Three Essays on Policy Activation and Information Inequality in Presidential Campaigns

Do campaigns help voters use policy preferences in their electoral decisions? Recent campaigns research emphasizes this role, known as policy activation, but evidence remains mixed. The inconsistency results from frequent reliance on models that smooth over important sources of variation in message receptivity. In the first essay of this dissertation I examine policy activation in 2008. Because voters with preexisting attachment to candidates tend to resist contrary information, I focus on voters who enter the general election phase undecided between the major party candidates. And because most voters have little incentive to pay attention to any single policy, I use a more flexible approach---a scale based on multiple measures of policy preference. Approaching the study of policy activation in this way reveals that these attitudes, measured during the primary phase of the election, indeed predict change during the campaign but the magnitude remains modest.;Theories of activation emphasize the campaign's information role. In this view campaigns provide information that helps voters connect policy preferences to vote choice, but scholars continue to debate voters' capacity to learn and use this information (especially among less sophisticated voters). In the second essay I compare dynamic perceptions of candidates' policy positions across the spectrum of voter sophistication. On highly publicized policies, learning concentrates among the less sophisticated voters because their more sophisticated peers have already reached a saturation point. Changes in the 'knowledge gap' on a specific issue depends on the prominence of that issue in campaign discourse and the timing of observations.;In the third essay I examine how contextual salience of an issue and its personal importance to certain voters jointly produce heterogeneous reductions in the cost of learning. Using the case of Social Security policy in the 2000 and 2004 campaigns, I demonstrate that learning among seniors outpaces that of non-seniors in 2000 when the campaign prominently featured this issue. The learning difference vanishes four years later when the issue received far less attention. In the earlier election, seniors in the middle-to-low stratum of sophistication were especially likely to learn, challenging the notion that highly sophisticated voters most reap the rewards of campaign information.