SPSS Statistics

I assume the proportions are different but I want to know formally if :

84 positives out of 99 (Proportion around 0,848) are different than

11 positives out of 24 (proportion around 0,458)

Using Analyze -> nonparametric testing -> legacy dialogs -> binomial

I have entered my 4 numbers as variables but cant make the program calculate the confidense interval for the difference between the two proportions. You enter a fixed comparator proportions but how to compare two different proportions ?

If you install the PROPOR extension command via Extensions > Extension Hub (it will appear under Analyze > Descriptive Statistics > ,Proportion Confidence Intervals), you can enter these numbers and get the ci for the difference.

For your data, the 95% interval is (-.602, -.179)

Hi Jon

Thanks - Grateful ! It works.

BTW...is there a shortcut in SPSS from CI to P value (the observed type 1 error risk for this comparison). I guess I need to make an assumption of the distribution of the difference between these two binomials ?

Kind regards

René Worck

I'm not sure what you are looking for in ahortcut, but what you see is pretty much what you get. This page goes through the steps for some typical calculations.

How to obtain the P value from a confidence interval | The BMJ

It wouldn't be hard to encapsulate this in a custom dialog box or extension command. Some meta-analysis tools might help with this.

OTOH, don't get too hung up over p values.