Mixed‐effects models in psychophysiology

Abstract

The current methodological policy in Psychophysiology stipulates that repeated‐measures designs be analyzed using either multivariate analysis of variance (ANOVA) or repeated‐measures ANOVA with the Greenhouse–Geisser or Huynh–Feldt correction. Both techniques lead to appropriate type I error probabilities under general assumptions about the variance‐covariance matrix of the data. This report introduces mixed‐effects models as an alternative procedure for the analysis of repeated‐measures data in Psychophysiology . Mixed‐effects models have many advantages over the traditional methods: They handle missing data more effectively and are more efficient, parsimonious, and flexible. We described mixed‐effects modeling and illustrated its applicability with a simple example.

Keywords

Affiliated Institutions

• Columbia University US
• Columbia University Irving Medical Center US
• NewYork–Presbyterian Hospital US
• New York Psychoanalytic Society and Institute US
• New York State Psychiatric Institute

Related Publications