Framing morality policy issues: state legislative debates on abortion restrictions

Abstract

Scholars of “morality policies” have often assumed a signature characteristic of such policies is that advocates will frame them as clashes between fundamental moral and religious principles. Recent studies of issues typically considered under the “morality policy” rubric have found that advocates often frame these issues along multiple dimensions and that they do not necessarily favor frames that emphasize moral principles over other considerations. This paper examines this issue for abortion policy. We analyze verbatim records of debates over 26 recent proposals to restrict abortion rights in the 16 states for which data are available. We found that both sides in debates over abortion restrictions framed the issue along several dimensions with no single frame dominating most of the debates. While there is some empirical support for the morality policy perspective, the frequency that advocates employed morality frames was less than we expected given the disproportionately high levels of evangelical Protestant membership in the states we examined. Rather than simply casting the debate as one over irreconcilable moral principles, the two sides’ strategies often converged by framing the issue in terms of various consequences of abortion and abortion restrictions for women. Advocates propensity to frame the issue in terms of “right to life” versus “woman’s choice” principles rose when one side or the other escalated rhetoric about “life” or “choice” principles (inducing the other to respond in kind). Our data thus conform to the logic of a game of tit-for-tat in which individuals follow a strategy of “retaliation” if their opponents frame issues in highly moralized, judgmental terms, or they “cooperate” by emphasizing how their preferred policy will promote some widely shared value (like women’s welfare or the authoritativeness of medical research). “Morality talk” was also more prevalent when the debates were about bans on abortion rather than other types of restrictions. The broad implication of our findings is that the propensity of advocates to frame issues in terms of fundamental moral principles has less to do with the general subject matter or issue area (e.g., abortion) and more to do with the context of debate and strategic considerations.

Introduction

An essential characteristic of public policy-making is how advocates “frame” issues. We have recognized the importance of framing (or “issue definition”) in public policy for decades (Stone 1988; Riker 1996; Rochefort and Cobb 1984). The literature on framing in policy studies focuses on who frames, what they frame, and why they do so. We have learned that policy entrepreneurs, the media, and other actors frame issues, proposals, and the groups that policies target (Kingdon 1984; Mintrom 2000, 31–33; Pierce and Miller 2004, 38–39; Rochefort and Cobb 1984) and that framing can shape the “scope of conflict,” and ultimately, who wins and loses policy battles (Schattschneider 1962; Bosso 1989; Baumgartner and Jones 1993; Mucciaroni1995).¹

Framing is particularly critical in debates over “morality policies.” Scholars have identified morality policy as a separate type of policy that is marked by a distinct way in which advocates frame issues, along with other shared characteristics.² For Haider-Markel and Meier (1996, 333), “If at least one advocacy coalition involved in the debate defines the issue as threatening one of its core values, its first principles, we have a morality policy.” According to Mooney (2001, 3, 4, 172), “[W]hat ties these issues together is that each involves a controversial question of first principle…. Morality policy involves issues on which there is significant disagreement about first principles.” The abortion issue, in particular, has been frequently understood through the lens of morality policy (Luker 1984; Goggin 1993; Meier 1994; Mooney 1999, 2001; Mooney and Lee 1995, 1999; Tatalovich and Daynes 1981; Strickland 1998; Meier and McFarlane 1993, 205, 255; Studlar 2001; Roh and Berry 2008). Public policies aimed at abortion “seek to regulate social norms” and “evoke[s] strong moral responses…” (Mooney and Lee 1995, 600). Indeed, abortion is regarded as perhaps the quintessential and “most widely discussed” morality issue (Studlar 2001, 44–46).

This article examines whether, in fact, policymakers frame the abortion issue in terms of moral and religious principles, how often they do, and under what conditions they are more likely to do so. It builds upon an emerging literature that challenges the assumption that morality policies are only, or mainly, contests over conflicting and seemingly irreconcilable moral principles. Mucciaroni (2011) found that gay rights opponents typically do not use moral and religious arguments to frame LGBT rights issues even in states with high proportions of Christian evangelicals, where we would expect to find the highest rates of “morality talk.” Instead, they more often frame gay rights in terms of allegedly adverse consequences for groups or institutions in society (e.g., families, the military) or in procedural terms by talking about which level or institution of government should make policy decisions or how they should be decided. Ferraiolo found similar patterns in her studies of gambling (2013) and marijuana decriminalization (2014). State lottery and marijuana legalization opponents largely avoid condemnations of personal behavior, focusing instead on the appropriateness of government sanctioning lotteries or endorsing decriminalization measures as well as negative consequences that would result from policy change. Similarly, in a six-decade-long study of drug use and gambling policy in Germany and the Netherlands, Euchner et al. (2013) find that legislative majorities have shifted away from morality framing and instead emphasize security and public order, health, and fiscal concerns; in turn, this frame shifting generated greater permissiveness in policy outputs.

