Trends in amendment lobbying

Most amendments of the national constitution were considered and approved in state legislatures, where lobbyists regularly work. Among the 27 existing amendments, Congress sent all but one to state legislators for approval.Footnote ² Scholars have sought to identify interests that affected the decisions of state legislators regarding proposed amendments. For example, Baack and Ray (Reference Baack and Ray1985) and Barney and Flesher (Reference Barney and Flesher2008) found that states with more veterans and agricultural interests approved of the proposed income-tax amendment sooner than other states. Urban and Catholic interests may have helped encourage the repeal of the prohibition amendment (Dinan and Heckelman Reference Dinan and Heckelman2014), and organized Catholic and farming interests may have undermined the proposed Child Labor Amendment (Greene Reference Greene1988). None of these studies directly observed lobbying related to the proposed amendments. Rather, the studies employed proxy variables for mobilization.

Research produced by political scientists and others suggests several trends that might be observed among lobby efforts related to the ERA or even other proposed amendments. Perhaps most obviously, organized interests are said to have targeted states in which their efforts could have presumably had some effect. Since legislators could vote on the ERA multiple times prior to approving but not retract their decisions to ratify, organized interests would have directed their efforts toward states that had not yet approved of the ERA, but with some distinctions. Brasher, Lowery, and Gray (Reference Brasher, Lowery and Gray1999) present evidence that more interests lobby state legislators when there is some chance of policy change. In other words, there is little reason to lobby when the stakes are low. Moreover, according to Denzau and Munger (Reference Denzau and Munger1986), there is little reason for a group to lobby if legislators are likely to support the group’s cause in any case, including due to favorable public opinion. This suggests that, where public support for the ERA was strong, there was little reason to lobby. Indeed, 30 states approved of it within one year of its formal proposal by Congress. There likely was little mobilization regarding the ERA in these states. Over time, however, public approval of the ERA gradually declined (Soule and Olzak Reference Soule and Olzak2004). This suggests that advocates and detractors may have focused their efforts on remaining states, particularly those where legislative outcomes were unclear.

Historical accounts suggest that legislators in some states, indeed, were lobbied more often than legislators elsewhere. By the spring of 1977, or after Indiana’s legislators had approved of the ERA, the national coalition ERAmerica identified several states as being most likely to ratify: Florida, Illinois, Missouri, Nevada, North Carolina, and Oklahoma (Dinges Reference Dinges1977). Illinois proved to be a particular hotbed for ERA activism (see Mansbridge Reference Mansbridge1986, 165–77), and legislators in Florida, Georgia, Kentucky, North Carolina, and Virginia all experienced intensive lobby efforts (Baker Reference Baker2015; Boles Reference Boles1979, 72–82; Carver Reference Carver1981; Matthews and De Hart Reference Matthews and De Hart1990; White Reference White1989; Young Reference Young2007). Not all of the states that failed to approve of the ERA were targeted equally. No or few accounts suggest that legislators in Alabama, Arizona, Arkansas, Louisiana, Mississippi, South Carolina, or Utah were likely to approve, and no accounts accordingly suggest that legislators in those states experienced large numbers of lobby efforts.

There is historical evidence that different organizations engaged in lobbying to different degrees, but the accounts of such lobby efforts do not imply hypotheses that may be tested easily. Lobbying was not the only tactic that organizations relied upon for advocating for or against the ERA, but it was a prominent tactic. After Congress proposed the amendment in 1972, an “ERA Action Committee” composed of delegates from NOW, the National Woman’s Party, Common Cause, League of Women Voters (LWV), and National Federation of Business and Professional Women’s Clubs (BPW) formed to coordinate state advocacy efforts. By June 1973, the committee had assigned roles to the member organizations: NOW would use outside pressure tactics, the Party would seek to elect pro-ERA candidates to state legislatures, Common Cause would distribute issue papers, and the League would train lobbyists. Although “not much coordination occurred” (see Berry Reference Berry1986, 66–67), these roles reflected the historical strengths of the organizations and may have shaped subsequent mobilization efforts. NOW and the National Women’s Political Caucus (NWPC) were new and aimed to recruit members. These new organizations spurred members to take direct action on the ERA (Conover and Gray Reference Conover and Gray1983; Mansbridge Reference Mansbridge1986), which brought greater attention to the new organizations. Long-standing women’s groups (e.g., LWV and American Association of University Women [AAUW]) were more comfortable with traditional lobby tactics, given the size of their existing membership bases, and also endorsed the ERA more slowly (Boles Reference Boles1979, 48–50). Organizations associated with Schlafly and the New Right did not rely exclusively on lobbying but also organized protests and sit-ins (Conover and Gray Reference Conover and Gray1983, 90). Groups of the New Right pioneered political action committees and direct-mail fundraising (91). Although these accounts provide rich detail into the efforts of the various organizations, there are unclear implications regarding which organizations relied on lobbying more than others.

