Hi all,

I'm going over past papers in preparation for my stats exam. I've got stuck on one question that has data presented in a table - sample size, incidents of heart disease and number of years since diagnosis. The first two (out of 8) lines look like this:

156 3 1

252 17 2

where the first column is sample size, then incidents of heart disease, then number of years since diagnosis.

The question wants us to explain how we would analyse this data to assess the relationship between incidents of heart disease and years since diagnosis.

I'm guessing I have to do a type of regression since part b asks if this model is preferable to one where a separate proportion is fitted for each number of years since diagnosis.

I can't see it being MANOVA since there is not categorical information in my data set, ANOVA just doesn't work, but I don't understand how I would analyse it using multiple linear regression. I'm thinking the "separate proportion" is a big clue but it's just not ringing any bells with me.

Does anyone have any tips on how they would analyse this?