A Generalized Hosmer–Lemeshow Goodness-of-Fit Test for Multinomial Logistic Regression Models
Morten Wang Fagerland, David W. Hosmer
Abstract
Open-access reader
Morten Wang Fagerland, David W. Hosmer
Abstract
Open-access reader
Testing goodness of fit is an important step in evaluating a statistical model. For binary logistic regression models, the Hosmer–Lemeshow goodness-of-fit test is often used. For multinomial logistic regression models, however, few tests are available. We present the mlogitgof command, which implements a goodness-of-fit test for multinomial logistic regression models. This test can also be used for binary logistic regression models, where it gives results identical to the Hosmer–Lemeshow test.
OpenAlex reports 282 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Testing goodness of fit is an important step in evaluating a statistical model. For binary logistic regression models, the Hosmer–Lemeshow goodness-of-fit test is often used. For multinomial logistic regression models, however, few tests are available. We present the mlogitgof command, which implements a goodness-of-fit test for multinomial logistic regression models. This test can also be used for binary logistic regression models, where it gives results identical to the Hosmer–Lemeshow test.
Key concepts: Multinomial logistic regression, Goodness of fit, Logistic regression, Statistics, Logistic distribution, Factor regression model, Regression diagnostic, Multinomial distribution