Explaining Movement and Countermovement Events in the Contemporary U.S

Explaining Movement and Countermovement Events in the Contemporary U.S. Women's Movement

Lee Ann Banaszak (The Pennsylvania State University)

and

Heather Ondercin (Louisiana State University)

Presented at the American Political Science Association Meetings, September 2-5, 2010, Washington D.C. The authors would like to thank Christina Wollbrecht, Debra Minkoff, Sarah Soule and the other principal investigators of the Dynamics of Collective Action data for their willingness to share their data. Naturally any errors are the responsibility of the authors. Introduction

While feminist movements have been the subject of careful scholarly research, there has been less attention to the study of countermovements – defined as those movements that arise in opposition to feminist movements (but see Schreiber 2008; Staggenborg and Meyer; Staggenborg 1989, 1991). Nonetheless, both pundits and scholars of the women’s movement have often talked about the role that anti-feminist movements played in reducing the impact of the women’s movement1 starting in the 1980s (see for example Faludi 1991; Marshall 1984; Ryan 1992; Schreiber 2008). Such movements are not just important in their own right but also because they engage in such interactive struggles with feminist movements, affecting women’s movements’ mobilization and activity as well as the resulting public policy (Staggenborg and Meyer; Staggenborg 1989, 1991). Women’s movements often end up mobilizing to oppose the demands of countermovements and may even end up incorporating opposing countermovement goals, such as the rollback of certain policy achievements, in their own goals as well. Thus, where countermovements develop, movements and countermovements become entwined in a careful dance with countermovements responding to the mobilization and success of existing social movements (Zald and Useem; Staggenborg and Meyer; etc.), and social movements then responding to the activities of counter movements (Freeman 1983; Gamson 1975; Staggenborg).

The interaction between movements and counterparts also likely influences public opinion and public policy, especially given that some scholars have already found ties between movement activity on the one hand and public opinion or legislation on the other(Banaszak 1996a; Banaszak and Ondercin 2009; Soule et al. ;Burstein 1999). However, much of this literature finds only weak relationships between movement mobilization and political outcomes (Burstein and Linton 2002). But if movement mobilization is also a response to countermovement activity, the weak relationships may be a result of countermovement influence. Thus, we need to understand the dynamic relationship between movements and countermovements to clearly connect movement activity to policy and public opinion. Specifically, policy change may not just the result of mobilization of a movement but rather the net of movement and counter movement activities (Jasper and Poulsen 1993; Rosenfeld and Ward 1991; Zald and Useem 1987). However, beyond theoretically discussions of the nature of this interaction (Lo 1982; Meyer and Staggenborg 1996) or careful qualitative assessments of movement-countermovement

1 We utilize here the term women's movement as a synonym for feminist movement because this is the language of feminists themselves. However, we acknowledge that the alternative definition of women’s movements as those movements “whose definition, content, leadership, development or issues are specific to women and their gender identity” would place both the movements and the countermovements discussed here into the category of women’s movements (Banaszak, Beckwith and Rucht 2003, p. 2; Beckwith 2000).

relations in particular cases (Dixon 2008; Jasper and Poulsen 1993; Staggenborg 1991), the endogenous relationship between social movements and countermovements has gone largely unexplored. 2

This paper takes a step in the direction of separating out the effects of movements and countermovements, as well as the relationship between the two, using quantitative event data and time series analysis to examine the causes of feminist and anti-feminist mobilization (in the form of public events). In addition to analyzing movement-countermovement interactions for the larger feminist movement in the United States, we also examine two specific campaigns associated with these movements: abortion and the Equal Rights Amendment. Both campaigns have historically been associated with strong movement-countermovement interaction, and been important issues in the contemporary feminist movement. We expect to find a strong endogenous relationship between movement and counter movement activities, but also explore how other factors external to the movements -- such as political opportunities, gender opportunities, and previous policy successes– influence movement and countermovement events.

The paper proceeds as follows. We begin by discussing the role of countermovements against the U.S. women’s movement, particularly the cases of the Equal Rights Amendment and abortion. We then put this particular movement-countermovement relationship in a wider perspective by examining the scholarship on countermovements more generally. These discussions provide the hypotheses which we test using a data set on the U.S. women’s movement and its countermovement comprised of movement and countermovement time series, combined with additional measures – all of which is described in our section on data and methods. In the final two sections of the paper we analyze this data using autoregressive distributed lag models and discuss the implications of this analysis for our understanding of countermovements.

Battling for Control on Issues of Gender

Histories of the second wave of the U.S. women’s movement3 differ in the way they draw the timeline of the mobilization of the second wave (cf. Ryan 1992 and Freeman 1975 with Roth 2004 and Thompson 2002), yet most indicate that the movement mobilized in the late 1960s and early 1970s. As women’s movement scholars have suggested, that initial mobilization, chronicled by coverage of the events in newspapers and magazines of the time, was initially divided into two somewhat different (albeit overlapping) sectors (Freeman 1975; Ryan 1992; Marx Ferree and Hess). The first is often described as a small group sector, which developed among college students and others who had been active in the civil

2 The few exceptions include Andrews (2002) analysis of reactions to the civil rights movement in Mississippi and Werum and Winders (2001) on anti-gay activism; as well as a body of work on the interactions between movements and countermovements in making claims and framing their positions (Fetner 2001; .McCright and Dunlap 2000).

3 I utilize the term second wave here to acknowledge the historical continuities of the women’s movement in the United States (see Banaszak 2010, Chapter 2; Rupp 1985; Rupp and Taylor 1987).

rights, anti-war or New Left movements of the early 1960s. The events staged by this wing of the women’s movement were often smaller in size but dramatic in nature and employed such tactics as guerrilla theater or dramatic protest. One example might be the 1968 and 1969 demonstrations against the Atlantic City Miss America Pageant where feminists protested the objectification of women by shouting protests in the event, tossing high heels and bras into a trash can and staging guerilla theater on the boardwalk (Freeman n.d.). Despite the drama of these events, small group sector events often do not garner media coverage, perhaps in part because they mobilize a relatively small number of people, presenting a message that is provocative. For example, no contemporary coverage of the Atlantic City events occurred in the New York Times although it was mentioned in later articles by feminists and letters to the editor. The second sector – what Freeman (1975) labels the older, more established sector of the contemporary women’s movement-- was more focused on policy change. Much of the action of this sector occurred in the form of lobbying, testifying before Congress or behind the scenes work on changing laws and policies. When this sector mobilized public events, it was often in the form of larger demonstrations, pickets, and boycotts demanding the enforcement of existing laws or seeking the creation of additional ones. One such example might be the series of pickets of EEOC offices that the National Organization of Women and other groups engaged in to protest the Commission’s ruling (eventually reversed) that separate help wanted ads were not a violation of Title VII law. These protests often garnered more press coverage. For example, the EEOC pickets were covered in multiple New York Times stories including a longer story in the Sunday Magazine section (see Bender 1968; Weinman Lear 1968). 4

Taken together these two sectors led to an overall increase in the mobilization of women into the women’s movement and an increase in the number of women’s movement events reported in the newspaper. The later is illustrated by Figure 1, which reports the number of movement and countermovement events in each year that were reported in the New York Times5. As Figure 1 shows the number of women’s movement events reported in the New York Times rose steeply in the late 1960s and peaked around 1975. Despite a drop off in the number of events after 1981, the movement continued to coordinate collective action events even into the 1990s at a level that far exceeded the last years of the 1960s.

(Figure 1 about here)

Moreover, the events covered by the media in these later years increasingly focused on two issues: abortion and reproductive rights generally, and the Equal Rights Amendment. While the movement

4 In fact, the New York Times had covered the issue of how the sex clause of Title VII would be implemented even before it became a political issue, although much of that coverage focused on its implication for employers.

