Stata can compute odds-ratios. It sounds that you want to perform a logistic regression. You need a binary outcome coded 0 or 1. Use the command logit with option or; or use the command logistic.

Orvalho, the right command is logistic

When we perfom logit, we only get the coefficient... Odds ratio or marginal effect can be obtained after.

Dear Remelien,

Careful with what I mean. As I said on my answer you may use the command logit with or option to obtain the odds-ratios:

logit y x1 x2 x3, or

that is equivalent to

logistic y x1 x2 x3

or to

glm y x1 x2 x3, family(binomial) link(logit) eform

I make note that both command are estimating the same regression coefficients in log-odds. To get the odds-ratios we need to e-exponentiate. Stata can e-exponentiate for you by "logistic" or "logit, or" or "glm, eform".

Also note, old versions of Stata would avoid to produce the exponentiated constant coeficient on the logistic command. That is why logit come to hand.

I hope it clarified.

Orvalho

hi Orvalho & Remelien,

Thank you for instant replies. i have the applied the same command that you recommended. i got results after adjusting age, gender, tobacco consumption but whenever i put alcohol variable in the command , results show some kind of error. is it normal error? what one can conclude with this kind of result?

please reply.

Dear Vertica,

I don't know what kind of error you mention. Can you share the output please?

Also note that Stata has its own forum for more help.