Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Jessie C <jessiecoh@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: Nonparametric testing between two series |
Date | Sun, 13 May 2012 17:06:02 -0400 |
I have 2 series of data that are time series. I want to test whether the 2 series trend in a similar direction or not statistically. I'm not interested in the differences in the levels between the two series but only the difference in how the series is trending. For example, if one series is trending upwards and the other is trending flat, then I'd clearly like the test to reject that the two series trend similarly. If, however, the two series are both fairly flat, then potentially the test would fail to reject that the two series trend similarly (or maybe it would reject depending on how similar the flatness between the two is). And, moreover, if the one flat series is pretty much 200 for every data point and the other series is pretty much 20 for every day point, I wouldn't want the big difference in their levels to affect the test. I am blank on ideas. I'd imagine something like the Kolmogorov-Smirnov test is what I'm looking to do. Does that account for the levels of the series? I would appreciate any suggestions! Thank you for your time and help. * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/