5 The data were taken from the Dynamics of Collective Action data set available at http://www.stanford.edu/group/collectiveaction/cgi-bin/drupal/. The authors thank Doug McAdam, John McCarthy, Susan Olzak and Sarah Soule for making these data available. Naturally all errors in the analysis are the fault of the authors.

remained diverse in its organizational forms, purposes, and ideology, the public face of the movement chronicled in the newspapers was dominated by these larger campaigns. The Equal Rights Amendment campaign occurred in two phases: the first was a protracted battle, ending in 1972 but starting in the 1920s, to have Congress adopt the amendment (see Banaszak 1996; Rupp and Taylor). Even though the Senate approved the amendment in 1946 by a bare majority, it would not be until the issue became the focus of the re-mobilized women’s movement that it would pass Congress in 1972. As Figure 2 shows, women’s movement mobilization around the Equal Rights Amendment rose sharply in 1971, just prior to the debate in Congress. The second phase was the period that began with Congressional ratification and ended in 1982, when the deadline given for states to ratify the amendment expired. That deadline had already been extended once, when the original ratification deadline occurred and supporters were only three states shy of ratification. Pro-ERA events show the attempts to influence the ratification process. Local maxima in the number of events occur in 1978 – the year before the initial ratification deadline – and then peaked in 1982, the year the time extension ended. Although a few scattered events occurred after 1982 as feminists tried to revive the Equal Rights Amendment, by 1990 the issue was clearly not in the public eye.

(Figure 2 about here)

The issue of abortion, on the other hand, has continued to inspire mobilization by both opponents and the women’s movement. Indeed, while abortion initially constituted a very small proportion of feminist events, as Figure 3 shows, the number of feminist events related to abortion has risen over the years. A comparison of Figure1 and Figure 3 also suggests that increasingly the feminist agenda is dominated by the issue of abortion; while abortion events constituted only about 15% of all feminist events reported in the New York Times in the 1970s, by the 1990s abortion events were a half of all events. Mobilization around the decriminalization of abortion laws or greater access to abortions began prior to the 1973 Roe v. Wade decision, although these events were relatively few in number. Much of the early push for abortion decriminalization came from doctors concerned about the lack of clarity in existing law and lawyers and legal scholars seeking to rationalize state legal codes (Luker 1984; Staggenborg 1991). In 1962, for example, the American Law Institute (ALI) opened up a new wave of reform by calling for new legislation. As the women’s movement mobilized, feminists increasingly took up the cause of abortion (Luker 1984 Chapter 3; stateline.org n.d.).

(Figure 3 about here)

Many of the policy effects of the women’s movement occurred relatively early in its mobilization process. The Equal Pay Act was adopted in 1963 and in 1964 Title VII of the Civil Rights Act of 1964 prohibited discrimination on the basis of sex, and made sexual discrimination the purview of the Equal Employment Opportunity Commission. When Title IX of the Education Amendments of 1972 passed a few years later, the National Organization for Women had just celebrated its fifth anniversary. Two events were particularly important for the mobilization of a countermovement. First, in March 1972, the Senate approved the Equal Rights Amendment, marking the beginning of a campaign for ratification among the fifty states. Second, in January 1973, the Supreme Court handed down the Roe v. Wade 4

decision, which stated that a woman had a right to seek an abortion without state intervention during the first trimester. Naturally both events were preceded by considerably activity among activists. Before the decision on Roe v. Wade, 12 states had adopted the ALI model law and another four states had adopted legislation which allowed a woman to have an abortion whenever she or her doctor thought it necessary (stateline.org). Many of these policies were of course also engendered by feminist activity and lobbying (see for example Banaszak 2010 Chap. 6).

Thus, feminist activity started in the late 1960s, peaked in the 1970s and early 1980s but never totally disappeared. Indeed, the number of feminist events related to the issue of abortion continues to rise over the entire time period although it is cyclical in nature. Events related to the ratification of an Equal Rights Amendment were also an important part of women’s movement events in the 1970s and early 1980s but then disappeared totally from the agenda. How then do the activities of those opposed to the feminist agenda compare?

Countering the women’s movement.

Many scholars of the women’s movement acknowledge that the second wave’s periods of high mobilization and policy change led to organized attempts to protest the movement or roll back its policy achievements. As Figure 1 shows, anti-feminist actions began to rise in the mid-1970s (see also Ryan 1992: 56 and Marshall 1984). As was the case with the women’s movement itself, some of the countermovement activity was localized activity by small group. For example, Ryan (1992:68) notes that groups supporting traditional gender roles such as Protect our Women (POW) organized in the mid- 1970s. On the other hand national groups opposed to the women’s movement also developed in the 1970s. For example, Beverly LaHaye founded Concerned Women for America in 1979 after watching an interview with Betty Friedan (Schreiber 2008: 30).

Two particular victories for the women’s movement served to mobilize the bulk of countermovement activism. The first was the adoption of the Equal Rights Amendment (ERA) by Congress in 1972. Facing the push for ratification of the ERA, Phyllis Schafly’s Stop ERA was founded in 1972. By that year, thirty of the necessary 38 states had already ratified amendment . Stop ERA organized anti-amendment ratification lobbies in several states including Illinois and Florida. The organization was perhaps best known for its lobbying tactics including the distribution of homemade bread and jam to legislators by women asking them to vote against ratification (Klemesrud 1975). After the defeat of the Equal Rights Amendment, Stop ERA demobilized although it continues to fight attempts to revive the amendment.

The second important event was the increasing legalization of abortion, culminating in the Roe v. Wade decision of January 1973. By the 1970s, over 2/3’s of the anti-feminist events in Figure 1 were pro-Life events, with the proportion rising even more in subsequent decades. While the Roe v. Wade decision mobilized many opponents to abortion, local anti-abortion organizations had already begun forming in the late 1960s (Luker 1984; Staggenborg 1991). The National Right to Life Committee was organized soon thereafter in 1971 (Staggenborg 1991: 35). Initial pro-Life reaction to Roe v. Wade

focused on reducing the availability of abortion through legislation like the Hyde Amendment, but soon it also moved into the streets, and began engaging in confrontational and violent tactics. Much of the violence including clinic bombings and arson attempts peaked in the 1980s and early 1990s (Staggenborg 199: 132-133; National Abortion Federation n.d.) but picketing and clinic blockades continued after 2000 (National Abortion Federation n.d.). Pro-Choice activists responded by mobilizing clinic defense and securing additional legislation designed to limit the impact of these countermovement tactics.

This brief description of the U.S. women’s movement and the anti-feminist movements that arose in their wake helps to illustrate the way that both movements are intertwined. It also shows that much of that interaction occurred within two issue areas of the women’s movement: equal rights and abortion. Before we move to a statistical analysis of these two movements, we need to examine theories of countermovement development and movement-countermovement interaction.

Understanding Movement and Countermovement Events

Studies of movements and countermovements have generally seen two different types of effects that influence the mobilization of these two groups. In the section below we focus first on how movements influence countermovements and vice versa; we then focus on factors that affect both women’s movements and their countermovements, specifically the resources, political opportunities and gender opportunities available to the two movements. While much of the literature has focused on the interactive claims or framing of movements and countermovements, we focus here on a specific dependent variable: the mobilization of public events.

Interactive Influences between Movements and Countermovements

As is inherent in their name, countermovements arise in the first place to counter existing movements. Hence, much of the theoretical literature on countermovements focuses on the influence of movements on countermovements and the effect that countermovements can have on the movements they are fighting (Zald and Useem 1987; Meyer and Staggenborg 1996; Lo 1982). This interactive feature can take two forms: each type of movement can react to the actions of the other or each may react to the success of the other.

Several scholars (Andrews 2002; Zald and Useem 1987: 247-248; Lo; Meyer and Staggenborg) argue that it is largely movement events, particularly visible events that create the conditions that mobilize countermovements. Movement events mobilize countermovements by creating grievances as an existing population begins to realize that what they had considered widely accepted positions and practices are being challenged. As movements begin to focus on specific issues they also make it easier for countermovements to mobilize, around the same specific set of goals or issues. Finally, the actions of movements demonstrate concretely that collective action can have an effect, if not specifically on policy at least in capturing the public eye. Hence we expect that:

H1: An increase in movement events causes an increase in countermovement events.

Alternatively, we might expect that countermovements react less to movement mobilization than to a movement’s ability to create changes in policy (Meyer and Staggonberg 1996; see also Zald and Useem 1987)6. In this case, opposition to the movement mobilizes as movements gain policy successes. This theory suggests it is not mobilization of the movement but actual policy change that creates the grievances and suggests opportunities for change. Unfortunately, it is difficult to determine final policy successes easily. Here we utilize the passage of legislation in Congress to measure the degree of policy success, but recognize that such a measure provides a limited view. 7 In particular, policies that are likely to be implemented are more dangerous than those being considered (Andrews 2002). We anticipate that as legislation is actually passed by Congress, this increases countermovement activity, because Congressional legislation even once adopted can be repealed or modified. Thus we expect that,

H2: The more feminist legislation adopted by Congress, the more anti-feminist groups will mobilize.

