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�).