Also, as Frank Harrell writes (in the context of model building): "When Y is binary or continuous (but not censored), a good general-purpose measure of association that is useful in making decisions about the number of parameters to devote to a predictor is an extension of Spearman's rho rank correlation. This is the ordinary R^2 from predicting the rank of Y based on the rank of X and the square of the rank of X. This rho^2 will not only detect nonlinear relationships (as will the ordinary Spearman rho) but nonmonotonic ones as well."
SR Millis "Wuensch, Karl L" wrote: > > Rich said "Spearman's, especially, serve no end that I think of." > > I'm probably taking this out of the context of the thread, but Spearman rho > is a potentially useful statistic if you wish to measure the monotonicity of > a relationship rather than the linearity. ~~~~~~~~~~~~~~~~~~~~~~~~~~ Scott R. Millis, PhD, ABPP (CN, RP) KMRREC Research 1199 Pleasant Valley Way West Orange, NJ 07052 http://www.rehabtrials.org NOTICE: (1) The foregoing is not intended to be a legally binding or legally effective electronic signature. (2) This message may contain legally privileged or confidential information. If you are not the intended recipient of this message, please so notify me, disregard the foregoing message, and delete the message immediately. I apologize for any inconvenience this may have caused. . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
