Model Specification and Operationalization

Abstract

Similar to the selection of cases, the selection of variables must be guided by theoretical criteria. Here, in contrast to the case selection, we are confronted with the opposite situation, namely, a potential abundance of factors to be considered. Given the state of social science theory for most empirical questions, a large number of variables often cannot be excluded a priori. The conventional way to select a limited number of variables is to test any relevant hypothesis for the problem concerned in a strictly Popperian falsificatory manner. Thus, for example, the well-known Lipset thesis that “the more well-to-do a nation the greater the chances that it will sustain democracy” (Lipset 1960: 31) can be tested in this way. In the contemporary world, this hypothesis is verified for about 70% of the cases that are relatively well-established and consolidated democracies. However, the hypothesis does not account for poor countries with relatively stable democracies like Botswana, India and Papua New Guinea or for the failure of democracy in relatively well-developed ones like Weimar Germany.