We conceive of framing in policy-making as a series of strategic decisions that legislators make for building support or opposition to proposals, for explaining their votes to constituents and for other purposes. Whether advocates gravitate toward or away from “morality talk” and deontological reasoning when they frame issues has more to do with their perception of strategic advantages, often shaped by the specific political context, than by the general subject matter (e.g., abortion policy).

Most studies of issue framing focus on framing in the mass media because data on media coverage are easy to obtain and because framing is often studied as part of agenda-setting (Kellstedt 2003; Entman 2004; Reese et al. 2003; Baumgartner and Jones 1993; Weiss 1989). Fluctuations in media and governmental attention to issues and the tone of media coverage signal issues coming on to or off of the agenda (see Baumgartner and Jones 1993). We know less about how policymakers frame concrete proposals within particular institutional contexts when they are up for adoption (Dery 2000). The incentives that legislators face differ from those that influence interest groups and the media. Legislators need to explain their policy decisions to voters and seek to build reputations for being effective and credible with their colleagues, which may shape their framing strategies. Legislators and interest groups also play different roles in the process that may shape how they view legislative compromises, which also might lead to different framing strategies.

While previous studies have not ignored the questions about strategic framing that we examine, few of them have systematically collected and analyzed data on policymakers’ discourse. Our study employs a unique dataset of speeches delivered by state legislators in the USA to examine how policymakers frame the abortion issue. We collected data on 45 debates on three key kinds of abortion restrictions in 16 states between 2011 and 2014. Abortion has been high on the agendas of many state legislatures recently. They have approved a wide array of new restrictions, including three that we examine here: regulations on abortion providers to upgrade their facilities and/or require doctors to have admitting privileges, ultrasound requirements, and bans on abortion after 20 weeks or the detection of a fetal heartbeat. Within the past decade, legislatures have adopted technologies to make verbatim, online records of debates increasingly available to researchers and the public. And because a large number of states have debated abortion restrictions within a few years of one another, we can observe the impacts of variation across states and restriction types, while holding constant the context of abortion policy at the federal level.

In the next section, we review recent efforts to pass abortion restrictions at the state level. Then, we develop some theoretical expectations about how advocates frame the abortion issue under the morality policy perspective as well as under an alternative set of assumptions. After that, we discuss our methods and data. Finally, we present our findings and discuss their implications.

Recent developments in abortion policy

Abortion restrictions have increased sharply in number and stringency as social and religious conservatives have mobilized to test the limits of Roe v. Wade and post-Roe court decisions (Eckholm 2014). Led by Tea Party activists, conservative Republican majorities took control of 25 state legislatures (up from 14) and strengthened their grip on others after the 2010 elections. Over half of the eleven new GOP-controlled legislatures approved restrictions.³ Abortion restrictions began to proliferate in 2011 and the number of new restrictions from 2011 to 2014 exceeded by far the total for the entire previous decade (Boonstra and Nash 2014).

None of the restrictions are new types, but they are more stringent and numerous and have spread to states where they did not exist previously. Notable have been bans on abortions at earlier gestational ages; by mid-2014, nine states had followed Nebraska’s lead in banning abortion at 20 weeks. Abortion foes have also favored “TRAPs”—Targeted Regulation of Abortion Providers—which include mandating hospital admitting privileges for abortion doctors and expensive upgrades of clinics to meet the standards of ambulatory surgical centers. Many of the TRAP laws are on top of other requirements that many states already had in place. Only 11 states had such laws in 2000; today over half do, covering well over half of the American female population (Nash et al. 2014).

The permanent impacts of the restrictions on those who seek and provide abortions remain unclear. They appear to be reducing access to abortion in many states (Tillman and Eckholm 2014). Ultimately, their effects on access will depend upon how well providers and women seeking abortions adjust to them and how courts rule on legal challenges to them.⁴ Some of the restrictions have already reached the Supreme Court and could shape future legislative and judicial actions on the topic. In a 5-3 decision in 2016, the Court invalidated the Texas law that imposed TRAP restrictions in Whole Women’s Health v. Hellerstedt (Liptak 2016). Other TRAP laws may be challenged, and challenges to other kinds of restrictions may reach the Court. It remains to be seen whether Whole Women’s Health will become a durable precedent or whether President Trump’s appointments to the Court will overturn that decision and others that have upheld abortion rights. Legislative actions induce courts to reconsider earlier rulings. Courts may interpret the spread of the restrictions as evidence of a change in society’s attitudes toward the regulation of abortion. Courts often adopt advocates’ arguments for and against abortion regulations that first appear in non-judicial arenas.⁵

Theoretical expectations

Advocates face several strategic decisions in how to frame public policy issues. One decision is whether to pursue a unidimensional strategy (deploying a single dominant frame) or one that is multidimensional (deploying several different frames). The evidence from case studies is mixed, but because many studies rely on media coverage for data, and that coverage has a “tendency to focus attention only on one dimension of a policy” (Baumgartner and Jones 1993, 109), they may overlook additional dimensions or more specific and nuanced frames that the media may filter out. Ascertaining how many frames advocates put forward and their relative importance presents measurement challenges. Few studies measure with any precision the relative importance of each frame that they report, making it impossible to determine whether a frame dominates a debate or is one among several that participants assert frequently.

