Dear List, If I calculate the C-index after 10-fold cross-validation based on Cox model, which statistic test should I used to compare two C-statistc values.I means I use cross-validation to obtain a predictive probability for each individual, then combine all the probabilities to calculate the C-index.If I get two C-index values(0.73 vs. 0.80) based on two different model, then how to test the difference is significant. For the C-index values got from a single dataset, we may use bootstrap to get the standard error, and use wald test to measure whether they are significant different,but if c-index got from cross- validation,whether I can still use this method to compare them. -- View this message in context: http://r.789695.n4.nabble.com/how-to-compare-C-index-in-Cox-model-tp4548251p4548251.html Sent from the R help mailing list archive at Nabble.com.