To determine just how much lobbying pertained to the ERA and if some states saw more lobby efforts than others, lobbyist registration records were collected from archives in all states for 374 legislative sessions (i.e., sessions with existing records) that occurred between 1969 and 1982. The records identify the organizations that lobbied legislators and the number of lobbyists they employed during each session. Based on other studies that identify pro- and anti-amendment groups (e.g., Boles Reference Boles1979; Conover and Gray Reference Conover and Gray1983), numbers of lobbyists that registered to advocate for or against the ERA explicitly during each session were calculated, as well as numbers of lobbyists who represented women’s organizations that took stances for the ERA but who may not have lobbied legislators specifically regarding the amendment. Forty-five states required lobbyists to register with state authorities by the time Congress proposed the ERA in 1972. The five states that did not require lobbyists to register by that year were Idaho, Hawaii, Nevada, Utah, and West Virginia; however, all these states adopted registration statutes by 1975. The registration records produced by lobbyists provide insight into the scale and locations of lobbying on the ERA. Lists of registered lobbyists and clients, or registration records, were collected from archives and libraries in all states where records were available or from newspaper articles that published lists of registered lobbyists. Some lists from 1973 and 1975 are from Reitman and Bettelheim (Reference Reitman and Bettelheim1973) or Marquis Academic Media (1975), respectively. In total, the lists span 374 legislative sessions that occurred between 1969 and 1982 in all states, but with unequal coverage across states. Whereas lists were found for every session that occurred during the period for 12 states, no lists were found for Hawaii.Footnote ³

Lists of registered lobbyists are useful for two reasons. First, the lists indicate which individuals were authorized to represent which client organizations during each legislative session. Second, the lists provide a glimpse into the individuals and interests active during sessions on an unprecedented scale. As noted previously, one other study examines registered lobby efforts related to the ERA, but the study examines lobbyists active during 1975 (see Conover and Gray Reference Conover and Gray1983, 176). There are some drawbacks to using the lists. The lists include only individuals who personally solicited legislators. Therefore, the lists do not include individuals who organized protests or testified before committees but who did not meet individually with legislators, such as Phyllis Schlafly herself. Registration criteria also differed across states somewhat, with a small number of states requiring only paid agents to register. The lists continue to provide useful information, however, since lobby laws changed rarely in the states and the effects of such criteria may be parsed statistically.

To demonstrate the scale of ERA lobbying, Table 1 reports the total number of lobby contracts or efforts that appeared in registration records for each year between 1969 and 1982. Contracts are combinations of individual lobbyists and individual clients. As shown in Table 1, numerous advocates registered to lobby. Since individuals could register as representing causes instead of organizations (and indeed numerous individuals registered regarding the ERA but did not provide organization names), contracts capture the total scale of lobbying in a capitol regarding an issue (see Strickland Reference Strickland2019). As shown in Table 1, for guidance, lobbyist lists from 25 sessions were located for 1969. In that year, a total of 7,488 lobby contracts were registered. Among those, zero were related explicitly to the ERA. This should be expected, given that Congress did not propose the ERA until 1972. In 1973, lobbyists active on the amendment began to appear. In that year, there were nine explicitly pro-ERA lobbyists and 24 anti-ERA lobbyists in the states. Such contracts consist of those that mention the ERA as a cause, include the ERA within the organization’s name, or are related to an organization formed exclusively for the purpose of ratifying or defeating the ERA, as listed by Boles (Reference Boles1979, 200–2) or Conover and Gray (Reference Conover and Gray1983). Examples of organizations that focused primarily on the amendment include ERAmerica, Happiness of Womanhood, STOP ERA (later Eagle Forum), and Women Who Want to be Women, among others. The lobbyists for these organizations were active in addition to the 316 representatives of various women’s organizations who may have lobbied legislators regarding the ERA in 1973: the AAUW, BPW, LWV, NOW, and NWPC, among others. During the 374 sessions for which lobbyist lists could be located, a total of 928 explicitly ERA-related lobby efforts were registered, or roughly 0.46% of all lobby efforts registered.

