The Ordinal Count Factor Model: An Improved Latent Variable Model For Ordinal Count Items Public Deposited

Analytics

Downloadable Content

Download PDF
Request Version for Screen Reader
Last Modified
  • March 20, 2019
Creator
  • Markiewitz, Nathan
    • Affiliation: College of Arts and Sciences, Department of Psychology and Neuroscience
Abstract
  • Much of the measurement of human behaviors relies on the reporting of a rate of behavior. Common measures use items that ask participants to select between given intervals of counts—these items are called ordinal count items. I present the ordinal count factor model (OCFM) as a latent variable model for ordinal count item responses in a single population and across multiple groups. OCFMs represent the underlying latent response as a count, instead of the logistic or normal distribution used by current latent variable models for ordinal data. In addition to representing the data generating process more faithfully, OCFMs allow for inferences on the metric of the underlying rate of behavior. I evaluate the OCFM through two empirical examples using the Rutgers Alcohol Problem Index. These studies demonstrate that OCFMs may fit better than standard models, produce more precise factor scores, and may be fit using widely available, open-source software.
Date of publication
Keyword
DOI
Resource type
Rights statement
  • In Copyright
Advisor
  • Bollen, Kenneth
  • Curran, Patrick
  • Bauer, Daniel
Degree
  • Master of Arts
Degree granting institution
  • University of North Carolina at Chapel Hill Graduate School
Graduation year
  • 2017
Language
Parents:

This work has no parents.

Items

Thumbnail Title Date Uploaded Visibility Actions
Markiewitz_unc_0153M_16943.pdf 2019-04-11 Public
PREMIS_Events_Metadata_0_c4a91c82-e1dd-4a21-a440-8bf4ecab6903.txt 2019-04-11 Public
original_metadata_file_c4a91c82-e1dd-4a21-a440-8bf4ecab6903.xml 2019-04-11 Public