SAS Analyst Part 2

deleted_user — Fri, 16 Jan 2009 03:17:24 GMT

First, thank you very much for your comments to my last post, they were very helpful! Second, after I find the differences, which is my independent and which is my dependent variable?? Do I just create a column of all ones and call that my independent variable and then the difference column is the dependent variable?? Thanks!!!!

Re: SAS Analyst Part 2

Doc_Duke — Fri, 16 Jan 2009 13:58:11 GMT

You said earlier that you had paired data on two independent samples. "Independent" means separable or distinct (like males and females or two drug treatments in a clinical trial), so you should have an indicator variable that demarks the two independent samples.

The dependent variable is the difference score.

If you are looking for a non-parametric ONE-way test of location, then you can do it in SAS/Analyst, but it is a pain in the butt. The test is called a Wilcoxon Signed Rank test (and is part of PROC UNIVARIATE in base SAS). You can read an introductory non-parametric statistics book to see how to do the analysis by hand and then reproduce those steps in SAS/Analyst. You are probably better off using a different tool than SAS/Analyst.

Doc Muhlbaier
Duke

topic SAS Analyst Part 2 in SAS Procedures

SAS Analyst Part 2

Re: SAS Analyst Part 2