P.Mean: SPSS won't compute my odds ratio! (created 2011-11-08). News: Sign up for "The Monthly Mean," the newsletter that dares to call itself average, www.pmean.com/news.

Dear Professor Mean, Help! I want to compute an odds ratio and confidence interval in SPSS, but I can't get SPSS to help because it refuses to compute the odds ratio anytime there is more than two rows or more than two columns.

If your table has more than two rows or more than two columns, then there is more than one odds ratio to compute. Here's an example

     X  Y
 A  22 21
 B  33 62
 C  50 23

You could look at the odds ratio between A and B

     X  Y
 A  22 21
 B  33 62

or between A and C

     X  Y
 A  22 21
 C  50 23

or between B and C

     X  Y
 B  33 62
 C  50 23

or you could combine categories, such as A+B versus C

     X  Y
A+B 55 83
 C  50 23

or A versus B+C

     X  Y
 A  22 21
B+C 83 85

or even A+C versus B

     X  Y
A+C 72 44
 B  33 62

Change your data (using RECODE, for example, or SELECT CASES) so that your table has exactly two rows and exactly two columns. Then, if you select the STATISTICS button and the RISK option, you will have your odds ratio and confidence interval.

The only possible problem is if it swaps the numerator and denominator from the way you want it. But this is a trivial pocket calculator exercise. If your odds ratio is 4.0 with a confidence interval from 2.0 to 10.0, then the inverted odds ratio is 0.25 with a confidence interval from 0.10 to 0.50.

Now you might want to consider adopting an approach like binary logistic regression or ordinal logistic regression, as this would allow you to estimate odds ratios without having to combine or remove some of your rows or columns.

This page was written by Steve Simon and is licensed under the Creative Commons Attribution 3.0 United States License. Need more information? I have a page with general help resources. You can also browse for pages similar to this one at Incomplete pages.