Proportional odds logistic regression model: F-statistics, Multiple R-squared and Adjusted R-squared

[MAY90]

Hello
I'm running a "polr" model, but in the output of course, there is no F-statistics, Multiple R-squared and Adjusted R-squared.

Does anyone has an idea, what could be an other alternative to measure them and the goodness of the fit of the model?

• For the R-squared I thought maybe a pseudo R-squared would be an alternative but for the F-statistics I don't have any idea.

Thank you very much for your help.

Best regards

[technocrat]

See https://stats.stackexchange.com/questions/143024/how-can-i-explain-proportional-odds-models-to-a-layman

The differences in assessing the output of a proportional odds logistic regression and an ordinary least squared regression are analogous to the differences with logistic regression (discussed at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1065119/) . Basically, you get the ammunition for applying test statistics, which it's up to the analyst to decide which are appropriate for the data and purpose, which is often domain-specific.

If you want to take a deep dive, a standard text is Applied Logistic Regression, 3rd Edition
by David W. Hosmer Jr., Stanley Lemeshow, and Rodney X. Sturdivant (2013) (Wiley).