• Practice your interview with the new SDN AI Interview Coach. Choose a school, answer by voice or typing, and receive a personalized feedback report. Available now to all SDN members. Try the AI Interview Coach.

test for median comparison

Started by pjmrg
This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.

pjmrg

New Member
10+ Year Member
15+ Year Member
Advertisement - Members don't see this ad
hi, i hope someone could help me, i am doing some research to find a test that could compare medians, like te t-test that compares means, could you help me please? and if it exists... how do i use it in the spss??? hey thanks for any help you can give me.
 
hi, i hope someone could help me, i am doing some research to find a test that could compare medians, like te t-test that compares means, could you help me please? and if it exists... how do i use it in the spss??? hey thanks for any help you can give me.

You will need to use non-parametric statistics, therefore look around for the Mann-Whitney test, or Wilcoxon Rank-Sign test. The conditions for Mann-Whitney are that the samples are independent, while the Wilcoxon Rank-Sign test requires that the samples are dependent. Suggest you google or wikipedia it up, or even better talk to a statistician. Most universities provide free consultation to their students.

I've used SPSS but I'm not too familiar with it. Suggest you use a program like SAS if you intend on publishing results or if you talk to statisticians. Most of the stats profs use SAS. If you're in a bind, then MiniTab works too..however that program has a nasty habit of reporting P<0.0001 as P=0.000...which is silly.