Policymakers must also decide whether to frame issues so that they converge on broadly similar values and considerations, or, whether to pursue a strategy of divergence, in which each side focuses upon distinct topics and often difficult-to-reconcile values. Since many policy decisions feature broad disagreement over policy, we often assume that policymakers follow a strategy of frame divergence, in which “when one side dominates in the volume of rhetorical appeals on a particular theme, the other side abandons appeals on that theme” (Riker 1996, 6). For example, Bali (2009) found that supporters of a national identification system frame it as a national security and immigration control issue, while opponents frame it as a threat to civil liberties. Frame divergence also appears in other policy areas, such as tobacco (Baumgartner and Jones 1993, 114–117), nuclear power (Weart 1988), pesticides (Bosso 1989), wilderness protection (Ginger 2000), and government paperwork (Weiss 1989), but some evidence for convergence exists (Jerit 2008).⁶

A final consideration for policymakers deciding how to frame is choosing which frame(s), from among several, they think will have greater influence on their audience. To broaden support for their positions, advocates must “align” or “link their interests and interpretive frames” with widely shared values and beliefs that “resonate” with target audiences (Benford and Snow 2000, 624, 622, 629; Snow and Benford 1988; Bosso 1989, 183–187; Chong 2000; Davies 1999; d’Anjou and Van Male 1998; Jasper and Poulsen 1995; Williams and Kubal 1999).⁷ We often conceive of “culture” as a fixed constraint on advocates’ framing strategies, but cultures are “not unified system[s] that push action in a consistent direction”; they include “diverse, often conflicting symbols…and guides to action” instead (Swidler 1986, 277). Policy options are sufficiently malleable so that they can be framed in ways that are consistent with a number of diverse (yet widely shared) values. Policymakers, thus, have some discretion in choosing from a repertoire of culturally congruent frames and in synthesizing new frames that resonate with the culture.

The morality policy perspective, at least as applied to the abortion issue, assumes that advocates will pursue a unidimensional framing strategy in which values diverge; both sides should be expected to frame the issue on a single dimension that reflects their profound moral disagreement over whether public policy should either protect fetal life, or, permit women to make their own choices about whether to terminate a pregnancy. Much of the literature on abortion politics emphasizes these conflicting principled claims (Luker 1984; Tribe 1990) and studies of how advocates framed the issue in the 1970s, 1980s, and 1990s often found evidence of an emphasis on the conflict between protecting “life” and preserving women’s rights, or “choice” and “privacy” values (Condit 1990; Rohlinger 2002; Ferree et al. 2002; Dillon 1993; Williams 2013). And if advocates choose frames that have cultural resonance, one would be hard pressed to find values that fit better than individual autonomy and protecting human life.

However, legislators might develop framing strategies that depart from these assumptions. Advocates on the same side may pursue a multidimensional framing strategy, for example, with some abortion restriction opponents arguing that they violate the principle of choice, while others that they are unconstitutional, and still others that they threaten women’s health. Advocates may disagree about which frame will be the most effective or may simply prefer their own “take” on the issue. Advocates also resort to multidimensional framing because they need to target multiple audiences, such as when they vary the frames that they use in communicating with adherents, potential allies, “bystanders,” or “elite decision-makers” (Klandermans 1992; Evans 1997). Movements “align” their own understandings of an issue with those of target groups in order to build support beyond their most reliable adherents and to neutralize their opponents’ support (Snow et al. 1986). Legislators, similarly, face multiple audiences—core supporters, interest groups, re-election constituents, and legislative colleagues (Fenno 1978). The frames they may find most useful for mobilizing issue activists who agree with them may be different from those for expanding their base of support, explaining their positions to constituents, appealing to moderate swing voters, or building credibility with colleagues. They may not be able to build majority support if they rely primarily upon frames that reflect their core moral beliefs, but may need to deploy frames that resonate with the plurality of Americans who hold ambivalent or nuanced opinions on abortion in order to build such support (Williams 2013).

