In this essay, I start to sketch a research agenda about an electoral backlash against women's descriptive representation. Few scholars have considered the possibility of such a backlash. By “backlash,” I mean resistance to attempts to change the status quo. As Jane Mansbridge and Shauna Shames argue in this issue, when actors disadvantaged by the status quo work to enact change, they may be met by a reaction by those seeking to maintain existing power arrangements. In the following pages, I introduce the idea of a backlash against women's representation, propose several preliminary hypotheses about a backlash, and discuss ways of testing them.
A backlash against women's descriptive representation may seem unlikely. We typically assume that increases in women's representation are inevitable and that gains for women will always beget further gains. The conventional wisdom is that as more women run for and hold elective office, the public becomes more accustomed to women's officeholding, voters have additional opportunities to learn about and experience women's leadership firsthand, and they are less likely to rely on gender stereotypes. Thus, increases in women's descriptive representation and historic firsts for women should lead to further gains. Indeed, the general trend is one of increased women's representation over time, with record highs for women in the U.S. Congress and in state legislatures in 2008 (CAWP 2008b). Arguably, voter attitudes ceased to be a barrier to women's candidacies decades ago, with incumbency and the scarcity of women candidates the remaining reasons for women's continued descriptive underrepresentation (Burrell Reference Burrell1994; Darcy, Welch, and Clark Reference Darcy, Welch and Clark1994; Seltzer, Newman, and Leighton Reference Seltzer, Newman and Leighton1997).
Could gains in women's descriptive representation yield a backlash in the electorate in some cases, rather than always fueling more support for women? Although further liberalization in attitudes about women's representation is probably more likely, there are reasons to believe that such a backlash could occur. For example, the 1992 “Year of the Woman,” in which a record number of women—mostly Democrats—won seats in Congress, was followed in 1994 by what some termed the “Year of the Angry White Male,” in which white men increased their support for the Republican Party. The rise of women in state legislatures stagnated between 1999 and 2006 (CAWP 2008b)—the reason for which is unclear. More recently, observers noted misogyny in some media coverage and public responses to Senator Hillary Clinton's campaign for the Democratic nomination.Footnote 1
Despite widespread public support for women candidates and the similar success rates of men and women candidates, challenges remain. Voters hold gender stereotypes about the personality traits of politicians that tend to advantage male candidates (Huddy and Terkildsen Reference Huddy and Terkildsen1993; Kahn Reference Kahn1996). Women's ability to win office seems to depend on place: The pattern of women's officeholding is uneven across the United States (Palmer and Simon Reference Palmer and Simon2008; Sanbonmatsu Reference Sanbonmatsu2006). Meanwhile, according to a 2006 national survey, the public prefers that greater numbers of women hold office than currently do, but the public also prefers that men constitute 60% of elected officials (Dolan and Sanbonmatsu forthcoming). A recent poll found that nearly half of registered voters would like to see more women in positions such as governor or member of Congress, but about 40% think that the current number of women in those positions is about right (“Breaking the Glass Ceiling” 2008).
A backlash against women's descriptive representation could take one of several forms. Mansbridge and Shames argue that a backlash is a reaction to a group's loss of power or threat of loss. Because candidacy and officeholding are male-dominated roles, opposition to changes in women's roles could explain a backlash against women in electoral politics (Eagly and Diekman Reference Eagly, Diekman, Dovidio, Glick and Rudman2005). Men and socially conservative women might feel a loss of status or power in response to the women's movement—including the movement to elect more women to office. And because gender intersects with other categories, such as race/ethnicity and sexuality, backlash may be greater when members of particular subgroups of women are elected. Electoral gains by doubly disadvantaged women, such as women of color or lesbian, bisexual, or transgendered women, may create more voter opposition than white, heterosexual women might encounter. The reason for backlash might concern an individual's personal gender identity, rather than a group's loss of power or status. Women leaders violate the traditional female stereotype of passivity; by definition, women politicians as a group are agentic—a typical male trait—rather than passive (Huddy and Terkildsen Reference Huddy and Terkildsen1993). Individuals may feel a threat to their personal identity as more women hold office. Backlash may reward individuals because enforcing gender norms may enhance self-esteem (Rudman and Fairchild Reference Rudman and Fairchild2004).