Table 1. Lobby contracts by year

To determine if ERA lobbyists focused on particular states, Figures 1 and 2 present scatterplots of total ERA lobby efforts (i.e., pro- and anti-amendment) per session across time. Note that the vertical axes are not scaled linearly. Figure 1 presents lobby efforts by year in states that approved of the amendment at some point before 1983. Each dot represents a single observed legislative session (although dots from different states overlap), and the dashed line provides the average number of lobby efforts for each year among the sessions observed. Sessions are ordered according to the number of sessions before or after the session in which legislators approved of the ERA. As shown in Figure 1, there were few lobby efforts in states that approved of the amendment, with the exceptions of Indiana and Texas. Indiana was the last state to approve of the ERA before 1982, doing so in 1977 after a protracted lobby battle lasting five years. After that, practically no ERA lobbyists appeared in the state. In Texas, numerous members of Women Who Want to Be Women, an anti-ERA group, registered as lobbyists in 1977. The legislature had considered a rescission resolution two years earlier. The other states in which ERA lobbyists appeared after ratification include Kansas, Minnesota, Montana, and West Virginia, but lobbyist numbers never exceeded four. In contrast, from the second figure, it is evident that non-ratifying states saw the bulk of ERA lobby efforts. In accordance with Mansbridge (Reference Mansbridge, Snow, della Porta, Klandermans and McAdam2013, 421) and others, lobbying on the ERA did not begin in earnest until 1973. Some states are outliers. During two sessions in Florida, more than 200 individuals registered to lobby for or against the amendment. Lobby efforts spiked in Arkansas in 1975, Nevada in 1977, and Georgia in 1982. Despite Illinois presumably being a battleground, fewer than 10 individuals registered in that state each year.

Figure 1. Lobby efforts across ratifying states.

Figure 2. Lobby efforts across non-ratifying states.

Expectations for influence

The previous section demonstrated that hundreds of individuals registered to lobby state legislators regarding the ERA and that their numbers were concentrated in particular states and sessions, but it remains unknown if the various lobbyists’ efforts had any effects on state legislators’ votes. In examining voting trends on the amendment, I propose that the effects of various factors, including lobby efforts, were contingent on the partisanship of legislators. Partisanship itself may have had direct effects on ERA votes by state legislators due to the stances of the parties. Between its initial introduction in Congress in 1923 and its failure in 1983, the parties and organized interests that supported or opposed the amendment shifted. The Republican Party platforms endorsed the ERA between 1940 and 1980, while the Democratic Party platforms endorsed the ERA in 1944 and after the early 1970s (Freeman Reference Freeman2000). Support for the amendment was largely bipartisan when it was formally proposed by Congress in 1972, but the Democratic Party warmed to the amendment further as support among Republicans faded. Importantly, partisanship may have interacted with other factors, particularly lobbying, to determine legislator votes. Soule and Olzak (Reference Soule and Olzak2004) argue that “…policy outcomes are likely to be the result of a number of contingent and interactive forces over and above the main effects of [social] movements” and other factors. They attribute the success of movements regarding the ERA to the political-opportunity structure in each state, and, in particular, the presence of co-partisan allies in legislatures. Such success may be due to movement activists providing information to allies (Burstein and Linton Reference Burstein and Linton2002). Indeed, Soule and Olzak (Reference Soule and Olzak2004) find evidence that the efforts of the AAUW and anti-ERA groups were more effective in legislatures with more Democrats or Republicans, respectively.Footnote ⁴

