Conventional wisdom long held that there was a bias against women in elections. Subsequent research indicates that men and women who challenge for elective office confront a common barrier: incumbency. In this article we extend our previous research on women in open seat elections by examining the performance of women who compete in special elections. Female candidate emergence in special House contests is slightly higher than in regular open seats. Using multivariate regression models, our analysis uncovered no bias against women in special elections. Overall, the performance of women in special elections and open seats indicates that disruptions of the political status quo by the sudden vacancy creates opportunities that women exploit with effectiveness, although the low level of female candidate emergence limits the growth of descriptive female representation.
A sustained argument in the debate over representation centers on the existence of sexual bias in congressional elections. Although some studies indicate an historical bias against women, other researchers have found these biases due largely due to incumbency which discriminates against all challengers (Cook, Thomas, and Wilcox 1994; Darcy and Choike 1986). More recently, Gaddie and Bullock (1995) found no bias against women in open seat congressional elections. In this article, we extend our previous research on sex in open seats to ascertain whether bias is evident in a littlestudied category of congressional election: the special election. The political dynamics of special elections in the U.S. are compared to previous findings in regularly scheduled open seat contests to determine whether the competitiveness of women in regular open seats carries into special elections. These elections often receive little attention, even from the constituents of the vacated district. Still, while forty-eight elections across fourteen years may seem to be a small population, it is important to consider that special elections constitute less than 1.5 percent of congressional contests between 1981 and 1995, while producing nearly 10 percent of new House members during that period. During this time only 154 members of Congress came to the House by defeating incumbents, while in 2585 instances the incumbent was returned. Of the 508 first-time members of Congress elected between 1981 and 1995, two-thirds were the product of regular open seat contests.
Patterns of female candidate competitiveness are examined in 48 special House elections held from 1981 to 1995. A multivariate model is specified to determine whether a significant sexual bias occurs in special elections, and those results are compared to our previous analysis, expanded to 282 open seat races by the addition of the 1994 elections (Gaddie and Bullock 1995). We then discuss the implications of our findings for understanding female competitiveness in congressional elections.
SPECIAL ELECTIONS, OPEN SEATS, AND WOMEN AS CANDIDATES
Sigelman (1981) argued that special elections were special precisely because they guaranteed sudden, unexpected changes in representation. Although the level of partisan change is not greater in special elections than in regularly occurring open seats, the change in legislators represents a departure from the past. Sigelman (1981) and Studlar and Sigelman (1987) indicated that special election victors tend to win by comfortable margins, although special elections are still more closely contested than the preceding general election. When party change occurred, it usually involved a loss by the incumbent president's party Studlar and Sigelman (1987) found that special elections were manifestations of the normal partisan forces operating in general elections. None of this work examined the role of candidate sex in these elections.
Should we expect a sexual bias in special elections? Historically, special elections prompted by the death of legislators have …