Application of a Multivariate Multilevel Growth Model With Dyadic Data

• In Event: Poster Session 6
  In Poster Session: Advanced Issues in Multilevel Modeling

Abstract

This paper illustrates an application of a multivariate dyadic growth model using multilevel modeling. The data used in this paper were collected from couples participating in a randomized control trial of a relationship education program. The couples were assigned to treatment or control conditions and completed the same set of four assessments at four time points, from pre-treatment to a 180 day follow period. A two-level multilevel model with random effects specified for indistinguishable dyads that also controls for multiple outcome measures is described. The model tests whether treatment and control groups have similar assessment scores for each measure at baseline and whether the treatment has a significant impact on growth rate for any or all of the four assessments.

Authors

• Tammiee S. Dickenson, University of South Carolina - Columbia

• Ryan G. Carlson, University of South Carolina - Columbia

• Sejal Barden, University of Central Florida