Legislators who frame abortion restrictions starkly in terms of “life” versus “choice” also risk credibility given the limited reach of most abortion restrictions. State legislatures rarely debate outright bans on abortion because Roe and other Supreme Court decisions have confined abortion opponents’ options to incremental limits on abortion rights. Abortion rights opponents have been unable to overturn the basic holding in Roe for many decades with appeals simply based on the “right to life” principle; meanwhile, supporters have seen state after state reduce access to abortion despite their insistence that “it’s a woman’s choice.” “Life” and “choice” principles are powerful in the abstract, but in the abortion debate that has dragged on for decades, they are closely identified with two extreme, uncompromising camps.

Hence, each side may have a strong incentive to abandon an exclusive reliance on the “life” versus “choice” binary and engage in an alternative frame alignment strategy. Legislators who frame abortion proposals may downplay their core convictions and supplement them with frames reflecting other moral principles or practical consequences. Knutson’s (2011, 32) findings from a study of press releases of religious groups on abortion policy are consistent with this line of thinking. She found that “while religious groups did offer a number of arguments classified with the moral [pro-life] frame…they offered even more arguments not based in morality” and that anti-abortion groups put forward a wider range of arguments in their press releases than appeared in media reporting (which over-represented religious objections to abortion).

An alternative framing dimension that could allow advocates to move beyond the “life” versus “choice” dichotomy and converge on similar themes, are frames rooted in the medical and scientific aspects of abortion. Medicalization was the dominant frame that legal reformers, medical groups, and political elites used in the pre-Roe era of legislative abortion reform. In framing of the issue as a medical procedure, they limited the perceived moral implications of abortion and portrayed it as a non-controversial medical decision, a “limited reform” devoid of “sweeping moral implications” (Burns 2005, 205). No longer seen as a struggle between opposing worldviews, reform was much easier to accomplish. Legislators might also find this strategy attractive because scientists, physicians, and other “non-partisan” experts are accorded widespread respect and can give them “cover” to avoid seeming to favor one moral position over another. One anti-abortion group, the Elliot Institute, began pushing advocates in the 1980s to emphasize “scientific” findings related to abortion’s allegedly negative impacts on women’s health (Rose 2011). Today, abortion restriction supporters point to recent research on the “science of fetal pain” and ostensible “harms to women” from abortion (Belluck 2013).

Another way in which framing strategies may converge on similar dimensions and values is for one side to appropriate a dimension used by the other. In the case of abortion, both sides may converge on the implications of abortion and abortion restrictions on the health and welfare of pregnant women. This strategy broadens each side’s appeal beyond its core “life” and “choice” constituencies and neutralizes their opponents’ arguments. Legislators can show how their policy positions are consistent with the values of their opponents, and at the same time, expose their opponents’ ideological inconsistency. Using such a frame appropriation strategy, we would expect supporters of abortion restrictions to emphasize how they protect women’s interests (values typically associated with liberal politics and abortion rights). Shifting from a fetal-life frame to a gender-based frame reconciles the interests of the fetus with that of the woman. Siegel (2007) found that pro-life advocates of a 2006 abortion ban in South Dakota emphasized the harms to women that come from abortion. Some restriction opponents, similarly, have appropriated the “big government” rhetoric of conservatives to argue against the regulation of abortion as an unwelcome intrusion into “people’s bedrooms” (Saletan 2003). While consistent with a woman’s right to privacy, this frame shifts the focus from women to the government.

Our predictions up to this point have ignored some key contextual variables that might shape legislators’ calculations. First, we expect that each side’s framing strategy will have some effect on the other side’s strategy. One side’s rhetoric may escalate or de-escalate talk about core principles—resembling a “tit-for-tat” game strategy (Axelrod 1984). If one side resorts to rhetoric that accuses the other side of threatening “life” (or “choice”), it should induce the other side to engage in similar moralistic rhetoric.

Another aspect of the context is the distinction between challengers and defenders of the status quo. The literature suggests that the side that advocates policy change should be more inclined to gravitate toward multidimensional framing strategy because they must overcome people’s aversion to the perceived risks of change and expand the scope of conflict in order to alter the balance of power in its favor (see Jerit 2008; Baumgartner et al. 2009). We expect the supporters of abortion restrictions to meet these challenges by putting forward multidimensional frames and frames that converge with those of their opponents more than abortion rights advocates.

Framing strategies may vary, thirdly, with the type of abortion restriction under consideration. Debates over bans on abortion should induce each side to frame the issue as an all-or-nothing struggle in which its core convictions are under threat. We expect less reliance on the life versus choice binary in debates over proposals that raise the barriers to abortion, such as additional tests for patients and licensing requirements for providers, because these restrictions should not be perceived as threats to core values as seriously as bans would be.

