The Origins of Critical Mass and “Critical Mass Theory”

To advocate giving up on critical mass is neither an insignificant nor a purely academic concern: It runs the very real risk of leaving supporters of women's representation without a crucial tool for increasing the number of women elected to political office. A closer look at the academic literature, however, reveals that it has been interpreted and applied in many different ways over the last 20 years. In her seminal contributions to these debates, for example, Rosabeth Moss Kanter outlines three expectations regarding increased numbers of women in corporate life: 1) With an increase in their relative numbers, women are potentially allies, can form coalitions, and can affect the culture of the group; 2) with an increase in their relative numbers, women can begin to become individuals differentiated from one another; and 3) with an increase in their absolute numbers, despite few changes in their relative numbers, women can develop supportive alliances and affect the culture of the group so long as they are “feminist” or “women-identified-women” (Kanter 1977a, 966; 1977b, 238). Bringing these insights into the field of women and politics, Drude Dahlerup (1988) tests the first conjecture but ultimately uses her evidence to reject the notion of critical mass in favor of “critical acts,” on the grounds that relative proportions of women seem to be less important than individual policy entrepreneurs in explaining women-friendly policy gains.

Following Dahlerup, many scholars reduce Kanter's three conjectures into one, anticipating that increased numbers will facilitate coalitions among women. At the same time, nearly all present Dahlerup's work as if she had made a strong case in favor of critical mass, when she, in fact, argues for focusing on critical acts. We label these misrepresentations “critical mass theory,” an approach to the study of women's legislative behavior that distorts the work of Kanter and Dahlerup but—through strong consistencies across accounts—plays a central role in structuring research on women and politics. Its applications concentrate exclusively on opportunities for women to form coalitions with one another, finding that legislatures with high proportions of women introduce and pass more bills on women's issues than do their female counterparts in low-representation legislatures (Saint-Germain 1989; Thomas 1994). In contrast, its critiques emerge from cases where policy change does not occur with increased numbers of women (Childs 2004; Rosenthal 1998), or where policy reforms take place despite relatively small numbers of women (Reingold 2000; Towns 2003). Thus, applications adopt Kanter's first claim and Dahlerup's notion of critical mass, while critiques present evidence that corroborates Kanter's second and third claims and Dahlerup's idea of critical acts (Childs and Krook 2005). This disjuncture between the origins of critical mass and how it is represented in critical mass theory suggests a need to revisit relations between numbers and outcomes in research on women's political representation.

Dimensions of the Substantive Representation of Women

Turning our backs on critical mass theory, we propose two means for rethinking this debate: 1) reformulating the central research question from when women make a difference to how the substantive representation of women occurs, and 2) refocusing the investigation from the macro level (i.e., what do “women” do?) to the micro level (i.e., what do specific women do?). Combined, these two moves open up a series of new possibilities for analyzing legislative behavior by relaxing overly restrictive analytical frames regarding the form and content of “acting for women.” To facilitate the first shift, we identify five dimensions that merit further consideration when analyzing women's substantive representation.

Anticipated effects of increased proportions of women in political office. As already noted, the most common assumption is that as women grow more numerous in legislative chambers, they will be increasingly able to form strategic coalitions with one another in order to promote legislation related to women's concerns (Thomas 1994). Existing research, however, presents at least four other scenarios that might prove to be more compelling starting points for analyzing individual cases: A rise in the number of women may influence men's behavior in a feminist direction, causing both male and female legislators to pay more attention to women's issues (Bratton 2005; Flammang 1985); the increased presence of women may provoke a backlash among male legislators, who may employ a range of tactics to obstruct women's policy initiatives and keep them outside positions of power (Hawkesworth 2003; Heath, Schwindt-Bayer, and Taylor-Robinson 2005); a lower proportion of women may be more effective than a higher number because female legislators may be able to specialize in women's concerns without appearing to undermine male domination (Crowley 2004; Dodson and Carroll 1991); and a rise in the overall number of women may result in the election of an increasingly more diverse group who may or may not be interested in pursuing women's issues, either because their priorities lie elsewhere or because they believe that other female legislators will continue to lobby on behalf of women (Carroll 2001; Schwindt-Bayer 2004).