Importantly, although hundreds of individuals registered to lobby for or against the ERA explicitly, insurance interests are said to have opposed the amendment on the grounds that it would prevent insurance companies from charging men and women different premiums (Critchlow and Stachecki Reference Critchlow and Stachecki2008). Smeal and Steinem (Reference Smeal and Steinem2020) argue explicitly that insurance lobbyists successfully undermined the ERA in state legislatures: “[h]ealth insurance interests alone stood to lose billions if the ERA forced it [sic] to stop charging women more for less coverage, and since that industry was largely state-regulated, it had lobbyists in every state capital.” Further, “…Schlafly was only window dressing. The anti-ERA fix was already in…. Phyllis Schlafly and her antifeminist homemakers had been brought in to cover for legislators who were voting against the ERA anyway.” Smeal and Steinem also argue that corporate interests opposed the ERA in general since it would have required them to pay men and women equally for the same work. Hence, in addition to the efforts of pro- and anti-amendment activists, the lobby efforts sponsored by the insurance industry might have had some effect on legislators’ votes.

The amendment may have created a context ripe for counteractive lobbying, and that such lobbying, if it occurred, has the potential to obscure correlations between lobby efforts and votes cast by ostensible legislator allies. Ainsworth (Reference Ainsworth1997) suggests that counteractive lobbying, or when competing advocates seek to influence the votes of the same legislators, may occur when issues are “contentious and openly debated, [thereby] reducing the opportunities for legislative bargaining and log-rolling,” when outcomes are binary or involve stark legislative choices, and when issues do not recur (526). As an example, Ainsworth (Reference Ainsworth1997) and Austen-Smith and Wright (Reference Austen-Smith and Wright1994) claim that such lobbying occurred during the confirmation battle over Robert Bork. Decisions regarding the ERA did not enable legislators to bargain with one another easily (as with tax policy) or spur lobbyists to focus only on their ostensible allies, but the amendment was a recurring issue in some states. If advocates and detractors of the ERA solicited the same legislators frequently, then correlations between their lobby efforts and co-partisan legislators’ votes should be less discernible statistically.

The personal traits of individual legislators might have mattered for their support for the ERA. Hill (Reference Hill1983) theorized that partisan affiliations may have had weaker predictive power over the votes of women legislators on the ERA than over the votes of men legislators. Hill suspected that women legislators demonstrated weaker attachment to party cues on issues, such that votes cast by women legislators on the ERA would reflect partisan cues less strongly than votes cast by men. Moreover, legislators’ racial or ethnic backgrounds might have determined votes. Legislators from populations traditionally excluded from politics might have been more inclined to support the ERA due to generalized support for civil rights, but such a difference might also have been due to constituency effects (see Wohlenberg Reference Wohlenberg1980). Surveys reveal that African-American voters were more supportive of the ERA in the 1970s than others (see Burris Reference Burris1983; Huber, Rexroat, and Spitze Reference Huber, Rexroat and Spitze1978), so legislators representing those voters might have been more supportive of the ERA as well. Moreover, the link between party and ERA votes may have been weakened somewhat by race or ethnicity, given that non-white Republican legislators supported the ERA due to generalized support for civil liberties or constituency effects. No existing studies determine if partisanship and race have interactive effects on ERA votes.

Finally, levels of public support may have mattered for votes on the ERA. If state legislators seek to be reelected, and if votes on the ERA were salient or highly publicized, then such votes should have presumably been correlated with public opinion on the ERA in legislators’ districts. There is significant evidence that legislators seek reelection (Erikson, Wright, and McIver Reference Erikson, Wright and McIver1993; Mayhew Reference Mayhew1974). With regard to salience, roll call votes on the ERA were often published in newspapers, so legislators had some incentives to vote according to their constituents’ preferences. Public support for the amendment differed over time, particularly across parties; therefore, the effect of public opinion itself may have varied across parties. Recall that support for the ERA was largely bipartisan when Congress formally proposed it in 1972. Over time, support declined (Soule and Olzak Reference Soule and Olzak2004), and Critchlow and Stachecki (Reference Critchlow and Stachecki2008) attribute this decline, in particular, to the publicity efforts of Phyllis Schlafly. Schlafly was adept at sowing doubt regarding the effects of the ERA. During televised interviews, she argued that the amendment’s ratification would force Congress to require women to sign up for the draft and lead to universal unisex bathrooms. By the late 1970s, Democrats and Republicans had very different views of the ERA: 62% of Democrats and 42% of Republicans favored the ERA (Mansbridge Reference Mansbridge1986, 215). Unfortunately, district-level statistics regarding ERA support are unavailable, but if legislators reflected the majority parties of their districts, then Democratic legislators might have continued to support the amendment despite broader declines in public support as Republican legislators came to vote increasingly in line with public support in general (which was brought down over time by declines in support among Republicans specifically).