Finally, framing strategies should vary across states according to several political characteristics. First, we expect states that have been more hostile to abortion rights historically to feature debates with a greater emphasis on the “life” frame because they are likely to have cultures that stress moral traditionalism and well-organized pro-life movements. Conversely, they are likely to have weaker pro-choice movements, which should induce restriction opponents to emphasize frames that go beyond their commitment to women’s autonomy. Second, while the vast majority of abortion restrictions since 2010 have been legislated in states that have more conservative public opinion, some states are considerably more conservative than others. The size of these states’ memberships in Protestant evangelical denominations also varies, which should be relevant for issues like abortion. A recent Pew study found that 8 of the 10 most anti-abortion religious groups in the USA are evangelical Protestants and they are among the largest denominations in the country. Twice as many evangelicals oppose legal abortion as support it (66–33%). Some evangelical sects had legal abortion opposition rates as high as 77% (Masci 2018). We expect that pro-restriction advocates in states with more conservative public opinion, and higher rates of evangelical Protestantism will frame the issue more unidimensionally by stressing the “life” frame. Our expectation for restriction opponents is less clear, but we expect less unidimensional framing around the “choice” frame in these states. Legislators in states with larger proportions of evangelicals and more conservative publics should find the “choice” frame less appealing because of its identification with liberal secularism and abortion rights.

Data and methods

Legislators engage in framing through verbal and written communication. Public debate is a critical venue for framing within legislatures and the mode of communication that is likely to reach the largest audience of colleagues and constituents. We observed how state legislators frame abortion restrictions by listening to what they said in formal legislative debate. Floor debates are part of a stream of legislative discourse on policy issues and a readily available source of data for indicating how legislators frame issues that come before them. Unlike media reports, official, verbatim records of debate are unabridged and unfiltered by editors and journalists. We chose to examine floor proceedings rather than committee hearings because larger numbers of legislators usually speak during floor debate. Hearings mainly feature how outside witnesses frame issues and they are recorded and preserved less often. Legislators are unlikely to frame issues on the floor in drastically different ways than they do in hearings, press interviews other legislative settings.

Unlike interviews or surveys, legislators’ statements in public debate are aimed at political audiences and thus should reveal the frames that they believe will have the greatest impact. Although the final votes on most legislative measures that come to the floor are known in advance, legislators may share their thoughts “on the record” for a host of reasons: to demonstrate their expertise to their colleagues; to signal legislative intent to agencies and courts that will implement and interpret legislation; to curry favor with interest groups; or to explain their votes to their constituents. Using the floor to explain, a vote may be particularly important on controversial and salient legislation like abortion measures. In short, there may be several audiences for floor debates, regardless of whether the speeches change enough minds to alter the outcome.

We found many debates that lasted for hours and involved large numbers of legislators. We doubt that so many legislators would spend their scarce time discussing issues on the floor if no one was paying attention or if what they had to say had no impact. On issues as controversial as abortion, the mass media report on what legislators have to say in legislative debate. For example, floor debate on Texas House Bill 15, a mandatory sonogram policy, was extensively covered by local media, including the major daily newspaper, the Austin American-Statesman, and a non-profit media organization, the Texas Tribune. Texas representatives were quoted from the floor in a March 2011 front-page Statesman article that detailed Democrats’ efforts to amend and stall the bill (Eaton 2011). Media outlets in Florida extensively documented floor debate before votes on a series of abortion bills in Florida in 2011. One story describes pro-restriction House members quoting the Bible, likening abortion to genocide, and telling personal anecdotes of teen motherhood (Zink et al. 2011). Coverage of debate on abortion restrictions is regularly picked up by national media, as well (e.g., Condon 2011; Peralta 2012).

We obtained lists of all of the abortion restrictions adopted in the American states during 2011–2014 from the Guttmacher Institute (www.guttmacher.org), a research and education organization that tracks policy changes in sexual and reproductive health. Since there were more restrictions debated and adopted than we could possibly study, we decided to examine three restrictions: bans on abortions after 20 weeks of gestation or the detection of a fetal heartbeat⁸; more stringent licensing requirements for abortion clinics and providers (i.e., “TRAP” provisions); and “informed consent” requirements for women seeking abortions to undergo ultrasound exams. We chose these restrictions because they are among the most common types of restrictions adopted during this period and they provide variation on the type of restriction. We chose bans because they go beyond regulating access to abortion by prohibiting certain classes of abortion altogether. TRAP laws target providers of abortion and reduce access disproportionately to women living away from urban areas. TRAP provisions tended to be more controversial than ultrasound exam requirements, although mandatory “trans-vaginal” ultrasounds could be highly controversial as well.

Our data include the debates on all 26 proposals that reached the floor on these restrictions for all of the states for which data were available. Sixteen (16) states recorded their debates and made them available for this period: Arkansas, Arizona, Florida, Georgia, Louisiana, Michigan, Minnesota, New Hampshire, North Carolina, Ohio, Oklahoma, Tennessee, Texas, Utah, Virginia, and Wisconsin.⁹ Thus, for the restrictions that we are examining, we have examined the universe of cases for which data are available, not simply a sample of them.