Backlash could be directed at individual women, rather than at women politicians as a group. For example, voters may react negatively toward a woman who violates gender norms by exhibiting masculine personality traits (Eagly, Makhijani, and Klonsky Reference Eagly, Makhijani and Klonsky1992; Rudman Reference Rudman1998); at the same time, because most politicians are male, women politicians could suffer a negative reaction if they display very feminine traits. A woman politician who pursues a strong feminist agenda could suffer backlash as well if voters believe that she should forsake women's collective interests. It is now common for women candidates to “run as women,” rather than avoid identification with women's issues (Witt, Paget, and Matthews Reference Witt, Paget and Matthews1994). However, running as a woman could lead to backlash, depending on voter attitudes toward women and women politicians.
Alternatively, backlash could stem from policy positions and political opposition to women politicians as a group. Voters who are opposed to the agenda of women candidates and officeholders may react negatively to the election of greater numbers of women. Studies typically reveal gender differences in the policy priorities and issue positions of candidates and officeholders—particularly on “women's issues.” Women's issues are usually defined either as issues traditionally associated with women and the private sphere, such as children's issues, or as issues that affect women as a group, such as women's rights issues (Carroll Reference Carroll1994). Because of gender stereotypes, voters might assume that a woman candidate will differ from her male counterparts—regardless of whether her issue positions differ. Voters may also assume that women will only represent women's issues. Thus, the election of more women could mobilize those opposed to a women's political agenda—whether that be an expectation of a women's agenda or an actual women's agenda.
Changes in the climate toward women in general or toward the women's movement could affect the public's reaction to women's descriptive representation. As Donald Haider-Markel (Reference Haider-Markel2007, 122) posits, “general social, political, and economic victories, not just electoral victories, for a previously marginalized group might lead to backlash against the group in a variety of venues.” Therefore, a backlash could have broader origins than changes in the level of women's representation.
Thus far, I have discussed women's descriptive representation in general terms as the election of more women. The election of more women is a departure from the status quo because women are underrepresented in office compared to their numbers in the population. Today, women are 16.3% of members of Congress and 23.7% of state legislators (CAWP 2008b). Eight women serve as governors, and 23.8% of all statewide elective executives are women. In no state have women ever constituted a majority of state legislators. Women have served as governor in only 22 states (CAWP 2008a).
A departure from the status quo with respect to women's representation is more complex than the election of “more” women, however. Progress in women's candidacies or officeholding could take a variety of forms because there are different levels and types of office. At least three types of changes or challenges to the status quo might lead to backlash: numbers, rate or size of change, and office type. Any of these changes could foster greater support for women candidates. But these changes could have the opposite effect.
Drawing on Rosabeth Moss Kanter's (Reference Kanter1977a; Reference Kanter1977b) research about gender ratios within organizations, scholars have debated whether a “critical mass” of women legislators yields better substantive representation of women (e.g., Thomas Reference Thomas1994). However, a critical mass of women could have the opposite effect because a larger presence of women could stimulate a backlash (Bratton Reference Bratton2002). Numbers may also matter for understanding the electorate's response to a change in women's officeholding. A very simple hypothesis is that any increase in the number of women candidates or officeholders could spark a backlash. However, the proportion may be more important than raw numbers. We simply do not know how voters would react to a state legislature or Congress that was half female, majority female, or even supermajority female. We might hypothesize that gender ratios similar to those investigated within legislatures, such as 25%, 30%, or 40%, could prompt an electoral response.Footnote 2 More broadly, the presence of women candidates or election of women to a large proportion of offices, such as statewide offices, might lead to backlash.
Even if the overall number of women officeholders remains low, a substantial increase or perception of rapid change in women's officeholding nationally or within a state or locality could lead to a reaction (Yoder Reference Yoder1991). In other words, the number of women may be less important than the rate of change.
Another change to the status quo concerns the entry of women into new offices. Richard Fox and Zoe Oxley (Reference Fox and Oxley2003), for example, found that women tend to run for statewide offices that are more consistent with traditional female roles. The election of women to new offices, or historic bids for office—such as Clinton's competitive bid for the presidency—might lead to backlash.
