Online Program Home
My Program

Abstract Details

Activity Number: 603
Type: Contributed
Date/Time: Wednesday, August 3, 2016 : 2:00 PM to 3:50 PM
Sponsor: Biometrics Section
Abstract #320404 View Presentation
Title: Missing Data Approaches in Categorical Latent Growth and Multilevel Proportional Odds Models
Author(s): Karen Traxler* and Niloofar Ramezani
Companies: University of Northern Colorado and University of Northern Colorado
Keywords: Missing data ; Growth models ; Categorical ; Multilevel ; Imputation
Abstract:

Missing data presents unique challenges when building growth models (Cheema, 2014). The purpose of the current study is to compare three methods of handling missing data; listwise deletion, mean imputation, and multiple imputation, in a two-level categorical latent growth model and a multilevel proportional odds model conducted to assess mental health recovery over time. Data were collected from a residential treatment facility in North Carolina at five time periods. The strength of any growth model over other regression models is that it can differentiate between individual starting points (intercepts) as well as individual change/recovery over time (slopes) (Raudenbush & Bryk, 2002) based on explanatory variables added to the model at level two. The explanatory variables include four classifications of primary diagnoses, age, and gender. The outcome variable is represented by categories of recovery established by Young and Ensing (1999) mental health recovery measure (MHRM). Each missing data method will be assessed within the two aforementioned multilevel models and the most appropriate techniques will be discussed.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2016 program

 
 
Copyright © American Statistical Association