Our data show that Republican majorities passed virtually all of the abortion restriction measures (over 96% of the speeches that we analyzed came from Republican-controlled legislatures) and that very few abortion restrictions brought up for floor debate failed to gain passage (less than 2% of the speeches that we analyzed are from debates that resulted in defeat). The twenty-six measures yielded 45 separate chamber debates (House and/or Senate) that included 736 speeches.

For each debate, our unit of analysis was an individual speech, but we also aggregated the speech-level data, making entire debates another unit of analysis. We identified each speech as either in favor of or opposed to the legislation. Some speakers gave multiple speeches during a single debate. The overwhelming majority of speeches included statements that contained frames, such as when speakers stated what they thought the issue or proposal was “about” or gave arguments in favor of or against the measure under discussion. We did not count speeches that contained no frames or arguments for and against adoption. In these instances, legislators typically offered parliamentary points of order or simply repeated the content of the bills or amendments under consideration. We followed a procedure consistent with studies of the content of media coverage and other textual material by developing a set of frames inductively as we listened and re-listened to the debates (see Chong and Druckman 2007, 106–108). We manually coded all of the speakers’ statements for each speech that contained frames.¹⁰ Two of us listened to the entire debates and then compared our transcriptions and coding. Each time that we heard a speaker assert a frame/argument, we recorded and counted it. If speakers reiterated the same frame in a single speech, we counted those instances as a single assertion of a frame. Our inter-coder reliability test revealed 91% agreement.

The Guttmacher Institute provides annual ratings of states according to their level of “hostility” toward abortion rights; we use the 2010 ratings that they published just before the outpouring of abortion restrictions that we studied.¹¹ We use the Association of Statisticians of American Religious Bodies’ (Grammich et al. 2012) 2010 US census of religion for measures of states’ level of membership in evangelical Protestant sects. The states in our dataset reflect a higher level of affiliation with evangelical Protestant sects than for all 50 states. The median was roughly 177 evangelicals per 1000 population for the states in our data set compared to 129 out of 1000 for all states. We include Lax and Phillips’s (2012, 152, 153) measure of state liberal public opinion. Overall, the states we examine are slightly less liberal than the 50 states. The median for these 16 states was 47% liberal public opinion across policy areas, while the median for all states was 48%.

Findings

Turning to our findings concerning how legislators frame the abortion issue, clearly both sides in the debate framed the issue along multiple dimensions. We recorded a large number and variety of frames—16 for the pro-restrictions side and 25 for the anti-restrictions side. Three of the restriction supporters’ frames and five of the opponents’ frames appeared in at least 20% of speeches (Table 1). While each side favored certain frames over others, no frame for either side reached the 40% mark, much less appeared in a majority of speeches overall. The “right to life” frame appeared in 38% of the speeches of restriction supporters and the “women’s choice” frame appeared in 28% of restriction opponents’ speeches.

Table 1 Top frames put forward by abortion restriction supporters and opponents (percent of speeches in which frame appears; frames that appear in less than 10% of speeches not shown)

Because these aggregate data could mask unidimensional framing at the level of individual debates, we looked at the 24 debates in which there were a total of 10 or more speeches to see how many featured a single frame that appeared in at least half of all speeches and where no other frame reached the 50% threshold. Using this rather low threshold, on the pro-restriction side, in only 29% (7) of the 24 cases did a frame reach the 50% threshold without any other frame also reaching it. Similarly on the anti-restriction side, only 25% of the cases met that threshold. If we increase the threshold to more than 50%, the proportions drop to 21% and 13% for the pro- and anti-restriction sides, respectively. We also looked at how consistently restriction supporters and opponents favored the “life” and “choice” frames across debates. In only one-third of the debates did restriction supporters use the “life” frame as their top frame, the same figure for opponents’ use of the “choice” frame. In a larger proportion of the debates those frames ranked below second place (in 46% of the debates for supporters and 54% of them for opponents).

Contrary to our expectations, the supporters of the status quo policy pursued a multidimensional framing strategy more than those who proposed policy change (i.e., new abortion restrictions). Restriction opponents put forward twice as many different frames as restriction proponents, using a threshold where frames had to appear in at least 10% of speeches (see Table 1).

