0
$\begingroup$

If I estimate a model with logistic function:

glm(Yi~Xi,Di, data=data, family=binomial(link="logit") )

where Di is dummy.

The estimated coefficients are: (Intercept) =0.9, Xi=2.3, Di=-0.33

If I exponentiated the estimated coefficients returned by this regression, are these considered odds or odds ratios?

$\endgroup$

1 Answer 1

1
$\begingroup$

The term (Intercept) is an odds for the response Yi when Di=0 and Xi=0. Depending the encoding, the other term(s) will represent odds ratios for the response for the specific (non-zero) levels of Di compared to 0. If Xi is continuous, the odds ratio is a ratio fo odds for groups differing by 1 unit in Xi.

$\endgroup$
2
  • $\begingroup$ added a clarification $\endgroup$
    – Jenny
    Apr 27, 2021 at 21:24
  • $\begingroup$ It shouldn't matter to know what the coefs actually are. $\endgroup$
    – AdamO
    Apr 27, 2021 at 21:38

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.