Constraining and enabling characteristics of legislative contexts. The literature similarly points to a multiplicity of factors that limit and enhance opportunities for women to translate policy preferences into legislative initiatives on behalf of women. Much of this research focuses on institutional rules and norms that reflect a bias toward men's experiences and authority (Hawkesworth 2003; Kathlene 1995), and as such, compel women to conform to existing masculine legislative practices in ways that undermine their ability to integrate women's perspectives into public policymaking (Carroll 2001). Many scholars also note the impact of party affiliation and ideology on women's legislative activities, observing that mechanisms of candidate selection, combined with pressures for party discipline, determine what kinds of women are elected, as well as the specific policy positions that they are likely to take once they accede to political office (Cowley and Childs 2003; Gotell and Brodie 1991). Others, however, draw attention to institutional norms that facilitate women's participation (Chaney forthcoming), including the presence of women's caucuses and women's policy machineries (Thomas 1994; Weldon 2002), and point out that some party ideologies offer greater opportunities for women to pursue feminist policy concerns (Reingold 2000; Swers 2002).

Identities and interests of female and male legislators. Nearly all work on women's substantive representation addresses individual features that facilitate and undermine cooperation among women in political office. While many political theorists aim to discern or define a shared perspective among women as a group in order to justify calls for their increased political presence, most empirical studies stress divisions among women—like race, class, age, and party affiliation—that prevent the formulation of a collective legislative agenda (Dodson and Carroll 1991; Swers 2002). Indeed, some argue that identity categories like “women” are inherently exclusionary and serve to reify one difference while erasing and obscuring others (Carroll 2001). Further, “gender” is not a prepolitical and fixed identity that women bring with them when they enter politics, but one that is partially produced and reproduced within the context of particular legislatures (Towns 2003; Whip 1991). Others question the elision of women's bodies with feminist minds, on the grounds that being female may matter less than “gender consciousness” does for achieving feminist outcomes (Childs 2004; Reingold 2000).

Feminist and nonfeminist definitions of women's issues. These variations and ambiguities are reflected further in competing definitions of “women's issues.” Scholars adopt various approaches that include policies that increase the autonomy and well-being of women (Bratton 2005; Wängnerud 2000); concerns that belong to the private sphere according to established views on gender relations (Meyer 2003); areas where surveys discover a gender gap in the population (Schwindt-Bayer 2004); and any issues of concern to the broader society (Dolan and Ford 1995). As such, some prefer feminist definitions that focus on role change for women through increases in autonomy and the scope for personal choice (Dodson and Carroll 1991; Reingold 2000), while others opt for more inclusive ones that capture a broader range of issues affecting women's everyday lives (Swers 2002). Yet others favor definitions based on the concerns articulated by women's movements at various moments in time, which allow women's issues to remain a priori undefined, context-related, and subject to evolution (Celis 2004), as well as a collective product that emerges as women interact with other women to identify their priorities (Weldon 2002).

Stable and contingent features of policymaking processes. The possibility to achieve gains for women depends closely on features of the policymaking process, which influence how and when women's issues reach the legislative agenda, as well as their ultimate prospects for being passed into law. Most research focuses on stable features, finding that women as a group have the greatest impact—or, more specifically, tend to differ most from men—in terms of setting the legislative agenda and proposing new bills that address issues of concern to women (Childs 2004; Swers 2004). For this reason, many criticize studies that focus exclusively on legislative voting, because these assume that enactment is the most important stage of the policymaking process (Tamerius 1995). While some suggest that the best solution is to examine the entire legislative process (Carroll 2001; Swers 2002), others point out that policymaking often involves numerous elements of contingency that make such models appear overly simplistic. On the one hand, complex combinations of actors—often in series of chance events—are generally responsible for moving an issue to agenda prominence and gaining its passage (Childs and Withey 2006; cf. Kingdon 1984). On the other hand, policy innovations rarely proceed in a vacuum, because policy cycles and demonstration effects strongly condition which issues enter and are kept off of legislative agendas, separate from any assumed prerequisites for change (Bratton and Ray 2002).

