Generalizing and theorizing

When we’ve completed our data analysis, it’s time to complete the loop, which entails at least three somewhat overlapping tasks. First, we return to our research question, not asking it, but answering it. What did we learn? To what extent can we generalize our findings—to a larger population, to other settings, to other cases, to other times? We do this humbly. In social research, the claims we make are almost always provisional. We rarely claim to wholly “answer” a research question, and we virtually never claim to “prove” anything. We state our conclusions tentatively, realizing that future research could improve on, expand, or even contradict what we’ve learned. We also acknowledge the limitations of our own research (which are always present) and suggest directions for future research. (See Appendix F about how to avoid the common error of generalizing from groups to individuals.) Second, we relate what we’ve learned to previous research. How is what we’ve learned consistent with previous research? How is it different? Where does it fit in to the larger body of knowledge? Third, we advance what we know about theory. From a deductive perspective, does the theory that drove our research seem to be a good fit with what we’ve observed? How might our observations suggest we should modify the theory? From an inductive perspective, what theory did we construct based on our observations? How might future research test this theory?