While we have focused on the effect of movements on countermovements, the literature on movement-countermovement relationships also notes that countermovements have strong effects on the movements that spawn them. In particular, countermovements are likely to mobilize the original movements as they realize that their positions and policy successes are being challenged. Thus, countermovements create new or revitalize existing grievances in movements which will likely lead to an increase in mobilization as well (Fetner 2001). Hence we expect:

H3: An increase in countermovement events will lead to an increase in movement events.

In some senses, this expectation directly contradicts another existing theory in the social movement literature that success breeds demobilization. This idea, made explicit in Zald and Ash’s seminal 1966 piece on organizational growth and decay, argues that when movement organizations are successful they must either take on new goals or demobilize. 8 Zald and Ash suggest that:

H4: the adoption of movement legislation leads to a decrease in movement events.

If we accept the previous three hypotheses as true however, we are faced with a serious contradiction. If movement success breeds both countermovements and movement demobilization, then we are not

6 McCammon et al. (2001) places movement policy successes, like the passage of suffrage, into the category of gender opportunities. While we agree that movements do affect both political and gender opportunities, we find it more useful to separate out those opportunities directly and immediately influenced by the movement from those that are more removed from movement influence in time and are less directly an outcome of movement activity.

7 We are currently collecting information on Supreme court decisions and executive orders as well.

8 Indeed, Zald and Ash (1966: 333-334) hypothesize that movement organizations that are a result of coalitions of different groups, and those which aim at changing society are more likely to dissipate after success.

likely to find that movements mobilize in reaction to the rise of countermovements. Staggenborg (1991) addresses this dilemma directly in her analysis of the pro-Choice movement, arguing that feminist pro- Choice activists were well aware of the rising pro-Life movement and that therefore Roe v. Wade did create the same incentives to demobilize as it might otherwise. Below we examine her analysis directly and explore it also in the case of the Equal Rights Amendment and the women’s movement more generally.

Responding to Opportunities.

In addition to the hypotheses that suggest that movement events affect countermovement events and vice versa, we also have expectations about how both of these types of events are affected by the opportunities available in the wider political system. There are two types of time-varying political context variables that we examine in this model. The first type is the access each type of movement has to the political system and the second set of variables are gender opportunities that influence specifically women’s movements (and women’s countermovements) ability to mobilize.

Political Opportunities. Political opportunities are openings in the political context that movements may take advantage of to gain the outcomes they seek. When movements observe changes in the political system that are to their advantage, they strategically utilize these openings by altering their activities. In our particular case, we expect movements to increase the number of public events when they believe that there are openings in the political system. Here we focus on one specific but important opportunity for women’s movements – whether they have political allies in government. One well-established opportunity for women’s movements has been the growing affinity between the Democratic party and women’s movements (Freeman 1987; Wolbrecht 2000). This affinity developed as the women’s movement arose in the late 1960s and early 1970s; prior to that time neither party had a stronger connection to the women’s movement.

While there has been a lot of focus on the political opportunities of movements themselves, there have been fewer pieces that focus particularly on the political opportunities affecting countermovements (but see Meyer and Staggenborg 1996 and Staggenborg 1991). Since countermovements are simply a specific type of movement, they are also affected by the existence of political allies in the federal government. The growing affinity between the Democratic Party and the women’s movement was matched by an increasing kinship between the Republican Party and traditional movements of all types, including anti-feminist movements. This means that we expect anti-feminist movements to find opportunities when Republicans are in power.

Even with a focus only on political allies, we should note that the American political system provides multiple arenas where such allies can be found. In particular, the federal system provides multiple opportunities for women’s movements (or anti-feminist movements) to develop political allies (Elman, Banaszak 2003; Meyer and Staggenborg 1996). Local political opportunities provide a haven in periods when political elites on the national scene oppose the movement. Hence, even if political allies are lacking on the national scene we expect that increasing numbers of local allies provide opportunities for 8

either movements or countermovements to mobilize. A few authors have argued that this opportunity is more likely to benefit countermovments because it prevents movements which have acquired policy successes from permanently closing the window to changing policy (Meyer and Staggenborg 1996: 1637). For example, Hausmann (2005) has argued that federalism was particularly important in the anti- abortion movements in the United States. This leads us to the following hypotheses:

H5: Periods where the Democrats are in power at the national level will have increases in the number of movement events.

H6: Periods where the Republicans are in power at the national level will have increases in the number of countermovement events.

H7: The higher the number of state legislatures that are democratic controlled the greater the number of movement events.

H8: The lower the number of state legislatures that are democratic controlled the greater the number of countermovement events.

Other social movements may also be part of the political opportunities available to movements. Sympathetic social movements create openings in the political structure by making the existing political alliances unstable. They also provide resources to sympathetic social movements that make it easier to mobilize. On the other hand, opposition movements sway potentially sympathetic political allies from supporting a social movement, and provide resources for countermovements. Here we look specifically at the role of conservative movements in influencing the dynamics of feminist and anti-feminist movements. Anti-feminist countermovements are part of a wider field of social movements that have included the rise of religious fundamentalism, as well as anti-tax and anti-immigrant activism. We expect that:

H9: The rise in the number of religious movement events increases the number of anti-feminist events.

Gender Opportunities. Equally important for women’s movements are opportunities related specifically to the gendered nature of the political system. Changing gender relations influence the opportunities available to women. As McCammon et al (2001: 53) points out men and women traditionally were viewed as having separate spheres with men dominating the workplace and politics. Gender opportunities for the women’s movement change as women increasingly enter these areas, blurring the traditional gender division. Our expectation is that many of the same gender opportunities that increase women’s movement activism in the form of events will also inspire the activism of opposition movements as well. In particular, anti-feminist movements may be reacting less to the activity of

women’s movements or to their ability to alter concrete policies than to the changing relation between the sexes more generally which increases the grievances that mobilize the movement.

In this paper we focus on three specific factors designed to capture the changing gender opportunity structure. First and foremost, we examine explicitly how public opinion about gender roles, equal rights for women, and abortion influences the mobilization of the movement and countermovement. One gender opportunity is the changing social norms and beliefs that underlie the general public. Movements may mobilize to capitalize on and as a result of changing attitudes toward women. Thus we expect that :

H10: Increases in support for feminist values in the general public increases the number of women’s movement events.

Countermovements may also react to these changes, mobilizing to oppose the status changes that come with changing attitudes. As Lo (1982) points out, however, changing values also increase the status discrepancy for many individuals. Rising feminist values may lead women in traditional positions to experience a loss of status as the values that supported homemaking as women’s calling disappear (see also Klatch 1987; Luker 1984). Hence we also expect that:

H11: Increases in support for feminist values in the general public increases the number of countermovement events

However, even if attitudes are slow to change, the movement of women in the arenas traditionally held by men may also prove to be gender opportunities. As women enter the workforce and politics, they have more opportunities to change the political agenda, at the same time as they may experience increased grievances regarding women’s status. In addition, women’s presence in business and politics may increase the willingness of male political elites to consider feminist demands. Finally, women in political offices may also act as insiders within the political system pushing change from within (Banaszak 2010; Santoro and McGuire 1997). Alternatively, we expect countermovements to benefit from traditional social arrangements since these not only reinforce the importance of traditional roles among political elites but also increase the pool of traditional women who can be mobilized (Klatch 1987). Thus, we expect that

H12: Rising participation in the workforce of women increases the number of women’s movement events. H13: Rising participation in the workforce of women decreases the number of countermovement events. H14: Increasing numbers of women in Congress increases the number of women’s movement events . and

H15: Increasing numbers of women in Congress decreases the number of countermovement events .

We examine each of these hypotheses in three different venues of the women’s movement. We begin by looking at all of the events related to the U.S. women’s movement and the countermovement that arose in reaction to it. In addition, we explore the specific campaigns surrounding the Equal Rights Amendment and the issue of abortion. Each of these campaigns encompass important issues for the women’s movement, yet as the discussion above suggests the history and dynamics of each is considerably different. The Equal Rights Amendment campaign involved a relatively short campaign that ended in a defeat for the women’s movement, while the abortion campaign has been a protracted campaign that began with a significant victory for feminists with attempts by pro-Life activists to pick away at the existing policy. However, before we discuss the results, we turn to a description of the data used in this paper and discuss the time series methods we employed.