Hence, I propose that partisanship did not temper the effectiveness of lobby efforts exclusively but also may have affected the links between ERA votes and legislators’ personal traits and public support for the ERA. Legislators in all states were members of the national Democratic and Republican Parties, and support for the ERA shifted throughout the 1970s within the traditional bases of support for the parties. The platforms of the parties came to reflect the preferences of their supporters regarding the amendment. As a result, the effects of lobbyists, traits, and public support for the ERA were all contingent on the partisanship of legislators.

Data

A data set consisting of roll call votes by state legislators on the ERA that occurred between 1973 and 1982 was constructed. The data set was restricted to those years since the ERA was immensely popular when it was proposed by Congress, and, from Table 1, lobbyists did not mobilize in earnest until a year later. Only after Congress proposed the ERA did Phyllis Schlafly organize STOP ERA and the amendment began to attract the attention of various lobbyists. It was not until July of 1972 that the “national campaign against ERA” began “when Schlafly called… [her] Illinois supporters to a one-day meeting at the O’Hare Airport Inn” (Critchlow Reference Critchlow2005, 219). Schlafly founded STOP ERA in September. The organization was renamed the Eagle Forum in 1975, and “[b]y the late 1970s, Schlafly was leading all the anti-ERA campaigns in the 15 unratified states” (Conover and Gray Reference Conover and Gray1983, 74). Moreover, the sample of votes is limited since roll call votes that occurred in 1972 were largely supportive of the amendment, so there is also limited variation in votes from that year. A few legislative chambers (e.g., the senates in Massachusetts, Michigan, and Oklahoma) even employed unrecorded voice votes to register their approval.

Table 2 lists the roll call votes in the data set. The roll calls were found in an issue brief by Gladstone (Reference Gladstone1980) and various other sources, including newspaper articles. There were a total of 87 roll calls comprised of 6,952 individual votes, including pairs in which legislators were absent but nonetheless indicated how they would have voted on the ERA. Legislators’ votes themselves were compiled from legislative journals and newspaper clippings. Only 25 states are listed since the others approved of the ERA prior to 1973. In the table, votes in lower chambers are separated from those in upper chambers. For example, representatives in Alabama never voted on the ERA, whereas senators voted on and rejected the amendment twice: in 1973 and 1978. Only in Illinois did two votes ever occur during a single session. The table does not include roll calls to rescind prior ratification decisions, present the ERA to voters, or on modified versions of the ERA.Footnote ⁵ The roll calls listed in Table 2 also do not include committee actions.

Table 2. ERA roll call votes, 1973–1982

Figure 3 presents histograms showing when ERA roll call votes occurred and the numbers of individual votes cast by legislators during each year. As shown in Figure 3, the bulk of votes were cast in 1973 or 1975. The bumps for odd-numbered years are due to state legislatures not convening during even-numbered years. No votes were cast in 1981. Although roll call votes decreased in frequency over time, there was a final burst of activity in 1982 when seven legislative chambers considered the ERA one last time, including the two in Illinois. The data set is the largest such set of roll call votes on the ERA compiled to date. In comparison, Hill (Reference Hill1983) examined only 37 roll call votes in which majorities in the parties disagreed.

Figure 3. ERA votes over time, 1973–1982.

Lobby statistics that are aggregated at the level of states were appended to the data set. Lobbyist lists were located for 80 of the 87 sessions in which ERA votes occurred, which include 6,682 or 96.1% of the 6,952 votes that were cast.Footnote ⁶ The first statistic consists of totals of pro-ERA lobbyists or only those who advocated for the amendment explicitly, including lobbyists for ERAmerica and various state-based coalitions, among others. The second statistic consists of anti-ERA lobbyists or stoppers, including those associated with Happiness of Womanhood, STOP ERA (later Eagle Forum), and Women Who Want to Be Women, among other organizations. To test the hypothesis of Smeal and Steinem (Reference Smeal and Steinem2020), the data set includes totals of insurance lobbyists, including all lobbyists for health, life, and property insurance companies, including the Farm Bureau. I control for the total number of lobby efforts during every session.

