tests for Longitudinal Mediation in Cross-Lagged Panel Models: A Methodological Review and Monte Carlo Simulation Study

Abstract

dc:description

Cross-lagged panel models (CLPMs) are frequently used in psychological and health research to examine longitudinal mediation, yet limited guidance exists regarding optimal methods for testing indirect effects. This Monte Carlo simulation study evaluated the performance of four commonly used indirect effect tests within the CLPM framework. Sample size (50, 75, 100, 200, 500), cross-lagged effect size (0, 0.03, 0.07, 0.12), and number of measurement occasions (3, 4, 5 waves) were manipulated across 60 design conditions, and performance was evaluated using Type I error rates, statistical power, convergence rates, and method agreement. All models achieved perfect convergence (100%). Type I error rates remained well-controlled across all methods (0% -1.73%), below the α = .05 level. The bias-corrected and accelerated (BCa) bootstrap demonstrated highest overall power (43.11%), followed closely by joint significance (42.47%), percentile bootstrap (36.80%), and Sobel test (33.78%). Power increased systematically with effect size and sample size, with the most substantial gains occurring between n = 100 and n = 200. For the small cross-lagged effects typical in psychological research (0.03-0.07), all methods showed limited power even with large samples, highlighting the challenge of detecting longitudinal mediation. Method disagreement was the highest (10%-15%) at the effect size of .07 and sample sizes of 75 and 100. The joint significance test demonstrated power comparable to BCa bootstrap while requiring 250-fold less computational time. These findings provide guidance for applied researchers designing longitudinal mediation studies, emphasizing the importance of adequate sample sizes (n ≥ 200) and realistic power planning based on meta-analytic effect size estimates rather than conventional benchmarks.

Degree

thesis:*
Grantor dc:publisher
University of North Texas
Year dc:date
2026

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Kulikova, Alexandra 1989-
Contributors dc:contributor
  • Middlemiss, Wendy
  • Uanhoro, James O.
  • Chen, Qi
  • Henson, Robin K.

Subjects

dc:subject × 3

Rights

dc:rights
Statement dc:rights
  • Public
  • Kulikova, Alexandra 1989-
  • Copyright
  • Copyright is held by the author, unless otherwise noted. All rights Reserved.
Language dc:language
English

Identifiers

dc:identifier.*
Identifier
local-cont-no: submission_4931
ark: ark:/67531/metadc2652441

Chain of custody

source
Harvested from
University of North Texas
Base URL
digital.library.unt.edu/oai/
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Kulikova, Alexandra 1989-. Performance of Indirect Effect Tests for Longitudinal Mediation in Cross-Lagged Panel Models: A Methodological Review and Monte Carlo Simulation Study. University of North Texas, 2026. https://digital.library.unt.edu/ark:/67531/metadc2652441/