Odds ratio in logistic regression on rare even

[t75wez1]

hello, 

my data include two variables with 20,000 records. One is whether or not received a product sample and the other is whether use the product or other products after received the product sample. I wanted to calculate the odds ratio to conclude that the consumer who received the product sample would be 2 times likely to use the product than the consumer who didn't receive the sample. However the product has only 3% of market share among all the products consumers use.

how to solve the problem in SAS?

I'm very appreciative of any helps?

Ethan

[Reeza]

Here's a good quick reference, Paul Allisons thoughts on the issue

http://statisticalhorizons.com/logistic-regression-for-rare-events

You can google with the following term to get more responses : Logistic regression rare events

[Ksharp]

Exact Logistic Regression

[StatDave]

Since you describe a 2x2 table, you could avoid fitting a model altogether and simply use PROC FREQ and request an exact estimate of the odds ratio with the EXACT OR; statement.