1
$\begingroup$

Can I compute Odds ratio coefficients and their 95% CI from a Poisson regression (GLM) using the exp function?

I have an ordinal dependent variable (i.e., a GS of a scale) and some independent variables which are either categorial or continuous (e.g., Age, Presence/absence of hypertension ...).

For each variable, in R, I did this: #Poisson regression model_name1 <-glm(dependentvariable ~ indepvariable1, family = poisson(), data=data_name)

OR and 95% CI

exp(coef(model_name1))
exp(confint(model_name1)) 

My question is: can estimate the ORs of a poisson regression (as for a logistic binomial regression)? Can I compute it with the function "exp" as I did? Thank you

$\endgroup$
  • 2
    $\begingroup$ Possible duplicate of Log-linear regression vs. logistic regression $\endgroup$ – AdamO Dec 15 '17 at 14:46
  • $\begingroup$ You seem rather confused about what you are trying to do. If the link which @AdamO provides does not sort out your problem can you edit your question to give us more details about what precisely your scientific problem is? $\endgroup$ – mdewey Dec 15 '17 at 14:49
2
$\begingroup$

If you have a Poisson regression then exponentiating the coefficients gives you the multiplier corresponding to a unit change in the predictor variable. So if you measure age in years and the coefficient was 0.1 then a difference of one year would correspond to being $e^{0.1} = 1.10$ times higher scoring or about 10% more. It is not an odds ratio. Some software calls it an incidence rate ratio but you do not seem to have an outcome variable for which that interpretation makes much sense.

$\endgroup$

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.