The tendency is to use residual-like diagnostics on the entire dataset that was available for model development. For test data we typically run predictive accuracy analyses. For example, one of the strongest validations is to show, in a high-resolution calibration plot, that absolute predictions (e.g., probability of survival at 2 years) are accurate.
Frank ----- Frank Harrell Department of Biostatistics, Vanderbilt University -- View this message in context: http://r.789695.n4.nabble.com/calculating-martingale-residual-on-new-data-using-predict-coxph-tp3050712p3052377.html Sent from the R help mailing list archive at Nabble.com. ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.