The previous graph showed the apparent validity of a regression model. It is of more interest to study the internal validity, that is the performance on an underlying population (also labelled ‘reproducibility’), or external validity, that is the performance on a different population (also labelled ‘generalizability’ or ‘transportability’).