Date of Original Version

4-2013

Type

Thesis

Rights Management

All Rights Reserved

Abstract or Description

A mixed effects multinomial logistic model is useful in understanding a response variable with more than two outcomes and its relationship with covariates for nested data sets. Because of the nested structure of the data, using random intercepts is needed to adjust for the dependency within subgroups. This paper addresses this method while analyzing a psychological study performed by the Carnegie Mellon Psychology Department. The data are collected from local elementary schools to better understand the environmental effects that cause off task behavior in the classroom. Logistic multinomial models with Bayesian analysis are used to better understand what behavior and activity is responsible for making students go off task in the classroom.

Comments

Advisor: Howard Seltman

Department of Statistics

Embargo Date

2013

Share

COinS