Where do I find odds ratios for a binomial logit generalised linear model?

[#1]

Hello,

I am trying to run a binomial logit generalised linear model and am having some difficulties finding where the odds ratio and p-values are. My independent and dependent variables are both binary (yes/no). I am trying to run an unadjusted model first, and then re-running the model but adjusting for 2 confounders.

However, when I run the model, there is no option to open odds ratios under the drop-down menu (I have attached a picture of what I see). If anyone knows where I can access the odds ratios and p-values, that would be much appreciated

[#2 · Accepted solution]

You are correct, there is no command in that platform for this result. But you have two choices. You can use the Nominal Logistic platform instead of the GLM platform. It will provide odds ratios with a command. You can also do the work in the data table. Save the current model as a column formula. You use the probabilities to compute the odds of an event and then the desired odds ratio for two conditions.

[#3]

Great thank you! I've tried doing it in the data table; I can find the odds ratio, but I'm still unsure how to find the p-value. Do you have any suggestions?

[#4]

The odds ratios are derived from the predicted probabilities. The model predicts the probabilities. Use inference on the model, then the odds ratios are significant, too.

[#5]

I'm also not sure how to calculate OR in the data table when I adjust for my 2 confounders. I can convert the probabilities into odds, but since each cell has a different value within it, I'm unsure how to calculate the adjusted OR

[#6]

I think that you have two choices. You can average the probabilities over the data or you can use the profiler to obtain the probabilities for specific levels of the covariates. It depends on your purpose or the meaning of the odds ratio you seek.