Evidence-appraisal glossary
External Validation
External validation tests a prediction model on data it has never seen, from a different place, time, or population, to check whether its performance holds up outside the setting where it was built.
A model can look excellent on its development data and still fail elsewhere because patients, measurements, or care patterns differ. External validation measures discrimination and calibration in this genuinely new sample, a far stronger test than reusing the original data. Without it, a model's reported accuracy is a promise rather than a track record.
This is a plain-language methodology definition for reading research. It is general education, not medical advice.