Quantitative independent variable with empty cells

[Malin Lumio]

Hi,

I am doing a logistic regression analysis with a binary outcome. I have an independent variable measured on ratio scale (age). Since it includes empty cells when crosstabulating with the dependent variable, I use age categories instead. Is this correct or are there better ways to do this?

Thanks in advance!

[Maarten Buis]

If you add age as a continuous variable, then empty cells are not a problem. Empty cells would only be a problem if you added age as a categorical variable. If the range of age is large enough, then age is very likely to have a non-linear effect.You could accommodate that with polynomials or splines.

[Nick Cox]

See https://stats.stackexchange.com/ques...ssion-in-stata

You were advised on Cross Validated (CV) to post here -- but you are requested here to tell us about cross-posting.
http://www.statalist.org/forums/help#crossposting applies.
Note that Maarten's excellent advice echoes advice already given on CV. If he had known about your posting there he might not have bothered.

Conversely if you have abandoned your thread on CV, you are well advised to delete it to clean up after yourself. (I write as someone also active on Cross Validated.) There, although not here, you can delete threads without upvoted answers.

[Malin Lumio]

Nick Cox Thank you for informing about that. I didn't know about this. I am a very new member to both Cross-validated and Statalist.

I have deleted the original post in Cross Validated.