9631640 Snyder In prior research these investigators began an empirical investigation of the productivity of campaign spending in House elections. Some early results from this work indicate that the "vote production functions" facing U.S. House incumbents and challengers are remarkably similar, and are not statistically distinct. This finding contrasts sharply with the conventional wisdom, which holds that incumbent spending is far less productive than that of challengers. The results suggest that the central issue in the study of campaign finance is not simply how money translates into votes but why some candidates are able to raise vastly more than others. In the current investigation the researchers proceed along two tracks. First, they validate their initial investigation of the effects of campaign expenditures in incumbent-contested U.S. House elections. To validate these estimates they extend their analysis to U.S. Senate races and to open seat U.S. House races. They also improve their measurement techniques through the use of LEXIS/NEXIS and are thus allowed to make the most complete and accurate data set of who runs for Congress. Second, they investigate how the productivity of campaign spending varies across different kinds of candidates, across electoral settings (such as congressional district, state, and media market), and over the course of political careers. Political lore and public opinion research frequently emphasizes that the effectiveness of a campaign depends very much on the context. With only a few exceptions, research on campaign finance has ignored these contextual factors, though there has bee much speculation. Among the specific hypotheses, the marginal product of spending is thought to be lower (1) if the candidate is already well-known, (2) if the campaign is conducted in an expensive media market, (3) if the challenger is running for the second time, (4) if the challenger is a political novice, and (5) if the electorate is composed largely of vote rs of the opposite party of the candidate. Using the instrumental variables approach presented in their earlier work, they test whether the productivity of campaign spending money varies in any of these ways. While the statistics are relatively straightforward, additional data collection is undertaken. This project produces three highly useful new data sets for use by other scholars: one on the personal and professional backgrounds of congressional candidates; one on the locations, resources, and activities of business and other interest groups; and, one tracking the histories of important federal legislation of the last two decades. ***