The partisanship and personal traits of legislators were identified and included in the data set. Partisan affiliations were collected from Klarner et al. (Reference Klarner, Berry, Carsey, Jewell, Niemi, Powell and Snyder2013), who compiled a data set of historical election returns in the states. In Minnesota, where legislators did not campaign as Democrats or Republicans prior to 1973, legislators were classified based on their caucuses (see Masket Reference Masket2016). Women legislators were identified using a comprehensive directory of women state legislators (i.e., Cox Reference Cox1995). The ethnic or racial backgrounds of legislators in the data set were collected from Klarner (Reference Klarner2021).

To detect the effects of public opinion, public approval ratings of the ERA for the states were collected from Soule and Olzak (Reference Soule and Olzak2004). I thank Sarah Soule for providing the public opinion data. In the words of Soule and King (Reference Soule and King2006, 1885), the “data come from Gallup polls conducted in 1974, 1975, 1976, 1978, 1980, 1981, and 1982… which were obtained from the Roper Center for Public Opinion Research.” Such data are available only for years prior to a state’s decision to ratify, so the closest available measures in states that ratified were used where needed.

Prior to conducting any statistical tests for influence, various descriptive statistics regarding the propensity to vote for the ERA among various groups of legislators based on party, gender, race, and public opinion are presented. Table 3 presents the percentages of legislators in each group who voted in favor of the ERA (as opposed to voting against it, abstaining, or being absent). The columns parse legislators by party, whereas the rows parse legislators by combinations of gender and race. Each cell indicates the percentage voting in favor of the ERA and the number of observations within each category. As shown in Table 3, for example, roughly 37% of the votes cast by Republicans who were white men were cast in favor of the ERA. There were 2,430 votes cast by such legislators in the data set. This percentage contrasts with that for Democratic women who were non-white. Nearly 93% of the votes cast by these legislators were cast in favor of the amendment. The data set includes no Republican women who were non-white. Similarly, Table 4 reports the percentages of votes cast in favor of the ERA by party and different quartiles of public support for the amendment, with the first quartile, including the observations with the lowest levels of support. Support among Democrats was high, at least compared to support among Republicans, for all four quartiles of public opinion. Even in the lowest quartile, 60% of Democrats voted for the ERA. This amount shifted slightly to 64.6% in the fourth quartile. Republican support increased from 33.1% to 53.2% across the quartiles.Footnote ⁷

Table 3. Percent votes for ERA by party and traits

Table 4. Percent votes for ERA by party and approval

The numbers presented in Tables 3 and 4 consist of votes and not unique legislators. Numerous legislators cast multiple votes since they served over multiple sessions. Among the 6,952 votes cast, there were 4,392 unique legislators. A majority of legislators (2,735 or 62.3%) saw only one roll call while in office. A small number (143 or 3.3%) of legislators served so long in the Illinois House that they had five, six, or seven opportunities to vote for the ERA. Among legislators who had two or more opportunities to vote for the amendment, 256 or 15.4% changed their stance at some point. Of those, 150 legislators voted for the ERA before voting against it or abstaining, and 106 shifted the other way.

Tests for influence

To determine if lobbyists and other factors were correlated with votes, a series of logistic regression models were estimated in which the dependent variable for each model is a binary indicator for whether a legislator voted in favor of the ERA. Given that fewer than 3% of legislators abstained or did not cast ERA votes at all, and for the sake of simplicity, a binary outcome variable (i.e., “yes” vote or otherwise) and logistic regressions were used rather than multinomial logistic regressions. Given the central role of the party in my expectations, three models are estimated: a model for Democrats, a model for Republicans, and a combined model. In the combined model, various interactive terms that test for whether the differences in coefficient values between Democrats and Republicans are statistically different from zero are included. Each model includes state and year effects. The state effects capture the effects of time-invariant factors unique to each state that are not included explicitly in the models, and the year effects capture national averages in legislators’ overall propensity to vote for the ERA for each year. The standard errors are clustered by session and chamber since some legislators may have changed their votes in response to the votes of other legislators. As an example, factional conflicts among Democratic lawmakers in Illinois led to the demise of the ERA in that state (see Critchlow and Stachecki Reference Critchlow and Stachecki2008). For interpreting the results presented in Table 5, note that legislators change from year to year. The coefficients for legislator-level variables (e.g., party, gender, and race) reflect overall differences across groups of legislators in propensities to vote for the ERA. The coefficients are not based on changes in votes within individual legislators or even among groups of legislators over time.

