ABSTRACT
Peter Hedstrom and Richard Swedberg correctly point out that models are not intended to give "theoretical accounts of what happens as it actually happens", but that they ought to incorporate "the essential elements of the concrete situation". Hedstrom and Swedberg's surprise, find that participation is determined by social class without any observable intervening variables. It seems as if Hedstrom and Swedberg would be satisfied with developing a tool box of mechanisms from which they can pick one or two which in a specific case may have produced the observed conditions. The hard training in methods at graduate schools has sometimes led to the unlucky development that students looked for problems which could be approached by help of the techniques they had learnt. Typical examples refer to patterns of class mobility, social selection to higher education, temporal variation in birth rates, gender segregation in the labor market, patterns of class voting, etc.