The status quo with respect to women's representation varies by place. Thus, women's descriptive representation may be more likely to lead to backlash in places where the status quo includes few women. In a state where women have already broken most barriers in politics, we might not expect backlash. For example, the election of a woman governor in a state that has previously had a woman governor does not represent a challenge to the status quo. Then again, backlash may be more likely to occur in those states where women have made the most progress in electoral politics because women could win a larger number of offices in those states.Footnote 3
In order to test these hypotheses about backlash, researchers could pursue a variety of strategies, including an analysis of public opinion. Voter backlash could be detected in public preferences over women's descriptive representation. For example, a decline in support for electing “more” women to office or an increase in support for majority male officeholding might indicate voter backlash—although such attitudes may also simply ebb and flow with changing levels of women's representation. A panel design would be ideal for this type of study. One problem with this approach, however, is that few public opinion surveys have queried the public on their preferences toward the level of women's officeholding.
One could also look for evidence of backlash in the success rates and vote share of women candidates. A change in women's representation might be followed by a decline in women's success rates or vote share. Or, studies of electoral behavior could show changes in the relationship between voters’ demographic characteristics or policy preferences and support for women candidates. In any analysis of public attitudes and election results, subgroup analysis would be necessary. For example, if our hypotheses are about particular groups of voters, one would look for evidence of backlash among those groups—be they men, voters with traditional attitudes about women, those politically opposed to women's interests, and so on.
Subgroup analysis would be important because lack of change in aggregate election results might obscure the existence of backlash. For example, we might observe stability or stagnation in women's electability and assume the absence of backlash. However, we might not observe aggregate change if the election of more women increases support among some groups of voters but increases opposition among other groups of voters. Subgroup analysis would also be necessary to determine if backlash is directed at certain groups of women candidates.
Evidence of backlash might be found in the level of women's officeholding and the emergence of women candidates. If the electoral climate facing women candidates becomes less favorable because of backlash, fewer women might seek office. We would therefore want to investigate any changes in candidate emergence. The emergence of women candidates and their electoral success are shaped by a wide range of factors related to both candidate supply and electoral demand. Thus, past levels of representation, contextual factors such as voter ideology and electoral rules, and institutions such as political parties would need to be taken into account.
Aggregate analysis of changes in women's officeholding patterns would also be a fruitful way to determine if voter backlash occurred. One could look at specific offices and the types of offices that women candidates are —or are not—contesting. Hypotheses could be developed about longitudinal effects as well, because backlash may occur gradually and may not be evident in the short term (Hughes and Paxton Reference Hughes and Paxton2007; McConnaughy Reference McConnaughy2007). Scholars could test for backlash by looking back at the historical record and patterns of women's officeholding.
Due to space constraints, I have outlined only one area for future research. But other areas of American politics warrant investigation. For example, one area of backlash research concerns the emergence and influence of the countermovement of socially conservative women opposed to the modern women's movement (Klatch Reference Klatch1987; Mansbridge Reference Mansbridge1986; Schreiber Reference Schreiber2008) and the subsequent alliance of organized feminism with the Democratic Party and organized antifeminism with the Republican Party (Freeman Reference Freeman, Fainsod Katzenstein and Mueller1987; Sanbonmatsu Reference Sanbonmatsu2002; Wolbrecht Reference Wolbrecht2000). More research is needed on the ability of politicians and political parties to court backlash.
Another area of research has tested the conventional wisdom that electing more women to office will yield greater substantive representation of women (Bratton Reference Bratton2002; Crowley Reference Crowley2004; Kathlene Reference Kathlene1994; Smooth Reference Smooth2001). If electing greater numbers of women yields a backlash within legislatures—either in terms of male hostility toward women legislators or policy resistance to women's interests—then increases in descriptive representation may not yield the expected gains in women's substantive representation. Women legislators may not pursue women's representation if they anticipate backlash from their colleagues and change their behavior accordingly. Such research will be difficult to conduct because it requires the ability to distinguish between sincere and strategic behavior.
In sum, the idea of gender backlash has received little attention in the American politics field. We typically assume that progress for women is inevitable. Yet without theoretical refinement of the backlash concept and empirical tests, we will not know if backlash occurs.