Data and Methods

Our analysis is conducted on quarterly data ranging from the first quarter of 1960 to fourth quarter of 1995.9 We use an autoregressive distributed lag (ADL) model to examine the relationship between movement and counter-movement events. The basic feature of an ADL is the inclusion of lags of the dependent variable on the right had side of the model. The lags of the dependent variables model the persistence of a series or how long changes influence future values. Additionally, the lags control for auto-correlation in the errors. The number of lags specified depends on the dynamics of the dependent variable; we used the minimum lag specification that gives us well behaved errors.

We model six different dependent variables: all women’s movement events, all events that oppose the women’s movement, pro-Choice events, anti-abortion movement events, pro-ERA movement events, and anti-ERA movement events. The event data are taken from the Dynamics of Collective Action data set which collected all collective action events from 1960 to 1995 the daily editions of the New York Times.10 The event data used here was created by aggregating by quarter the 891 events which made claims related to women’s movements or abortion; these events were then coded by whether they supported or opposed feminist positions. We recognize that this data set does not represent an unbiased collection of all movement events11, but nonetheless believe that it is important to examine

9 Data availability reduces the time frame for several of our models. Our measure for the passage of feminist legislation ends in 1988, significantly reducing the series. The data used in the abortion events analysis range from the first quarter of 1962 to the fourth quarter of 1995 and the ERA from the second quarter of 1972 to the fourth quarter of 1995.

10 For a extensive description of the data collection process, see Walker, Martin, and McCarthy (2008) pp. 45-48 and Earl, Soule and McCarthy 2003.

separately the events that achieve this high level of visibility since they are more likely to inform the public debate about these issues and to influence attitudes of the public and/or policy makers.

We are primarily interested the relationship between the women’s movement and its countermovement. As a result we include the event measure for the corresponding movement in the model. For example, when modeling the general women’s movement events we include the measure of general women’s countermovement events in the model. Additionally, our models include controls for political opportunities, gender opportunities, and women’s movement success .

We use three different measures of political opportunities: mobilization of the religious right, party control of state legislatures, and party control of the president. Our measure of religious right mobilization comes from the same Dynamics of Collective Action data set and represents the number of events where religious claims were made aggregated by quarter. Party control of the state legislature is the number of state legislatures under the control of the Democratic Party. Party control of the presidency is a dummy variable series with 1 indicating that Democrat is in the White House.12

We measure gender opportunities using three different measures designed to capture both political and social opportunities. First, we use a public opinion measure that is specific to the particular type of events we use as a dependent measure. When modeling all women’s movement and countermovement events we use a general measure of gender attitudes. When modeling the abortion and ERA movements and counter-movements we use measures of public opinion on abortion and ERA respectively. These measures were created using Stimson’s dyadic dimensional algorithm that allows us to combine multiple survey questions into a single measure of public opinion over time (Stimson 1991, Ondercin 2007). Second, we control for women’s increased workforce participation by using the ratio of women’s workforce participation to men’s workforce participation.13 The third measure of gender opportunities is a count of the number of women serving the U.S. Congress provided by Center for American Women and Politics (CAWP 2010).

To examine the effect of movement success on movements and countermovements , we use the number of pieces of feminist legislation which passed either the House or the Senate. We calculated

11 The extensive literature on the bias of event data gathered from newspapers suggests that our list of events are likely to represent large, dramatic events involving physical violence, formal organizations or elites involved in policy making (McCarthy, McPhail, and Smith 1996; Oliver and Myers 1999; Oliver and Maney 2000; Ortiz et al. 2005, p. 399).

12 We would have liked to include a measure indicating party control of the U.S. Congress. However the Democrats control the Senate and Democrats control the House for all but three sessions during the time period of our study. Thus we exclude the measure due to lack of variation.

13Men’s and Women’s workforce participation are the seasonal adjusted civilian labor force participation rates for individuals over the age of 16 (women’s series id LNS11300002Q, men’s series id LNS11300001Q). Data can be downloaded from the Bureau of Labor Statistics: http://www.bls.gov/data/#employment.

this measure from the Christina Wolbrecht’s (2000) roll call data on feminist legislation 14. Date is only available until 1988, so inclusion of the variables reduces the length of our time series.

We standardize all of the variables in the model by calculating their z-scores. Standardizing each variable also places all of the variables on the same metric allowing for easier comparison. Stimson, for example, argues that “the virtue of standardization is that it forces measures to discriminate within series over time rather than between series.” (Stimson 1991, 54). Additionally, standardizing each variable’s time series transforms them so that they theoretically range from negative infinity to infinity, thus adhering to the common assumptions in time series analysis. In the analyses reported below, we also provide different lag specifications for each variable. We determine each variable’s lag specification for each of the independent variable by testing alternative lag lengths. Significance tests and AIC information criteria were used to select the lag specification.

In time series analysis each of the variables need to be stationary in order to complete a multivariate analyses. If a variable or set of variables are not stationary then the analysis runs the risk of obtaining spurious regression results. Moreover, tests of significance are no longer valid because one cannot apply the usual assumptions about underlying probabilities in calculating estimates or standard errors (Nathaniel 2003, Chatfeild 1996). We ran both the Augmented Dickey Fuller and KPSS tests to assess the stationary of each of the variables in the model. The Augmented Dickey Fuller tests if a series has a unit root or is non-stationary. Alternatively, the KPSS tests if the series is stationary. Both tests have relatively low power resulting in us rejecting the null too often, thereby sometimes providing conflicting results. In addition to using these tests we also examined the auto-correlation functions and partial auto-correlation functions. The event series, general movement events, general counter-movement events, pro-abortion events, anti-abortion events, pro-ERA events, anti-ERA events, and the pro- religious right events all appear stationary. As a result we model these series in their levels form. The public opinion series, party control of the state houses, party control of the presidency, women’s workforce participation, women’s representation in Congress and the passage of feminist legislation appear non-stationary. As a result we use the first difference of the series in all the models. The first differences of these series are stationary. The drawback to modeling series in their first differences is that we can only talk about short run effects.

Analysis

Theoretically we expect to find an endogenous relationship between movements and countermovements. Time series analysis provides us with tools to test for endogeneity and then model the series to account for an endogenous relationship. To test for endogeneity we use the Granger Causality tests (Freeman 1983, Granger 1974). Granger causality ask the basic question: After controlling for its own history, how does the history of another variable contribute to our ability to predict the current (and future) values of itself? Granger Causality has two limitations. First, it cannot test if two series are contemporaneously correlated. Second, it may not detect endogeneity between

14 We especially thank Christina Wollbrecht for providing these data from her book!

two variables if the two variables react to one another at a faster rate than we observe the data. The results of the Granger Causality tests, not reported here, indicate a lack of endogentity for our movement and countermovment variables. Because we do not find evidence of endogeneity between movements and countermovements we proceed with a basic autoregressive distributed lag model and test for a contemporaneous relations between the variables.

We begin by examining the dynamics of movement events. Table 1 through 3 presents a set of models with dependent variables of general movement events, pro-ERA movement events and pro-abortion movement events. Because of the reduced time period that our measure of social movement success, passage of feminist legislation, is available we present two models for each of the dependent variables. The first model contains controls for political and gendered opportunities. Then the second model adds the control of feminist legislation to political and gendered opportunities.

(Tables 1, 2 and 3 about here)

Our first observation is that movements are indeed responding to countermovements and movements are quick to respond. Looking at the model without the control for feminist legislation first we find a positive and significant relationship between movement events and countermovement events. The relationship is contemporaneous for both the general pro-women’s movement series and the pro- abortion series. For the anti-ERA events we observe the relationship at one lag, meaning that anti-ERA events in the previous quarter influence pro-ERA in the current time period. Most likely our finding of the contemporaneous relationship for general women’s movement events and pro-abortion events is a result of our unit of analysis not being fine enough to capture the speed at which these movements are responding to countermovement activity.

The finding that movements are responding to countermovements holds even when we control for the passage of feminist legislation for both the pro-ERA series and the pro-abortion. The general anti- women’s rights events variable is no longer a significant predictor of general women’s rights movement activity once we add in the control for the passage of feminist legislation. However, this appears to be a result of losing information because of the shorter time period. In results not reported here, we ran the base model, restricting the time period to the limited series we have for the passage of feminist legislationtest to see if the relationship is still significant. We found that even without the control for the passage of feminist legislation variable, the effect of countermovement events is not significant in this model; however, the coefficients are similar.

