Hi Yaakov,

I'm no expert on this but there seems to be a place for odds ratios if you are doing logistic regression. Almost the same question as yours has been asked before on ResearchGate and here is the link to those more expert answers than I could give.

https://www.researchgate.net/post/Is_there_a_manner_to_obtain_an_Odds_ratio_or_relative_risk_by_using_linear_regression

Incidentally what are you plotting ? I'd be interested in seeing the plot(s) in question.

Best wishes

TomH ([email protected])

thanks

If your dependent variable is continuous, using odds ratios in the interpretation of regression coefficients wouldn't make sense - unless you are binning your dependent variable and doing an ordinal or multinomial logistic model. Theoretically you might be able to do it with a log link function in a generalized linear model - but again it doesn't seem appropriate and I've never seen it done.

Maybe I'm misinterpreting.

thank you very much

Hi Yaakov, From our replies it sounds to me that you now have good grounds to go back to the Journal Editor and ask for that reviewer of your paper to expand on what he/she is expecting from you !

Regards

TomH

I agree with what others have said. Odds ratios are more difficult for most people to understand, so I don't know the benefit of what you've been asked to supply. You might want to ask the reviewer what he/she envisions as the result. It may be that they're asking for something different from what you understood.

Thanks Thomas and Kevin for your suggestion about asking the reviewer,

I agree completely, and was thinking of asking the reviewer, but I wanted to be sure, and understand better. So thanks for this information. I feel more confident.

Thanks

Best

Yaakov

It depends on the variables that you are using.

Most likely, your reviewer wants you to use the logistic regression.

If you use binary independent variables then the corresponding coefficients will be logs of odd ratios.

But in order to say precisely, you need to describe your current model here.

Maybe this page will help:

https://stats.idre.ucla.edu/other/mult-pkg/faq/general/faq-how-do-i-interpret-odds-ratios-in-logistic-regression/

If your DV is for example pain you can recode it as pain= yes = 1 and no pain = no = 0 and then compute a logistic regression analysis.