How to determine statistical significance?....

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

How to determine statistical significance?....
Message posted by mike on November 12, 1999 at 12:00 AM (ET)

I want to compare results of a survey conducted two years in a row. The sample size is different. If last year 83% (n = 53) answered "yes" to a question, and this year, 89% (n = 40) selected "yes", how do I determine if the 89% is a statistically significant change??

Thanks!!

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: How to determine statistical significance?....
Message posted by JG on November 12, 1999 at 12:00 AM (ET)

There are many things you can do. Try a Chi-squared test for independence by setting up a 2 by 2 table, etc.

Re: How to determine statistical significance?....
Message posted by Jack Tomsky on November 12, 1999 at 12:00 AM (ET)

Since you're testing if p1=p2, you can use Fisher's exact test. This is based on the hypergeometric distribution.

Your $5 contribution helps cover part the $500 annual cost of keeping this site online.