| To: | s-news@lists.biostat.wustl.edu |
|---|---|
| Subject: | arcsine transformation with negative values |
| From: | Christoph Scherber <Christoph.Scherber@uni-jena.de> |
| Date: | Tue, 21 Feb 2006 20:48:31 +0100 |
| User-agent: | Internet Messaging Program (IMP) 3.2.3 |
Dear S list members, I am currently analyzing a dataset that consists of differences of percentage values - i.e., the values may range from -100% to +100%. How can I transform these data in order to meet the assumptions of parametric statistics? I´ve been thinking of adding a constant and then using a transformation like transformed<-asin(sqrt(p)) But the p values would then not lie in the interval [0..1]. I would be most grateful for any advice. Best regards Christoph. ---------------------------------------------------------------- This mail was sent through http://webmail.uni-jena.de |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Kolmogorov-Smirnov test for nested alternatives?, Bill Shipley |
|---|---|
| Next by Date: | Firm effects in estimating earning function, Patrick.Heer |
| Previous by Thread: | Kolmogorov-Smirnov test for nested alternatives?, Bill Shipley |
| Next by Thread: | Firm effects in estimating earning function, Patrick.Heer |
| Indexes: | [Date] [Thread] [Top] [All Lists] |