How about making hypotheses actually predict something?
May 9, 2010 § Leave a comment
In the course of reading the otherwise well-written front end of an article, “Pathways to Meaning: A New Approach to Studying Emotions at Work” by Grant, Morales and Salaz in AJS 2009, The Disgruntled Sociologist came across the following “Hypotheses”:
Hypothesis 1.—Net of vertical work relations, a worker’s adoption of a corporate‐sponsored meaning is influenced by his or her horizontal work relations, cognitive framing, and involvement in outside institutions.
Hypothesis 2.—A worker’s adoption of a corporate‐sanctioned meaning is conditioned by how his or her vertical work relations, horizontal work relations, cognitive framing, and involvement in outside institutions combine.
Hypothesis 3.—A worker’s sense of authenticity varies by the pathway that he or she takes to a corporate‐sanctioned meaning.
Really? These are hypotheses? “Is influenced?” “Is conditioned?” “Varies?”
If all your theory can do is generate the prediction “there will be variance,” maybe you should try to make the theory more precise. In fact, a set of hypotheses like this are prima facie evidence that the authors have no theory. Indeed, TDS can see how, in the absence of any theory, it would strike the authors as hazardous to actually place a bet.
But perhaps that’s the point of the whole enterprise.