We find mixed support for hypothesis 4 which states we expect a decrease in movement activity when the movement experiences success. The coefficient for the passage of feminist legislation is negative and significant only for the pro-abortion series. Thus in the case of the abortion movement success does appear to cause demobilization. In the model of all women’s movement events there is also a negative coefficient for the passage of feminist legislation; however, the coefficient does not reach traditional levels of significance. We find contradictory evidence for the pro-ERA series, where we see a positive coefficient. While the coefficient fails to reach traditional levels of significance we should not

immediately dismiss this relationship. The ERA model is estimated with only 51 time points meaning there is not a lot of information going into the calculations and we fall just short of significance (.11). However, it is actually not surprising that we might find a difference for the ERA campaign since the passage of legislation is most likely passage of the ERA amendment and the extension of the deadline for ratification. Thus we would expect these bills to mobilize the movement for ratification in the states.

Political opportunities appear to have little effect on movement activities. We expected that democratic presidents would result in increased movement activity because the Democratic Party is often seen as supporting movement goals (hypothesis 5). However, our results suggest that democratic presidents may actually reduce the number of movement events. For the models of all women’s movement events and of pro-ERA events, the coefficient for change in party control of the president is negative, but is not significant. In the pro-abortion models the coefficient for changes in party of the president is negative and significant; indicating that when we change to a Democrat in control of the White House we actually see less pro-Choice movement events.

We expect that when more state legislatures are under democratic control we will also see greater movement activity (hypothesis 7). In all but one of the models we find positive relationship between democratic control of state legislatures and women’s movement events, however, none of these relationships are significant. When we add the control of passage of feminist legislation to the pro- abortion model the sign on party control of the state legislature flips and becomes negative, but is still not significant.

We expected that the rise of increased activity in the religious right movement would decrease feminist movement activity. We find support for this expectation in that all the signs for the coefficient on the religious rights event series are negative across all the models. However, in none of the models does the variable reach traditional levels of statistical significance.

While we found little support for political opportunities shaping movement activities, we do see gender opportunities playing a role affecting the number of movement events. We expected to see increased in feminist event activity when public opinion increasingly reflected support for feminist ideas (hypothesis 10). We find support for this hypothesis for the women’s movement in general, the pro-Choice campaign, and the pro-ERA time series. In the models of all women’s movement events we see a significant and positive relationship between changes in gender attitudes and women’s movement events. This relationship is observed at the 3rd and 4th lags indicating that it takes a while for the women’s movement to sense the changes gender attitudes. We also observe a positive and significant relationship between changes in attitudes about abortion and pro-abortion events at the fourth lag. However this relationship is no longer significant once we control for the passage of feminist legislation. There is, however, a positive and significant contemporaneous relationship between changes in attitudes about the Equal Rights Amendment and pro-ERA events. An additional effect with a lag at quarter 3 in the model suggests that there may be an additional slower effect of ERA attitudes on movement events as well.

We also expected to see increased workforce participation among women to increase the number of events in the women’s movement (hypothesis 12). Overall we find strong support for this hypothesis. In the models of general women’s movement activity we find a positive and significant relationship between changes in women’s workforce participation and women’s movement events at the second lag, indicating the impact of the changes in workforce participation is only felt after two quarters. However, the relationship is no longer significant when we include controls for the passage of feminist legislation. The relationship between women’s workforce participation and pro-Choice events is slightly more complicated. We observe a negative but insignificant relationship between contemporaneous changes in women’s workforce participation in the model without the control for the passage of feminist legislation. When we include the control for the passage of feminist legislation the sign on this coefficient flips signs but remains insignificant. However, there is a positive and significant relationship between changes in women’s workforce participation lagged one quarter and pro-Choice events that remains significant with the inclusion of the policy success measure. There is also a positive relationship between changes in women’s workforce participation at the second lag and pro-ERA movement events. The relationship is insignificant without the passage of feminist legislation control, but becomes significant when we control for passage of feminist legislation.

Finally, we expect increases in women’s representation will result in increases women’s events (hypotheses 14). We find mixed and weak results for this hypothesis. For both the model of all women’s movement events and the pro-ERA series we find a positive relationship between changes in women’s representation and movement activity. However, these relationships fail to reach traditional levels of significance. With the pro-abortion movement models we find a positive and significant relationship at the first lag of changes in women’s representation. We also find a negative and significant relationship between the second lag of changes in women’s representation and the pro- abortion movement. The sign on the second lag flips and both the changes at the first and second lag are no longer significant when we include the passage of feminist legislation measure.

We now turn our attention to examining the influence movements have on countermovement activity. Models of countermovement activity are presented in Tables 4-6, with anti-women’s movement events in Table 4, anti-ERA events in Table 5, and Table 6 contains the models for anti-abortion events series. Once again we present two models for each of the dependent variables; one without a control for the movement success, passage of feminist legislation, and one with the measure of movement success.

(Tables 4, 5, and 6 about here)

As we discussed above we expected and found that movements respond to countermovements, we also expect that countermovements will respond to movements. However, the results suggest that countermovement activity has a larger influence on movement behavior than movement behavior has on countermovements. Looking at the models without the control for movement success we see some evidence that countermovements are responding to movement activity. The model of all anti-women’s rights events is positively and contemporaneously related to the general women’s rights movement series. Additionally, the anti-ERA movement responds to movement events albeit at a slightly slower 16

speed; anti-ERA events are positively and significantly related to pro-ERA events which occurred one quarter earlier. While the sign on the coefficient is in the predicted direction, our model of anti- abortion events shows that pro-Choice events had little effect. Thus we find only some preliminary support for hypothesis 3, that countermovements respond to movement activity.

An alternative hypothesis however was that countermovements may not be responding to the movement activity, but to the success of women’s movements (hypothesis 2). We find strong support that countermovements are mobilizing in response to movement success. Across all three of the dependent variables, all anti-women’s movement events, anti-abortion events, and anti-ERA events, we find a positive and significant relationship between the passage of feminist legislation and countermovement events. Interestingly the relationships between countermovement events and movement events are no longer significant after controlling for movement success. Thus, it appears that movement success is more important in determining countermovement activity than the actual activities associated with the movement.

We also examine the influence of political opportunities on countermovement activity by exploring the relationship between party control of the presidency, party control of the state legislatures and religious right movement events. In hypothesis 6 we state that a greater number of countermovement activities should take place under republican presidents than under democratic presidents. For two of our dependent variables, all anti-women’s movement events and anti-abortion events, the coefficients are signed in the opposite direction of what we predicted, suggesting that there were more events under democratic presidents. However, these coefficients do not reach statistical significance. The coefficient in the models of anti-ERA events is signed in the predicted direction, but is not statistically significant.

The second political opportunity we examined is party control of the state legislature. We expect that countermovements would increase their activity the fewer state legislatures were under democratic control (hypothesis 8). The results of the general model of all anti-women’s movement events fits with our expectation, larger numbers of democratic state legislatures decrease countermovement activity. Conversely, lower numbers of democratic state legislatures increase countermovement activity. We observe a similar relationship between party control of the state legislature and anti-abortion movement events; however, this coefficient does not reach traditional levels of statistical significance. The results are mixed when we look at the anti-ERA models. The coefficient for party control of the state legislature is positive and insignificant in the model without the control for women’s movement success. When we control for women’s movement success the coefficient for party control of state legislature in the anti-ERA model is now signed in the expected direction but is still insignificant.

The final aspect of political opportunity we examine is the influence of the religious right movement on countermovement activity. We expect the rise of the religious right will increase anti-feminist events (hypothesis 9). We find a positive but insignificant relationship between the religious right movement and our model of all anti-women’s movement events. Surprisingly our strongest relationship is found for the anti-ERA movement series, where we find a positive and significant relationship. The rise in the religious right appears to coincide with a rise in the number of anti-ERA events. The relationship 17

between religious right events and anti-abortion movement events, however, is the opposite direction of what we predicted. The relationship is insignificant, but reaches close to traditional levels of significance in the model without controls for movement success.