From Critical Mass to “Critical Actors”

Examining these five dimensions reveals multiple facets within what might be considered the substantive representation of women, opening up a range of possibilities for defining what acting for women might constitute in various places and times. They do not in themselves explain, however, how changes in the numbers of women in political assemblies translate more concretely into specific policy outcomes. To unravel these processes, we turn to studies in other fields that also employ the notion of critical mass to analyze collective behavior. We find that they shift focus from macro-level patterns to micro-level actions in order to theorize the contingent calculations of individuals within groups that may explain why similar distributions of preferences do not always translate into comparable collective outcomes. To take one example, studies of thresholds and human behavior frame individual decisions to act in terms of the proportion of others who also choose to act, on the grounds that the costs and benefits of an actor making a certain choice depend in part on how many others decide to make that choice (Schelling 1978). For subjective reasons, however, individuals view these costs and benefits differently and thus vary in terms of how many others need to act before they do so themselves: “Radicals” have low thresholds, as they will act even if many others do not, while “conservatives” have high thresholds, as they will not act unless many others do so as well. Because thresholds are situation-specific, two groups may be identical in composition but still experience distinct outcomes due to the particular process of aggregation. As such, collective results do not necessarily offer an accurate reflection of what individuals would do in every situation (Granovetter 1978).

Many of these ideas resonate with the original formulations of Kanter and Dahlerup, who point to diversity among women and the importance of individuals who resolve to act on behalf of women as a group. Adapting them, we argue for distinguishing between “critical actors” and “critical mass” in order to identify the concrete representatives—not vague imperatives of “sex” or “gender”—who put in motion individual and collective campaigns for women-friendly policy change. “Critical actors” are those who initiate policy proposals on their own, even when women form a small minority, and embolden others to take steps to promote policies for women, regardless of the proportion of female representatives. Indeed, they may not even be women; in some situations, individual men may play a crucial role in advancing women's policy concerns (Celis 2004; Tamerius 1995). Their common feature is their relatively low threshold for political action: Although they may hold attitudes similar to those of other representatives, they are much more motivated than others to initiate women-friendly policy reforms (Childs and Withey 2006). Although they may operate alone, they may also stimulate others to act, setting in motion a momentum for policy change, or alternatively, provoking a backlash among those opposed to fundamental reform. As such, their shape and impact are not absolute, as smaller numbers of women may join together in legislative caucuses to promote common goals with great success, while larger numbers may enhance the opportunity for critical acts but may also foil their effects.

Conclusion

For at least two decades, the concept of critical mass has provided optimism to scholars, activists, and politicians interested in promoting both the descriptive and substantive representation of women. In the eyes of most, the entry of a few women to political office has not yet resulted in the widespread feminization of politics because there are simply not enough of them to “make a difference.” The solution, therefore, has been to argue for measures to achieve a “critical mass” of women who could, in turn, press for women-friendly policy change. As these claims have grown increasingly more popular in the policy world, however, a growing amount of scholarly research has cast their validity into doubt.

Given these opposing pulls, should feminists give up on critical mass? We think both yes and no. On the one hand, we believe that critical mass theory should be firmly discarded: Expecting an automatic change once women attain a certain proportion of seats is theoretically dubious and, perhaps even more seriously, undermines the case for women's increased presence if existing female politicians appear to be failing women as a group. On the other hand, we recognize that the concept of critical mass has proven extremely useful in making concrete gains in the “real world,” as it insists that a few token individuals are not sufficient for provoking large-scale policy change. To reconcile these theoretical and practical concerns, we propose 1) opening up the research question to explore the various actors, strategies, and outcomes consistent with the substantive representation of women across space and over time, and 2) focusing on “critical actors” and their role in pursuing policy change, either alone or together with others, as a more precise theoretical and practical strategy for understanding which kinds of women—and men—are most likely to represent women's concerns in political office.