Table 5. Predicting legislators’ votes for ERA, 1973–1982

Note. State and year effects were included in all models but not reported. Clustered errors are reported in parentheses.

* p < 0.1.

^** p < 0.05.

^*** p < 0.01 on two-tailed tests.

As shown in the models, different factors were correlated with votes from Democrats and Republicans. Among Democratic legislators, no lobby efforts were correlated with votes. Rather, women and non-white legislators were more likely to vote than men and white legislators, respectively. Democratic legislators voted more often for the ERA as state-level public support was lower. For Republicans, lobby efforts mattered. Votes by Republicans were correlated positively with pro-ERA efforts and negatively with stopper efforts. As with Democrats, women and non-white legislators were more likely to vote for the amendment than other Republicans. The first model has slightly more explanatory power than the second. The results presented in Model 3 confirm the results from the first two models. The results are largely the same, and the coefficients for various interactive terms indicate that the differences in coefficient values for Democrats and Republicans (for the direct effects) are themselves statistically different from zero. Specifically, members of the parties reacted differently to explicitly pro- and anti-amendment lobby efforts, gender, and levels of public opinion. The difference between white and non-white voting for the ERA was similar across both parties, even if Democrats (Republicans) were generally more (less) likely to vote for the amendment.

Pro- and anti-ERA lobbyists had somewhat countervailing effects on party differences in voting. Whereas pro-amendment lobbyists helped to narrow the partisan gap in ERA voting, stoppers widened the gap. According to Model 3, for every additional ERA lobbyist who appeared, the odds of Republicans voting for the ERA relative to all other legislators increased by 5% on average, ceteris paribus. For each stopper lobbyist, the odds of Republicans voting for the ERA decreased by 1.7%. Note that these are changes in odds ratios and not relative risks. Figures 4 and 5 present the predicted probabilities that white, male legislators in different parties voted for the ERA for different numbers of amendment advocates and stoppers. All other variables are held at their means, except for public support, which is held at 50%. The probabilities are calculated based on the results of Model 3. As shown in Figures 4 and 5, ERA lobbyists narrowed the partisan divide in amendment voting, whereas stoppers widened it, and the effects of both groups of lobbyists were confined to Republicans. There is no evidence presented that insurance lobbyists affected ERA votes.

Figure 4. Predicted probabilities of voting for ERA by party and ERA lobbyists.

Figure 5. Predicted probabilities of voting for ERA by party and stoppers.

Given the binary measures of legislator traits, Table 6 presents the predicted probabilities of legislators in different groups to have voted for the ERA, along with 95% confidence intervals, based on Model 3. The probabilities are generated when all other variables are held at their mean values. As expected, these probabilities mirror the actual averages presented earlier, but the confidence intervals for some estimates are quite large due to small subsamples.

Table 6. Predicted probabilities of voting for ERA by party and traits

Figure 6 presents the predicted probabilities of white, male Democrats and Republicans voting for the ERA under different levels of public support and when all other variables are held at their means. The negative (positive) correlation between Democratic (Republican) votes and public opinion is evident. As mentioned, these correlations may be due to legislators representing majority parties from their districts and shifts in opinions on the ERA among Republicans.

Figure 6. Predicted probabilities voting for ERA by party and opinion.