We also looked to see whether legislative advocates pursued a strategy of conflicting frames that stressed the “life versus choice” value dichotomy, or, whether they converged on common values and themes. The speech-level data indicate that advocates on both sides pursued both strategies (Table 1). The respective sides emphasized the “right to life” and “women’s choice” frames, but each also converged with essentially the same frequency on frames that stressed the consequences for women of having (or not having) the restrictions in place. Restriction supporters talked about as much about how the restrictions would protect “women’s health and safety” as they did of “life”; restriction opponents talked as much about how the restrictions would place financial, practical and emotional “burdens on women” as they did the “women’s choice” principle.¹²

We also measured value convergence and divergence by grouping separate, but related, frames under what we call broad dimensions or “meta-frames.” The “life” meta-frame, for example, includes the frames “right to life/protecting life/abortion is murder,” “fetal development/fetus is human,” “respect for fetal remains,” and “abortion is never justified.” The “choice” meta-frame includes “women should decide/get to choose abortion/privacy,” “abortion is legal/constitutional right/restrictions are unconstitutional,” “restrictions constitute gender discrimination,” and “restrictions discriminate against rural women.” Similarly, the “consequences for women” meta-frame included several specific frames that both sides put forward related to a variety of impacts that women would experience under the restrictions, or if women chose to terminate or continue a pregnancy (Table 2).¹³ A third meta-frame relates to the “medical-scientific” aspects of abortion.¹⁴

Table 2 Meta-frame usage

We see a similar pattern with meta-frames as we did when we looked at how often speakers used the individual “life” and “choice” frames. Supporters asserted the “life”-related frames and the “consequences for women”-related frames with the same frequency—in a little over half of their speeches. Similarly, restriction opponents preferred the “consequences for women”-related frames over the “women’s choice”-related frames, 53–49% of speeches, respectively (Table 2). By framing the issue in terms of consequences to women, restriction supporters were embracing the idea of abortion as a “women’s issue” and thus appropriating a frame generally associated with abortion rights proponents. (Similarly, opponents appropriated the “excessive government intrusion” frame usually associated with their conservative opponents, but they did so only in 13% of their speeches, ranking it low compared to other frames (see Table 1).

We also proposed that both sides might converge on another theme—by trying to “medicalize” the issue by invoking the authority of science and medicine and discussing how the restrictions would affect the practice of medicine. About half of the opponents’ speeches fit under this “medical/scientific” meta-frame. Restriction supporters gravitated to the medical frame much less. Only about 9% of their speeches fit under this rubric, less than expected given the media’s attention to recent research on the “science of fetal pain.”¹⁵

Turning to the key contextual variables in our study, we tested several hypotheses related to when we expected legislators to frame the issue along the “life” versus “choice” dimension that the morality politics literature predicts would dominate debates over abortion legislation (Table 3). The results of the logistic regression predicting whether or not a speech contained the “life” or “choice” frames show that two of the model variables—the share of the opposition using “life” or “choice” arguments and the type of restriction were significantly associated with a speaker’s likelihood of using “life/choice” framing. Legislators often followed the “tit-for-tat” strategy. They were more likely to put forward the “life”/“choice” frames if a larger share of the preceding speeches of the opposing side contained these frames. For example, an anti-restriction speaker debating a ban would have a 29% predicted probability of using the “women’s choice” frame, holding other variables at typical values, if she had followed pro-side speakers who did not use the “right to life” frame. If the same anti-side speaker followed pro-side speakers, all of whom used the “right to life frame,” she would have a 58% predicted probability of using the “choice” frame. This finding suggests that each side’s discursive strategy strongly influences the other side’s behavior. In this case, the more often that one side in the debate frames the issue in terms of “morality politics,” the more it induces the other side to follow suit. More broadly, it suggests that advocates are not “talking past” each other by disregarding the arguments of the other side. Even when each side frames the issue in terms of very different and competing substantive principles and arguments, as in this case, both sides put forward the same type of frame (i.e., one that emphasizes basic moral principles rather than practical consequences). There is added empirical evidence from the stem cell research debate that when one side employs moral arguments, that prompts the other side to also respond with moral argument (Clifford and Jerit 2013).

Table 3 Logistic regression predicting whether a speech contained “life/choice” frames

The type of abortion restriction under consideration also shaped legislators’ framing strategies. Proposals to ban certain abortions induced speakers to stress the “life”/“choice” arguments at significantly higher rates than those that would reduce access to abortion (TRAP requirements) or discourage women from getting them (requiring ultrasounds). For example, a speech in a debate regarding a proposed ban had a 38% predicted probability of containing a “life/choice” frame when all other variables were set at typical values, while a speech in a debate over a restriction that did not ban abortions had a 22% estimated likelihood of using a “life/choice” frame. It is worth noting that the three debates that featured clear “life” versus “choice” frame divergence were over bills that included abortion bans. And among debates with at least 10 speeches, in all six debates over stand-alone bans, “life” was the top frame of the pro-restriction side.

None of the other variables in the model reached statistical significance, suggesting that the liberalism of states,¹⁶ the population-standardized count of evangelical Protestants in states, and state’s estimated hostility toward abortion rights had no consistent effects on whether or not speakers used “morality” framing.