We also look at how gender opportunities shape countermovement activity. We begin by looking at the influence of public opinion about women’s roles, the ERA, and abortion on their respective countermovement variables. We expect that as gender attitudes grow more feminist there should be an increasing number of countermovement events (hypothesis 11). Across the three different dependent variables we find moderate support for the hypothesis. For all the dependent variables there is a positive relationship between public opinion and countermovement events. For the model of all anti-women’s movement events the relationship is significant at both the 3rd and 4th lag when we do not control for movement success. When we control for movement success only the 4th lag remains significant. Public opinion about the Equal Rights Amendment is significantly related to anti-ERA countermovement events at the 2nd lag when we don’t control for movement success. The relationship no longer reaches levels of statistical significance when the controls for women’s movement success are added; however, a p-value of .134 suggests there may be some relationship, especially given the smaller sample size of the model. While we observe a positive but insignificant relationship between attitudes on abortion and the anti-abortion event, when we add the controls for success of the women’s movement the relationship flips signs but remains insignificant.

The next gender opportunity we look at is women’s workforce participation, where we expect to find that women’s increased workforce participation decreases countermovement events (hypothesis 13). We find some support for this expectation. Women’s workforce participation is negatively related to both all anti-women’s rights events and anti-abortion events. In the models without a control for movement success, these relationships are in the correct direction but not statistically significant. However, when we control for the passage of feminist legislation, the relationships remain in the expected direction and becomes statistically significant. However, the coefficient on women’s workforce participation is in the opposite direction than what we expected in the analysis of anti-ERA events. Without controls for women’s movement success the relationship between women’s workforce participation and the anti-ERA movement falls just short of traditional levels of statistical significance. Once we add in the controls for movement success the relationship becomes significant.

Our final measure of gender opportunity is the representation of women in Congress. Hypothesis 16 states that we expect to see an increase in women’s representation decreasing countermovement activity. We find little to support this hypothesis; in fact the majority of the models suggest the opposite relationship. The coefficients in the models of all anti-women’s rights events and anti-abortion events are positive and fall just short of reaching statistically significance when we do not control for the passage of feminist. When we add in the control for policy success, the coefficients are still signed opposite of what we expected but are now highly insignificant. Similarly when we do not control for the success of the women’s movement, we observe a positive but insignificant relationship with anti- ERA events. When we add in the control for women’s movement success the coefficient for women’s representation flips signs but remains statistically insignificant. 18

Conclusions

The above analyses provide insight into our understanding of movement-countermovement interactions, the influences on countermovement mobilization, and the nature of different issues within the women’s movement. In the conclusion we discuss each of these in turn.

The literature on movements and countermovements – both theoretical and (largely qualitative) empirical-- suggests that movements and countermovements are tied together in an endogenous interaction where movement events inspire the creation of countermovements, and then they themselves react to countermovement events. Our analysis, which controls for other possible effects, suggests two things about this relationship. First, in our analysis we saw clear effects of countermovement events on movement events particularly in the campaigns for the Equal Rights Amendment and to protect abortion rights. In these cases, across the entire time period of the campaigns movement events were influenced by countermovement events. However, movements had less influence on countermovements. Although inherent in the name of countermovements is the sense that they react to movement events and mobilization, countermovement mobilization was driven by other factors than movement events. Thus, if the movements are engaged in an intricate dance, it appears that countermovements take the lead even if they arrive late to the dance. Thus, once a countermovement arises, it appears to somehow be able to lead the issue, at least in the specific campaigns of the ERA and abortion rights.

Second, while most time series analysis examines endogenous effects of variables by looking at how previous time periods of one variable influence current levels of the other variable, there is a great degree of simultaneity in the interaction of movements and countermovements. Although we tested different lag structures and are looking at quarterly data which is already highly refined compared to most social movement data, we found that where movements react to countermovements and vice versa that reaction occurs in the same time period. Thus, we suspect that the reactions of movements to countermovements and vice versa are very quick – occurring on a daily, weekly, or monthly basis. Particularly in the abortion campaign, this quick reaction is not surprising given that feminist activists are often reacting to clinic violence and protest. One unanswered question, however, is whether the quick reactions we find in the case of the women’s movement are mirrored in other movements. The timing of such reactions deserve more investigation particularly as movement scholars collect more event data: one possibility is that timing varies depending on the movement and its relationship with the countermovement. Some countermovements – perhaps those that are acting in highly confrontational ways or endanger activists in personal ways – may be more likely to inspire immediate response, while other interactions may be more slow. All of this suggests the need to collect more finely grained event data.

The results above also provide additional evidence about the factors that influence countermovement mobilization. The passage of legislation appears more important than the events staged by women’s movement. This suggests that it is not the movements themselves that inspire countermovements but the practical effects that they have. It is important to remember that our event data consists only of 19

those highly visible events of movements – those that registered in a major national newspaper. Hence, we have already selected those events that reach the public sphere. However, countermovements appear less concerned with the effect that movement events have than on the effects that concrete policies will have. Given that women’s movements have occasionally been identified as differentiated from other movements because they focus on change throughout society and not only on policy changes within the state (see van Dyke, Soule and Taylor 2004), the policy changes they acquire nonetheless are important enough aspects to inspire opposition.

The results also suggest that local opportunities are important to anti-feminist movements in ways that are not true for the women’s movement itself. In particular, anti-feminist movement events increased when there were more state legislatures controlled by Republicans but feminist movement mobilization was not affected by Democratic control of state legislatures. This supports the idea that countermovements may have a different set of political opportunities than movements. Because they arise when movements are already better established, they may require pockets of resistance to gain a foothold.

For both movements and countermovements, though, measures of partisan control of the national and state governments were less important than the gender opportunities reflected in public opinion and women’s place in the workforce. Indeed, having a democratic in the White House reduced women’s movement events, suggesting that movements demobilize when there is an ally in office. Of the gender opportunity variables, only the number of women in Congress was insignificant across all six models, a result that surprised us. The lack of significance may reflect that our measure was poor; entrée into Congress is very difficult and the number of women in the institution has changed only slowly over time. Alternatively, it may signal that of the gender opportunities available to the women’s movement, positions in politics are less important than those in the workplace or family. Nonetheless, the overall importance of gender opportunities especially in comparison with our measure of political opportunities suggests that these movement specific opportunities are more significant than those available to all social movements. Changing public opinion and women’s position is vital to the mobilization of women’s movements even as they seek to alter these same factors.

Finally, while we found some similarities among movement and countermovement mobilization for all feminist events, the Equal Rights Amendment events and events related to abortion, there were significant differences as well. We see the effects of countermovement mobilization on movement events most clearly when we focus on specific campaigns, although the analysis of abortion and ERA campaigns both had shorter time-series. In addition, there were significant differences in the lag structures of the independent variables and the sign and significance of variables’ coefficients across the three dependent variables. This suggests that in looking at movement events, particularly those that are highly visible, there may be campaign level factors that are important. In particular, the dynamics of the battle around abortion and the battle around the Equal Rights Amendment differed in a number of respects. Perhaps foremost is that while opposition to the Equal Rights Amendment centered around the ratification campaign in state legislatures, pro-Life proponents have fought their battles in a number of different venues from state legislatures to local clinics. Moreover, because of the ratification 20

deadline, both sides in the ERA battle had a limited timeline in which to fight, while such limitations do not exist in the case of abortion. This suggests that social movement scholars should be considering the role that campaign characteristics play in determining how the battle between movements and countermovements is fought.

Bibliography

Andrews, Kenneth T. (2002). “Movement-Countermovement Dynamics and the Emergence of New Institutions: The Case of "White Flight" Schools in Mississippi.” Social Forces, 80(3) 911-936.

Banaszak, Lee Ann. 2010. The Women’s Movement Inside and Outside the State. Cambridge: Cambridge University Press.

Banaszak, Lee Ann. 2003. “The Women’s Movement and the Constraints of State Reconfiguration: Abortion and Equal Pay in Differing Eras” In Women’s Movements Facing a Reconfigured State edited by Lee Ann Banaszak, Karen Beckwith and Dieter Rucht. (2003: Cambridge: Cambridge University Press). Pp.141-168.

Banaszak, Lee Ann. 1996. "When Waves Collide: Cycles of Protest and the Swiss and American Women’s Movements.” 1996. Political Research Quarterly 49 (December): 837-860.

Banaszak, Lee Ann. 1996a. Why Movements Succeed or Fail: Political Opportunities, Culture and the Struggle for Woman Suffrage. Princeton: Princeton University Press.

