Population sampling in longitudinal suverys

In an opening paper Harvey Goldstein questions the need for observational studies to achieve representativeness for real populations, in particular for longitudinal studies. He draws upon recent debates and argues for the need to distinguish scientific inference from population inference. The points he raises are then debated in commentaries by Peter Lynn, Graciela Muniz-Terrera and Rebecca Hardy, Colm O'Muircheartaigh, Chris Skinner and Risto Lehtonen. These commentaries are followed by a response from Goldstein.