Item Details

title

Dueling Dyads: Regression Versus MLM Analysis with a Categorical Predictor

author

Hwang, Yoo Ri

abstract

Dyadic data is widely used in social and behavioral studies. However, specific techniques must be used due to the nonindependence of dyadic data. This study compared multilevel modeling (MLM) and the pooled-regression approach in the dyadic analysis context. Furthermore, within the context of dyadic analysis, the modeling of categorical explanatory variables is understudied despite its usefulness and importance (Yaremych et al., 2022). This simulation study investigated the effect of sample size (the numbers of dyads), the main effect of the categorical variable on the dependent variable (DV), the interaction effect between the categorical variable and the continuous variable on DV, and the intraclass correlation (ICC) on power and Type Ⅰ error rate of parameters of each model. The results indicated that overall, MLM showed higher power than the pooled-regression approach. When investigating interactions, the pooled-regression approach is not recommended due to the high Type Ⅰ error rate. Forcing individual categorical variable into a level-2 variable is also not recommended for MLM. Further implications and limitations were discussed.

subject

dyadic analysis

multilevel modeling

pooled-regression approach

contributor

Garrison, S. Mason (committee chair)

Brady, Shannon T (committee member)

D'Agostino McGowan, Lucy (committee member)

date

2022-09-17T08:35:42Z (accessioned)

2023-09-16T08:30:07Z (available)

2022 (issued)

degree

Psychology (discipline)

embargo

2023-09-16 (terms)

identifier

http://hdl.handle.net/10339/101252 (uri)

language

en (iso)

publisher

Wake Forest University

type

Thesis