linear regression as an alternative to cox proportional hazards regression model

Question

I want to apply linear regression on age at first birth (response variable) data instead of cox regression. I will apply it on data for only females who have had a birth. Does this make sense?

Answer 1

Since all of the women in your sample have given birth, you have no censoring and don't need Cox PH (or any survival method).

BUT!! Does this make sense for your substantive questions? Clearly, lots of women in the real world never have a baby. You are ignoring those women. Also, in any actual sample of women of child bearing age, you will have some women in your sample who have not had a baby yet.

So, given your sample, your proposal makes sense. But I'm leery of your sample. It may be right (you haven't said much about your hypothesis) but ... at least for most situations, I think it is wrong.