If there are more than two races being compared, a Kruskal-Wallis should work.
Rick Dr. Rick Froman Psychology Department Box 3055 x7295 [EMAIL PROTECTED] -----Original Message----- From: Claudia Stanny [mailto:[EMAIL PROTECTED] Sent: Tuesday, March 15, 2005 4:23 PM To: Teaching in the Psychological Sciences Subject: Re: Data Analysis At 04:57 PM 3/15/2005 -0500, you wrote: > > Truhon, Stephen wrote: >> >> Does anyone have an idea of how to analyze data where one variable is >> nominal (race) and the other variable is ordinal (problems in school from >> none to serious)? I could do a chi-square but I hope to make of the ordinal >> characteristic of the second variable. > > Most everyone uses a 2-sample t-test? Why should you demur? :-) > > -- > Christopher D. Green The ordinal scale of the dependent measure might make one reluctant to use a t-test. But a non-parametric version of the t-test (e.g., Mann-Whitney) might capture the difference in problem severity between the two groups. Claudia Stanny ________________________________________________________ Claudia J. Stanny, Ph.D. Associate Professor General Track Coordinator Department of Psychology University of West Florida Pensacola, FL 32514 - 5751 [EMAIL PROTECTED] Web Site: http://uwf.edu/cstanny/website/index.htm Phone: (850) 474 - 3163 FAX: (850) 857 - 6060 --- You are currently subscribed to tips as: [EMAIL PROTECTED] To unsubscribe send a blank email to [EMAIL PROTECTED] --- You are currently subscribed to tips as: [email protected] To unsubscribe send a blank email to [EMAIL PROTECTED]
