This dissertation examines the electoral impact of interest groups’ endorsements. I start by hypothesizing that endorsements may influence both vote choice and turnout. I theorized that voters may use the endorsement as a vote choice cue, particularly in elections when party identification is irrelevant. The nature of these cues may be either cognitive (in the form of a heuristic) or affective (following balance theory). Many factors may moderate this relationship, especially commitment to the group. I further theorized that endorsements also signal to voters that a given election is important, which will increase turnout, with commitment also moderating this relationship.
I used experiments and surveys to determine the degree of influence endorsements exert on vote choice and turnout, the effects of commitment on this relationship, and the relevant vote choice mechanism. First, I designed a laboratory experiment that randomly assigned an endorsement to one of two fictional candidates, or to neither of them, within the confines of a computerized mock primary campaign.
Participants experienced the campaign period, choosing whether to examine each of eleven discrete bits of information about the candidates, including the endorsement. Just prior to their "vote," I randomly activated either affect or cognition. The results showed a statistically insignificant increase in voting when cognition was primed.
Second, I included a survey-based experiment on a random telephone survey of Ohioans, which asked about three statewide races. Half the sample received only their names and partisan affiliations, while the other half also learned about endorsements by the Fraternal Order of Police (FOP) and the Ohio Education Association (OEA). Separately, they told whether they thought crime or education was a more important problem in Ohio. Results here were also tepid: no relationship survived a multivariate analysis, but some evidence did exist that the endorsement had some marginal effect.
Finally, to examine endorsements’ impact on turnout, I used the above survey experiment and the American wave of a international election survey. Partially due to measurement difficulties, no impact whatsoever is demonstrable here. In the concluding chapter, I reviewed the project and suggested refinements and additions.