I presented in this chapter several examples of bivariate and trivariate cross-tabulations involving voting in the 2012 elections. What additional influences would you recommend examining to explain voting in elections? Suggest some additional independent variables for bivariate analyses with voting as well as several additional control variables to be used in three-variable crosstabs.
When should we control . . . just to be honest? In the evaluation project that I will describe in Chapter 13, I analyzed with some colleagues the effect on cognitive functioning of living in group homes rather than individual apartments. I found that living in group homes resulted in gains in cognitive functioning, compared with living in individual apartments. However, this benefit of group homes occurred only for residents who were not substance abusers; substance abusers did not gain cognitively from living in group (or individual) homes (Caplan et al. 2006). Would it have been all right if we had just reported the bivariate association between housing type and change in cognitive functioning? Should social researchers be expected to investigate alternative explanations for their findings? Should they be expected to check to see if the associations they find occur for different subgroups in their samples?