The results presented in Model 3 cannot be used easily to determine if Schlafly “actually changed any votes,” as Smeal and Steinem (Reference Smeal and Steinem2020) deny, or if “Schlafly… [had] not entered into the fray, the ERA would have been ratified,” as suggested by Critchlow and Stachecki (Reference Critchlow and Stachecki2008, 165), because of statistical imprecision; however, they provide some predictive utility. Note that the logistic regressions estimated in Table 5 all predict at least 66% of ERA votes, or assign probabilities of voting for (against) the ERA greater (less) than 50% to legislators who actually did vote for (against) the amendment. Based on Model 3, in a counterfactual scenario in which no stoppers appeared in any states, the predicted probabilities of legislators voting for the ERA do not change for 78.96% of legislators. This is likely because most legislators were Democrats who were unaffected by any lobbyists and zero or few stoppers appeared in most states. For only 61 votes, all in Florida but during different years, did the predicted probabilities of legislators voting “yes” increase from under 50% to above that threshold. Most of those votes were cast against the ERA, whereas 18 were in favor. This suggests that stoppers might have persuaded or pressured 43 legislators to vote against the amendment. In a counterfactual scenario in which no pro-ERA lobbyists appeared, the predicted probabilities of legislators voting for the ERA do not change for 50.28% of legislators. This may be due to these lobbyists being more effective than stoppers (when one compares coefficients in Model 3) and having appeared in more states than stoppers. Nevertheless, the number of legislators who had a 50% chance or greater of voting for the ERA and whose predicted probabilities decreased below that threshold is only 32. These legislators all served in Georgia, Illinois, or Nevada, and 10 voted against the amendment. Hence, pro-ERA lobbyists may have persuaded or pressured 22 legislators to vote for the amendment.Footnote ⁸ Therefore, lobbyists likely had small effects on votes, if any, and did not change outcomes in the states. Reaching this conclusion requires us to assume that legislators would not have changed their votes based on how other legislators voted, including legislators in other states.

Robustness checks

The 6,952 votes in the sample were cast by 4,392 unique legislators since many legislators voted on the ERA multiple times. Although fewer than 16% of legislators who cast multiple votes changed their stance at some point, models in which only the initial votes of legislators are examined were estimated. The results from these models are each based on single votes from unique legislators such that no individual legislator’s multiple votes skew the estimates. The estimates for these models are reported in Table 7. The trends reported in the table largely mirror those reported earlier in Table 5 but are somewhat less discernible statistically. I continue to find that Democrats were largely unaffected by the efforts of lobbyists, while Republicans responded somewhat to stopper efforts. Differences in slopes are still not equal to zero, as expected. Gender and race mattered for legislators’ initial votes, just as in Table 5, but one of the most noticeable changes across the two tables is the estimated effect size for public opinion on Democrats’ votes. Whereas there is a negative and statistically discernible relationship between opinion and Democrats’ votes in Table 5, the relationship disappears in Table 7. The correlation between opinion and Republican votes also becomes weaker. Since legislators cast subsequent votes during later years in most cases (i.e., only one session ever featured more than one roll call), the findings regarding lobby efforts and public opinion suggest that Democrats continued to vote for the ERA despite declines in public support for the amendment, Republicans began to shift against the amendment as opinion changed, and lobbyists’ influence grew over time as Republican legislators began to reconsider their support for the ERA.

Table 7. Predicting legislators’ first votes for ERA, 1973–1982

Note. State and year effects were included in all models but not reported. Clustered errors are reported in parentheses.

* p < 0.1.

^** p < 0.05.

^*** p < 0.01 on two-tailed tests.

In Table 8, correlations between lobby efforts and votes among legislators separated by region are examined. Differences between non-Southern and Southern states are examined since legislators who served in office in the South, particularly Democrats, are commonly found to differ from their non-Southern counterparts in terms of ideology. In particular, such legislators were found to be more conservative than Democrats not from the South (Shor and McCarty Reference Shor and McCarty2011; Reference Shor and McCarty2022). The first three models reported in the table examine legislators from non-Southern states. As with the models presented earlier, logistic coefficients are estimated, and errors are clustered by session. Regarding lobby efforts, trends among non-Southern Republicans conform best to my expectations: No efforts are correlated with votes among non-Southern Democrats, but lobbyists, including insurance lobby efforts, are correlated with the votes of non-Southern Republicans. As shown in the additional models in the table (which examine votes only in Alabama, Florida, Georgia, Louisiana, North Carolina, South Carolina, and Virginia), such trends are absent or reversed among Southern legislators. Regarding traits, gender mattered for votes among all four groups of legislators, although differences in slopes are not discernible across parties. Also, only non-Southern Republicans voted in accordance with public opinion (see Table 8).Footnote ⁹

Table 8. Predicting legislators’ votes for ERA by region, 1973–1982

Note. State and year effects were included in all models but not reported. Clustered errors are reported in parentheses.

* p < 0.1.

^** p < 0.05.

^*** p < 0.01 on two-tailed tests.