Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 1 down vote favorite

For a school research project, I have been instructed to run a large number of tests comparing the proportions of different groups. (Like this, except Statistica will give me a p outright.) The process is laborious and I would like to automate it. I would however need the formula for this test, and internet searches have turned up only inconsistent and unhelpful results.

Is the formula especially complex? Am I missing some important point?

share|improve this question
You mention Statistica. Can you use it for your project? I never used it but it should provide some kind of "by-processing" feature for the Chi-Square-Test which is often used for testing for differences in proportions. – psj Dec 8 '11 at 13:54
The Chi-Squared test does look somewhat like what I'm looking for, but en.wikipedia.org/wiki/Chi-squared_test"seems to suggest that it tests for variance, whereas I just want to determine whether the difference in proportion of a binary variable between two groups is statistically significant. What I would really like is a formula that I can give n1(sample group 1) p1 (proportion of successes in sample group 1) and n2 p2 which will output the p-value – Alex Dec 8 '11 at 19:01
The Chi-square-statistic can be used for a number of tests. One use is to test if two proportions are equal. I have never seen the formula for the p-value but it should be possible if you first compute the value of the Chi-square statistic by computing the expected cell counts before. Afterwards you'd need the (complement of the) cdf of the Chi-square distribution with 1 degree of freedom. The cdf is built many programs (even Excel), if you cannot access one of them, your way leads to computing the incomplete gamma function which could be cumbersome if not built in somehow... – psj Dec 10 '11 at 21:09

Know someone who can answer? Share a link to this question via email, Google+, Twitter, or Facebook.

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.