Endorsements by groups in American politics have typically been studied as voting cues only for members of the given organization. Using both the formal theoretical and low-information cognitive voting literatures, this article argues for a broader electoral role for group endorsements. Specifically, if groups that have clear ideological or policy preferences endorse candidates, these endorsements should provide all voters with ideological or issue information about the endorsed candidates. This inferred information should then impact voters' behavior, especially in low-information scenarios. Using both an experimental test and a test with American National Election Studies (ANES) survey data, this study analyzes the hypothesis in terms of elections to the U.S. House of Representatives. It finds that when the AFL-CIO endorses Democratic candidates, voters behave as though a liberal message has been sent-liberals are significantly more supportive while conservatives are significantly less supportive than they are when no endorsement is given, regardless of whether or not they are union members. At the same time, however, the analysis finds no support that endorsements of Republicans have any ideological impact on voting.
Group endorsements of political candidates have largely been studied as a limited phenomenon-affecting the attitudes and behaviors of group members exclusively. There is good reason to believe, however, that these endorsements could have broader effects. Both the general voting behavior literature and specific research into group effects and electoral cues provide theoretical grounding for the idea that group endorsements could act as informational shortcuts for all voters in low-information elections.
Research into American voting behavior has long demonstrated how most voters lack the political sophistication thought necessary to fully understand the world of politics (Campbell et al. 1960). As a result, voters rely on familiar and simplified methods of understanding politics, and even choosing candidates.
Social groups have been shown to play a central role in aiding individuals' political judgments in such a complicated world (e.g., Wlezien and Miller 1997). Politically relevant groups help voters orient themselves and others in the political realm through their affiliations and actions-providing valuable cues to those who share or oppose their interests (Brady and Sniderman 1985; Sniderman, Brody, and Tetlock 1991).
At the same time, research into how voters cope with their lack of information in choosing among candidates has formulated theoretical and empirical models for when and how such groups may play a direct role. According to formal theoretical models of voting behavior, social and political groups can have an effect on electoral politics when they become involved through either endorsing or opposing candidates for office. Group endorsements should act as cues for voters, signaling where an endorsed candidate stands ideologically, especially relative to an unendorsed candidate. In this way, endorsements can help voters choose between candidates (Grofman and Norrander 1990; Lupia 1992).
Theories of low-information voting and empirical studies resulting from them also support this reasoning. Voters take information shortcuts by inferring political information from candidate personal or social information that is readily available and easily digestible (Popkin 1991). Group endorsements of candidates easily fit this bill, and as a result, this literature also predicts important electoral effects.
Despite this emphasis in the voting behavior literature on the importance of groups and on voters' reliance on them, specific research into group endorsements has been limited both in quantity and scope. Most such research restricts examination of group influence to the effects of group endorsements on group members alone (see, for example, Rapoport, Stone, and Abramowitz 1991). …