Goodness of fit for univariate logistic regression

[Nita Chan]

Hi,
I am wondering if it is possible to carry out goodness of fit tests on unadjusted univariate logistic regression? I found the estat gof command but I don't think it works purely for a univariate logistic regression with one predictor. Is there a different way to assess goodness of fit? I have a few predictors and want to assess their associations independently with my outcome variable to and need to assess and record their goodness of fit individually.

Kind regards,

Anita

[Steve Samuels]

What kinds of predictors-categorical, continuous?

[Nita Chan]

Hi,
The predictors are binary and categorical.

[Steve Samuels]

A logistic model with a single categorical variable predictor (or set of predictors, for a variable with >2 categories) will exactly reproduce the crude category proportions. The predictions perfectly fit the data, so there's no place for a goodness of fit test..