This study shows that a simple, empirically derived rule accounts well for the outcome of multi-ballot, majority-rule conventions in the period 1848-1948 and that the same rule would have enabled one to predict correctly the outcome of all contested presidential nominations since 1952 before any candidate had achieved majority support in the Associated Press'polls of delegates. The rule's success suggests two conclusions about the behavior of delegates: (1) that their guesses about outcomes are based on objective cues, similarly interpreted, and (2) that the trend in voting for candidates, in addition to the level of support they attract, is an important element of such cues. The study's findings also suggest that delegates in recent years do not contrast sharply with their predecessors in their predilection for bandwagon voting.

Past individual-level studies of economic voting (1) have incorrectly operationalized the model they employ by using past-oriented rather than future-oriented questions and (2) have failed to examine the level of economic voting in United States Senate elections. Using the 1978 National Election Study, we show that economic voting exists in Senate but not House elections, presumably due to the differences in electoral context. Even when economic voting occurs, however, there is no guarantee that the public will influence the direction of macroeconomic policy.

This study explores the hypothesis that voting in response to economic problems is policy-oriented: voters concerned about unemployment ore predicted to give greater support to Democratic candidates, while those concerned about inflation are predicted to vote more Republican. In light of evidence from previous research, this study investigates the electoral effects of inflation and unemployment as (1) problems directly experienced by the individual, and (2) problems deemed serious for the nation as a whole. Support is strongest for the unemployment side of the hypothesis. Voters personally affected by unemployment gave a modest boost to Democratic candidates in virtually every election. And in years of high unemployment the large percentage of voters who fell it was a serious national problem voted heavily Democratic as well. This study concludes by discussing the important implications these findings have for our understanding of how economic conditions influence voting behavior in American national elections.