In logistic regression model for binary data,

If the residual deviance(2*(likelihood of saturated model-likelihood of my model)) is 56.728 with df=117, what can I say about the lack of fit of the model?

I don't know whether the deviance should be small, or similar to the degree of freedom when the model fits well.

  • $\begingroup$ The deviance is interpreted in comparison to the null (intercept-only) model. ¿What is the deviance of the null model for your dependent variable? $\endgroup$ – Gregg H Apr 14 '18 at 16:18

Deviance is a model's goodness-of-fit statistic.

You first need to fit the null model. The difference between the deviances for the null model and your model follows an approximate chi-squared distribution with k-degrees of freedom(df). (k is the difference of dfs between two models.)

For example,

your model deviance (df) = 56.728 (117)

null model deviacne (df) = 70 (120)

=> 70-56.728=13.272 ~ chi_squred (3)

and then you can test the significance of your model.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.