Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
Help with which test to use
Message posted by Dwayne on September 9, 2000 at 12:00 AM (ET)
Can you tell me if you would use chi-square or t-test for data analysis of two separate groups that you are trying to find out if there is a significant difference in the frequency of something? I am a graduate student in Nurse Anesthesia. I did a randomized study comparing post operative nausea and vomiting in patients undergoing outpatient surgery. n=19 in group 1 and n=15 in group two. 2 were sick in group one and 5 in group two. Any suggestions will be greatly appreciated. Thank you.
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: Help with which test to use
Message posted by Bill on September 9, 2000 at 12:00 AM (ET)
You could compare the proportion of patients with the outcome in each group. A chi-square may be appropriate if the expected frequencies meet the minimum presented in most texts. The Fisher exact test would be appropriate in your situation. The result is Exact Test for: P1 = P2 Fisher Exact p-value(two-tailed) = 0.199. Thus you fail to reject the null-hypothesis of equal proportions per group.
Re: Help with which test to use
Message posted by JG on September 10, 2000 at 12:00 AM (ET)
A chi-squared test is usually done when comparing frequencies. In your case a test for independence may be appropriate.
Re: Help with which test to use
Message posted by JG on September 10, 2000 at 12:00 AM (ET)
You have a standard test for independece in a 2 by 2 table where your frequencies are 2,17,5,10 going row by row. There are many web sites that will do the computations for you. Search for chi-squred test for indepence.
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.
