A note on cross-validating prediction equations

Cross-validation typically refers to the application of regression weights derived in one sample (the derivation sample) to a different sample (the validation sample) to investigate the stability of relationships based on the original weights. When the standard cross-validation paradigm is thought to reflect accurately the use of a prediction equation in practice, the researcher should use an appropriate cross-validation strategy. Two superficially similar cross-validation procedures are presented. (12 ref) (PsycINFO Database Record (c) 2006 APA, all rights reserved).