Conclusion

Our findings have implications for our understanding of the politics of abortion, morality policy, and how policymakers frame issues generally. Obviously, our findings come with the caveat that we have to be careful in generalizing to other abortion restrictions. Although the three kinds of restrictions that we examined are found in many states and have been the subject of recent US Supreme Court rulings, we do not know if we would have obtained different results if we had looked at other restrictions. Also, we have examined only a single morality policy issue—abortion, although our results are consistent with recent studies of how advocates frame LGBTQ issues (Mucciaroni 2011), gambling (Ferriaolo 2013) and marijuana decriminalization (Ferriaolo 2014).

Overall, we found some empirical support for the morality policy perspective on how the abortion issue is framed. Many state legislators framed the issue in terms of a profound clash between the competing moral principles of “the sanctity of human life” versus “a woman’s right to choose.” However, morality frames did not dominate these debates on the whole for either side, certainly not at the level we expected given the disproportionately high levels of evangelical Protestant membership in the states we examined. First, the right to life frame appeared in only 2% more speeches than the “women’s health and safety” frame; the “choice” frame appeared in exactly the same proportion of speeches as the “restrictions burden women” frame. The two sides often converged by framing the issue in terms of the various consequences of abortion and abortion restrictions for women. Advocates were as likely to stress women’s physical, emotional and practical needs as they were the moral imperatives to protect life and women’s autonomy. Frames that focus upon women’s health and welfare are plausible, culturally resonant alternatives to the life versus choice binary. Second, morality frames did not appear in anything close to a majority of speeches for either side; nor did they rank first in a majority of the debates. Third, we found that both sides in the debate framed the issue along several dimensions, not just the life-choice binary. Therefore, we conclude that the morality policy perspective, at least as it applies to the battle over abortion restrictions in the American states, is incomplete and oversimplifies a more complex debate.

Whether they are trying to expand support beyond activists in the pro-life or pro-choice movements, explain their positions to constituents, or persuade undecided colleagues, legislators have plenty of reasons for moving beyond the “life versus choice” binary. While legislators may cast roll call votes along a single left–right dimension, as many studies have shown, the manner in which they frame issues and policies is much more varied and multidimensional, on this issue at least. While the mass media generally offers simplified accounts of debate as being about a single dimension or overriding argument for each side, actual debates are far more complex.

Even for an issue like abortion, which is perennially labeled a “morality policy” issue, the prevalence of “morality talk” varies with the context of debate. Advocates’ propensity to frame the issue in terms of life versus choice increased when one side or the other escalated rhetoric about “life” or “choice” principles (inducing the other to respond in kind). Our data thus conform to the logic of a game of tit-for-tat in which individuals follow a strategy of “retaliation” if their opponents frame issues in highly moralized, judgmental terms, or they “cooperate” by emphasizing how their preferred policy will promote some widely shared value (like women’s welfare or the authoritativeness of medical research). “Morality talk” was also more prevalent when the debates were about bans on abortion rather than other types of restrictions.

Thus, the broad implication of our findings is that the propensity of advocates to frame issues in terms of fundamental moral principles has less to do with the general subject matter or issue area (e.g., abortion) and more to do with strategic considerations. Those considerations in this case take into account the specific content of a legislative proposal (e.g., what kind of abortion restriction is under consideration) or the other side’s rhetoric (e.g., whether they frame the issue in moralistic language). Recent research on patterns of discourse on climate change suggests a growing trend in favor of framing that issue in terms of basic moral principles rather than scientifically informed arguments about the economic and health consequences of carbon emissions (Braun and Jorgens 2013; Nisbet 2009; Feinberg and Willer 2013). The rise in morality talk on the climate issue, which was once dominated by scientific and economic arguments, may be driven by advocates’ desire to find alternative frames in a highly competitive and stalemated policy debate or may reflect the growing polarization in American politics. Whatever may be the case, decisions to emphasize or de-emphasize fundamental moral principles and deontological reasoning have less to do with the fixed content of policy domains and more with strategic considerations that are shaped by changes in the policy-making context.

Finally, contrary to expectations in the literature on framing in policy-making, the side that preferred to maintain the status quo put forward a greater number and variety of frames. What appears to induce advocates to engage in more framing and to deploy a greater variety of frames is not whether they are challenging or favoring the status quo, but whether they are losing (or appear to have lost) the debate. Here, it was defenders of the status quo who engaged in more framing because they faced the threat of an imminent change in their state’s abortion policy. Because pro-restriction legislators controlled the chambers and were in practically every case heavily favored to prevail in the roll call vote, they had less incentive to debate than the anti-restriction legislators. Restriction opponents knew that their only chance to stop the legislation, or to simply register their disapproval, was to call attention to what was happening by speaking out as frequently as they could and by putting forward a variety of arguments.