Banaszak, Lee Ann and Heather Ondercin. 2009. “Examining Endogeneity in Social Movement Protest and Public Opinion: The Case of the U.S. Women‟s Movement .” Paper presented at the American Political Science Association, September 3-6, 2009, Toronto, Canada.

Banaszak, Lee Ann; Karen Beckwith and Dieter Rucht. 2003. Women’s Movements Facing a Reconfigured State. Cambridge: Cambridge University Press.

Berry, Mary Frances. 1986. Why ERA Failed: Politics, Women’s Rights, and the Amending Process of the Constitution. Bloomington: Indiana University Press.

Burstein, Paul. 1999. “Social Movements and Public Policy.” In Marco Guigni, Doug McAdam, and Charles Tilly, editors. How Social Movements Matter. Minneapolis: University of Minnesota Press. Pp. 3-21.

Burstein, Paul; Rachel Einwohner, and Jocelyn Hollander. 1995. “The Success of Political Movements: A Bargaining Perspective.” In Jenkins, J. Craig and Bert Klandermans,

editors. The Politics of Social Protest: Comparative Perspectives on States and Social Movements. (Minneapolis: University of Minnesota Press) pp. 275-295.

Burstein, Paul and April Linton. 2002. “The Impact of Political Parties, Interest Groups, and Social Movement Organizations on Public Policy.” Social Forces 81(2): 381-408.

CAWP, 2010. “Women in U.S. Congress 2010” Center for American Women and Politics, Eagleton Institute for Politics, Rutgers University. http://www.cawp.rutgers.edu/fast_facts/levels_of_office/Congress-CurrentFacts.php

Chatfield, Chris. 1996. The Analysis of Time Series An Intorduction. London: Chapman and Hall.

Costain, Anne and Steven Majstorovic. 1994. “Congress, Social Movements, and Public Opinion: Multiple Origins of Women’s Rights Legislation.” Political Research Quarterly 47(1): 111- 135.

Dixon, Marc. 2008. “Movements, Countermovements and Policy Adoption: The Case of Right- to-Work Activism.” Social Forces 87(1): 473-

Earl, Jennifer: Sarah A. Soule, and John D. McCarthy. 2003. “Protest under Fire? Explaining Protest Policing.” American Sociological Review 69:581–606.

Faludi, Susan. 1991. Backlash: The Undeclared War on American Women. New York : Crown.

Fetner, Tina. 2001. “Working Anita Bryant: The Impact of Christian Anti-Gay Activism on Lesbian and Gay Movement Claims.” Social Problems, 48(3 ): 411-428.

Freeman, Jo. N.d. “No More Miss America! (1968-1969).” Available at: http://www.jofreeman.com/photos/MissAm1969.html. Accessed on July 12, 2010.

Freeman, John R. 1983 “Granger Causality and the Time Series Analysis of Political Relationships” American Journal of Political Science. 27(2): 327-358.

Guigni, Marco. 2004. Social Protest and Policy Change. Boulder CO: Rowman and Littlefield.

Granger, CWJ and P. Newbold. 1974. “Spurious Regressions in Econometrics” Journal of Econometrics 2: 111-120.

Jasper, James M. and Jane Poulsen. 1993. “Fighting Back: Vulnerabilities, Blunders, and Countermobilization by the Targets in Three Animal Rights Campaigns.” Sociological Forum, 8: 4 (December): 639-657.

King, Brayden G.; Marie Cornwall; and Eric C Dahlin. 2005. “Winning Woman Suffrage One

Step at a Time: Social Movements and the Logic of the Legislative Process.” Social Forces; 83( 3): 1211-1234.

Klatch, Rebecca. 1987. Women of the New Right. Philadelphia: Temple University Press.

Klemesrud, Judy. 1975. “Opponent of E.R.A. Confident of Its Defeat.” New York Times, December 15, 1975 p. 53. Accessed from ProQuest Historical Newspapers, August 5, 2010.

Lo, Clarence Y. H. 1982. “Countermovements and Conservative Movements in the Contemporary U.S.” Annual Review of Sociology, 8:107-134 .

Luker, Kristin. 1984. Abortion and the Politics of Motherhood. Berkeley: University of California.

Marshall, Susan. 1984. “Keep Us on the Pedestal: Women against Feminism in Twentieth- Century America.” In Women: A Feminist Perspective 3rd edition. Jo Freeman editor . Palo Alto: Mayfield Publishing. pp. 568-581.

McCammon, Holly J.; Karen E. Campbell; Ellen M. Granberg; and Christine Mowery. 2001. "How Movements Win: Gendered Opportunity Structures and the State Women's Suffrage Movements, 1866-1919." American Sociological Review 66(1):49-70.

McCright, Aaron M. and Riley E. Dunlap. 2000. “Challenging Global Warming as a Social Problem: An Analysis of the Conservative Movement's Counter-Claims.” Social Problems, 47( 4): 499-522.

Meyer, David S. and Suzanne Staggenborg. 1996. “Movements, Countermovements, and the Structure of Political Opportunity.” American Journal of Sociology, 101:6 (May):1628- 1660.

Nathaniel, Michael. 2003. Unit-Root. In Encyclopedia of Social Science Research Methods, ed Alan Bryman Michael Lewis-Beck & Time Futingo Liao. Thousand Oaks, CA: Sage Publications.

National Abortion Federation. N.d. “NAF Violence and Disruption Statistics. http://www.prochoice.org/about_abortion/violence/violence_statistics.html Accessed August 4, 2010.

National Right to Life Committee. N.d. “Abortion History” http://www.nrlc.org/abortion/timeline1.html Accessed August 5, 2010.

Ondercin, Heather. 2007. The Changing Social Definitions of Men and Women and Their Effect on the Partisan Gender Gap, 1953-2003. PhD. Dissertation: The Pennsylvania State University.

Rosenfeld, Rachel and Kathy Ward. 1991. “The Contemporary U. S. Women's Movement: An Empirical Example of Competition Theory.” Sociological Forum, 6: 3 (September): 471- 500.

Roth, Benita. 2004. Separate Roads to Feminism: Black, Chicana, and White Feminist Movements in America’s Second Wave. Cambridge: Cambridge University Press.

Rupp, Leila J. 1985. “The Women’s Community of the National Woman’s Party, 1945 to the 1960s.” Signs 10(4):715-740.

Rupp, Leila J. And Verta Taylor. 1987. Survival in the Doldrums: The American Women’s Rights Movement, 1945 to the 1960s. New York: Oxford University Press.

Ryan, Barbara. 1992. Feminism and the Women's Movement. New York: Routledge.

Santoro, Wayne A. and McGuire, Gail M. 1997. “Social Movement Insiders: The Impact of Institutional Activists on Affirmative Action and Comparable Worth Policies.” Social Problems 44(4): 503-520.

Schreiber, Ronnee. 2008. Righting Feminism: Conservative Women and American Politics. New York: Oxford University Press.

Staggenborg, Suzanne. 1991. The Pro-Choice Movement. New York: Oxford University Press.

Stimson, James. 1991. Public Opinion in America: Moods, Cycles, and Swings. Boulder, CO: Westview Press.

Thompson, Becky. 2002. “Multiracial Feminism: Recasting the Chronology of Second Wave Feminism.” Feminist Studies: 8(2):336-360.

Van Dyke, Nella; Sarah A. Soule, and Verta A. Taylor. 2004. “The Targets of Social Movements: Beyond a Focus on the State.” Research in Social Movements, Conflicts, and Change 25: 27-51.

Walker, Edward T., Andrew W. Martin, and John D. McCarthy. 2008. “Confronting the State, the Corporation, and the Academy: The Influence of Institutional Targets on Social Movement Repertoires.” American Journal of Sociology 114(1): 35:76

Werum, Regina and Bill Winders. 2001. "Who's 'In' and Who's 'Out': State Fragmentation and the Struggle Over Gay Rights, 1974-1999." Social Problems 48:3:386-410 24

Zald, Mayer and Roberta Ash. 1966. “Social Movement Organizations: Growth, Decay and Change”, Social Forces 44(March): 327-40.

Zald, Mayer und Bert Useem. 1987. “Movement and Countermovement Interaction: Mobilization, Tactics, and State Involvement.” In Mayer N. Zald and John D. McCarthy, editors. Social Movements in an Organizational Society. New Brunswick: Transaction Books. Pp. 247-272.

