Using Multigroup-Multiphase Latent State-Trait Models to Study Treatment-Induced Changes in Intra-Individual State Variability: An Application to Smokers’ Affect

Abstract

Read online

Sometimes, researchers are interested in whether an intervention, experimental manipulation, or other treatment causes changes in intra-individual state variability. The authors show how multigroup-multiphase latent state-trait (MG-MP-LST) models can be used to examine treatment effects with regard to both mean differences and differences in state variability. The approach is illustrated based on a randomized controlled trial in which N = 338 smokers were randomly assigned to nicotine replacement therapy (NRT) versus placebo prior to quitting smoking. We found that post quitting, smokers in both the NRT and placebo group had significantly reduced intra-individual affect variability with respect to the affect items calm and content relative to the pre-quitting phase. This reduction in state variability did not differ between the NRT and placebo groups, indicating that quitting smoking may lead to a stabilization of individuals’ affect states regardless of whether or not individuals receive NRT.

Keywords

• experimental design
• multigroup confirmatory factor analysis
• latent state-trait models
• nicotine replacement therapy
• Ecological momentary assessment (EMA)
• smokers’ affect variability