Population sampling in longitudinal surveys

Abstract

The paper questions the need for observational studies to achieve representativeness for real populations, in particular for longitudinal studies. It draws upon recent debates and argues for the need to distinguish scientific inference from population inference.