Models for Ordered Categorical Variables: An Application

Ordered logit and ordered probit models, commonly used ordered categorical variable models, are used when the dependent variable is categorical and ordered. The validity of the predictions of these models depends on the assumption of parallel slopes. These models can be used if the assumption of...

Full description

Saved in:
Bibliographic Details
Main Authors: Öznur İşçi Güneri, Burcu Durmuş, Aynur İncekırık
Format: Article
Language:English
Published: ORDT: Organization for Research Development and Training 2024-11-01
Series:Journal of Interdisciplinary Sciences
Subjects:
Online Access:https://journalofinterdisciplinarysciences.com/wp-content/uploads/2024/11/2-Models-for-Ordered-Categorical-Variables-An-Application.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Ordered logit and ordered probit models, commonly used ordered categorical variable models, are used when the dependent variable is categorical and ordered. The validity of the predictions of these models depends on the assumption of parallel slopes. These models can be used if the assumption of parallel slopes is met. When this assumption is not met, the generalized ordered logit/probit model, which is more flexible in terms of assumption, or the multinomial logit model can be used by ignoring the assumption. In this connection, an exemplary data set was taken, and first of all, the assumption of parallel slopes was investigated. These models were compared using loglikelihood, Akaike Information Criteria (AIC), and Bayes Information Criteria (BIC) statistics for the validity of the models. In this way, the lowest log-likelihood, AIC, and BIC values were found for the generalized logit model.
ISSN:2594-3405