Table 1. Analysis of All Feminist Events

Variables Base Model With Feminist Legislation Coef. s.e. Sig. Coef. s.e. Sig.

Movement eventst-1 0.14 0.08 0.083 0.20 0.10 0.041

Movement eventst-2 0.29 0.08 0.000 0.32 0.09 0.001

Movement eventst-3 0.35 0.08 0.000 0.29 0.09 0.002

Countermovement eventst 0.13 0.06 0.033 0.04 0.10 0.678

Religious Rightt -0.06 0.06 0.275 -0.09 0.09 0.315

Δ Public Opiniont-3 0.49 0.28 0.086 0.45 0.33 0.175

Δ Public Opiniont-4 0.77 0.29 0.008 0.75 0.33 0.027

Δ Dem. Presidentt -0.48 0.35 0.174 -0.28 0.42 0.496

Δ # of dem. statest 0.16 0.14 0.246 0.17 0.15 0.260

Δ women's workforce participationt-2 36.23 20.07 0.073 22.02 24.00 0.361

Δ # of women in Congresst-1 0.02 0.03 0.497 0.04 0.07 0.527

Δ # of feminist bills passedt ------0.07 0.07 0.305

Constant -0.09 0.07 0.220 -0.08 0.09 0.396

N= 139 111 pseudo R-squared= 0.54 0.61 Alternative Durbin's statistic= 1.549 5.776 * (* significant at p=.05)

28 Table 2. Analysis of Pro-ERA events

Variables Base Model With Feminist Legislation Coef. s.e. Sig. Coef. s.e. Sig.

Movement eventst-1 0.50 0.10 0.000 0.47 0.13 0.001

Movement eventst-2 -0.12 0.11 0.283 -0.18 0.14 0.202

Movement eventst-3 0.17 0.10 0.098 0.20 0.13 0.132

Countermovement eventst-1 0.44 0.12 0.000 0.51 0.16 0.003

Religious Rightt -0.02 0.12 0.885 -0.39 0.33 0.246

Δ Public Opiniont 1.27 0.53 0.019 1.14 0.67 0.097

Δ Public Opiniont-3 0.78 0.49 0.119 0.99 0.62 0.118

Δ Dem. Presidentt -0.44 0.64 0.492 -0.74 1.08 0.498

Δ # of dem. statest 0.09 0.08 0.260 0.15 0.10 0.163

Δ women's workforce participationt-1 42.64 40.06 0.291 138.81 69.65 0.053

Δ # of women in Congresst-1 0.04 0.05 0.448 0.08 0.20 0.68

Δ # of feminist bills passedt ------0.23 0.14 0.112

Constant -0.08 0.16 0.625 -0.23 0.27 0.399

N= 79 51 pseudo R-squared= 0.45 0.38 Alternative Durbin's statistic= 0.74 0.59 (* significant at p=.05)

29 Table 3. Analysis of Pro-Choice Events

Variables Base Model With Feminist Legislation Coef. s.e. Sig. Coef. s.e. Sig.

Movement eventst-1 0.10 0.08 0.207 0.15 0.10 0.130

Movement eventst-2 0.11 0.07 0.139 0.19 0.10 0.062

Movement eventst-3 0.14 0.07 0.071 0.07 0.10 0.498

Countermovement eventst 0.50 0.07 0.000 0.27 0.09 0.003

Religious Rightt -0.11 0.07 0.106 -0.07 0.08 0.366

Δ Public Opiniont 0.03 0.03 0.185 0.01 0.02 0.807

Δ Public Opiniont-4 0.17 0.08 0.043 0.12 0.08 0.119

Δ Dem. Presidentt -1.03 0.39 0.009 -0.62 0.33 0.067

Δ # of dem. statest 0.00 0.03 0.919 -0.01 0.03 0.806

Δ women's workforce participationt -34.33 23.69 0.150 6.28 21.73 0.773

Δ women's workforce participationt-1 52.32 23.80 0.030 74.25 21.53 0.001

Δ # of women in Congresst-1 0.07 0.03 0.026 0.04 0.07 0.566

Δ # of women in Congresst-2 -0.08 0.03 0.016 0.06 0.07 0.343

Δ # of feminist bills passedt ------0.10 0.05 0.085

Constant -1.96 1.44 0.176 -0.58 1.20 0.628

N= 129 101 pseudo R-squared= 0.49 0.28 Alternative Durbin's statistic= 0.861 0.000 (* significant at p=.05)

30 Table 4. Analysis of All Anti-Feminist Events

Variables All Anti-Feminist Events Base Model With Women's Rep. Coef. s.e. Sig. Coef. s.e. Sig.

Countermovement eventst-1 0.18 0.08 0.022 0.41 0.09 0.000

Countermovement eventst-2 0.13 0.08 0.136 0.05 0.15 0.728

Countermovement eventst-3 0.26 0.09 0.003 0.24 0.13 0.074

Movement eventst 0.17 0.08 0.028 0.10 0.07 0.136

Religious Rightt-1 0.06 0.07 0.389 0.08 0.08 0.331

Δ Public Opiniont 0.48 0.35 0.183 0.30 0.27 0.280

Δ Dem. Presidentt 0.44 0.40 0.271 0.17 0.30 0.585

Δ # of dem. statest -0.29 0.17 0.095 -0.24 0.13 0.057

Δ women's workforce participationt-1 -29.67 24.58 0.230 -58.03 20.82 0.006

Δ # of women in Congresst 0.06 0.04 0.109 0.01 0.06 0.904

Δ # of feminist bills passedt ------0.18 0.05 0.001

Constant 0.05 0.09 0.602 0.11 0.09 0.206

N= 141 113 pseudo R-squared= 0.35 0.34 Alternative Durbin's statistic= 0.00 1.068 (* significant at p=.05)

31 Table 5. Analysis of Anti-Equal Rights Amendment Events

Variables Base Model With Women's Rep. Coef. s.e. Sig. Coef. s.e. Sig.

Countermovement eventst-1 0.05 0.12 0.674 0.06 0.12 0.625

Movement eventst-1 0.17 0.09 0.065 0.10 0.09 0.293

Religious Rightt-1 0.24 0.12 0.053 1.10 0.25 0.000

Δ Public Opiniont-2 0.88 0.49 0.078 0.71 0.47 0.134

Δ Public Opiniont-3 1.16 0.50 0.024 0.60 0.49 0.227

Δ Dem. Presidentt -0.15 0.64 0.821 -0.51 0.82 0.536

Δ # of dem. statest 0.13 0.08 0.087 0.08 0.08 0.322

Δ women's workforce participationt-3 67.10 41.93 0.114 174.90 57.29 0.004

Δ # of women in Congresst 0.02 0.05 0.672 -0.14 0.16 0.376

Δ # of feminist bills passedt ------0.31 0.10 0.005

Constant -0.30 0.16 0.061 -0.60 0.22 0.010

N= 79 51 pseudo R-squared= 0.15 0.47 Alternative Durbin's statistic= 0.00 0.001 (* significant at p=.05)

32 Table 6. Analysis of Anti-Abortion Events

Variables Base Model With Women's Rep. Coef. s.e. Sig. Coef. s.e. Sig.

Countermovement eventst-1 0.13 0.09 0.165 0.30 0.10 0.002

Countermovement eventst-2 0.19 0.09 0.034 0.17 0.16 0.308

Countermovement eventst-3 0.27 0.09 0.003 0.20 0.17 0.239

Movement eventst-1 0.07 0.10 0.475 0.02 0.12 0.844

Religious Rightt-4 -0.10 0.07 0.167 -0.13 0.08 0.130

Δ Public Opiniont-1 0.09 0.10 0.333 -0.04 0.08 0.629

Δ Dem. Presidentt 0.46 0.46 0.323 0.35 0.37 0.339

Δ # of dem. statest -0.05 0.04 0.228 -0.04 0.03 0.148

Δ women's workforce participationt-1 -34.45 27.90 0.219 -66.11 23.41 0.006

Δ # of women in Congresst 0.07 0.04 0.106 0.06 0.06 0.290

Δ # of feminist bills passedt ------0.18 0.06 0.003

Constant 0.08 0.10 0.419 0.13 0.10 0.206

N= 132 104 pseudo R-squared= 0.31 0.24 Alternative Durbin's statistic= 0.535 0.131 (* significant at p=.05)