The theoretical approach involves generalizations, and as there are few if any true generalizations in the social sciences, this means generalizations which are approximations of some kind. The difficulties, and attractions, of this approach can best be seen in a classic example provided in the most influential exposition of this approach, Hans Zetterberg’s Theory and Verification in Sociology (1965: 23). The construction of the theory begins with two studies: a social psychology experiment and a study of student voting at Bennington College. In the experiment, students in different groups were given a task of writing a story connecting pictures. Some groups were told that they were "model groups," while other groups were not told this. Those who were told that they were model groups were more likely to agree in the writing of the story (Newcomb, 1943). In the study of student voting, it was found that the students who were voted to represent the college in an intercollegiate meeting were likely to be more affected by the liberal values of their teachers and fellow students than those not selected (Back, 1951).
The theorist seeks some more general formula relating the determinants, i.e. elected by popular vote and told they were a model group, and the outcomes, i.e. being more affected by predominant liberal values and agreeing more in the writing of the story. To generalize the determinants, Zetterberg suggests the following: they received more favorable evaluations. For the outcomes, he suggests: their ideas converged more with other group members. This yields a theoretical formula: “The more favorable evaluations rank and file members receive in a group, the more their ideas converge with those of other group members” (Zetterberg, 1965: 24).
Question: Rethink the Bennington study, but this time come up with a more abstract generalization that fits the findings. Would one be able to find confirmation for this generalization in a different setting as well? Give some possible examples.