Incomplete understanding of the connection between campaign spending and election outcomes has hindered evaluation of enacted and proposed congressional campaign finance reforms. Reanalysis of the 1972 and 1974 House and Senate campaign spending data using both OLS and 2SLS regression models shows that spending by challengers has a much greater impact on the outcome than does spending by incumbents. A similar analysis of the effects of spending on voters' recall of candidates in the 1972 and 1974 SRC surveys supports the explanation that campaign expenditures buy nonincumbents the necessary voter recognition already enjoyed by incumbents prior to the campaign. The 1974 survey questions on Senate candidates indicate that, although the inability to remember candidates' names does not preclude having opinions about them, voters recalling candidates are much more likely to offer evaluative comments, and these more frequently refer to candidates personally. Aware voters offer more negative as well as positive evaluations (though positive outnumber negative); familiarity is not automatically advantageous. And voters' evaluations of candidates strongly influence how they vote. The implications of these findings for congressional campaign finance policy are readily apparent.

No abstract

Analysis of both district-level and aggregate time-series data from postwar House elections supports the thesis that strategic political elites play a pivotal role in translating national conditions into election results and therefore in holding members of Congress collectively accountable for the government's performance. More high-quality candidates run when prospects appear to favor their party; they also win significantly more votes and victories than other candidates in equivalent circumstances. Thus, strategic career decisions both reflect and enhance national partisan tides. The electoral importance of strategic politicians has grown over time in tandem with the trend toward candidate-centered electoral politics. This has rendered the effects of national forces less automatic, more contingent, thus threatening the capacity of elections to enforce some degree of collective responsibility.

No abstract

No abstract

No abstract

Scholars were quick to exploit the opportunities fo~ analysis offered by the new campaign finance data.Most of the early work focused on the question of how campaign spending influences election outcomes and the companion question of what determines how much money candidates raise and spend. A variety of models of campaign spending effects were proposed and examined. All took election outcomes to be a function of campaign spending and other variables but differed in how the outcome and spending variables were measured and in selection of control variables. These differences were reflected in several basic disagreements:i. On functional form. Proposed alternatives included linear functions (Shepard, 1977; Welch, 1974; Jacobson, 1976); semi-log (Welch, 1976); double-log (Welch 1974;Lott and Warner, 1974;Bental, Ben-Zion, and Moshel, 1977); and quadratic (Silberman and Yochum, 1978) for the money-votes relationship.2. On the measurement of relevant variables. For example, most models attempt to control the "normal vote"--the distribution of partisans--in a state or district. This has been approximated by party registration figures (Shepard, 1977; Welch, 1981;Caldeira and Patterson, 1982); district-level presidential votes (Welch, 1974(Welch, , 1981Silberman and Yochum, 1978), and previous votes for the same office or candidate (Jacobson, 1976(Jacobson, , 1978.3. On how the effects of incumbency are to be assessed. Incumbency is known to have a major impact on congressional election results and so, obviously, has to be taken into account. Most early models entered incumbency as a dummy variable (Welch, 1974; Jacobson, 1976;Shepard, 1977); others measured it as years in office, either logged (Giertz and Sullivan, 1977) or in quadratic form (Silberman and Yochum, 1978) to acknowledge diminishing returns.These early models generally overlooked the potential interaction between incumbency status and campaign-spending effects. But there was ample reason to expect that a given amount of spending would have different electoral effects for incumbents, challengers, and candidates for open seats. Later work confirmed that it did and indicated that the marginal returns on spending are much greater for challengers and other nonincumbents than for incumbents (Glantz, Abramowitz, and Burkhart, 1976; Jacobson, 1978;Silberman and Yochum, 1978).

No abstract