Binding Morality and Perceived Harm as Sources of Moral Regulation Law Support Among Political and Religious Conservatives

Abstract

Conservatives—both political and religious—are more likely than liberals to support laws regulating traditional or religious morality. The current study applies a moral psychological framework to argue that the association between conservatism and moral regulation law support can be explained in part by binding morality, or a moral orientation that privileges group needs above individual needs and includes concerns about absolutism, collectivism, and purity. Binding morality may directly promote support for moral regulation laws, which enforce traditional rules, discourage deviance, and require adherence to customs regarding bodily and spiritual purity. Binding morality may also indirectly promote support for moral regulation laws by motivating perceptions that “immoral” behaviors are harmful to innocent victims. I test these hypotheses using three analytic samples drawn from the General Social Survey (N = 3669, 7120, 369). Results show that, as expected, binding morality (especially purity) mediates the effects of political and religious conservatism on laws regulating pornography, marijuana, abortion, and euthanasia. Additional analyses exploring views about pornography show that binding morality (especially collectivism) is associated with perceptions of immorality as harmful, which in turn is associated with support for regulation.

How much of a country's moral character should be dictated by law? Is the regulation of morality a matter of culture and tradition, or does it protect citizens from harm? These questions have long driven social and political debates about laws governing the private lives of citizens. Moreover, the governmental regulation of “immorality” may have consequences for the well-being and autonomy of citizens. For example, at the same time that debates over whether prostitution is legitimate work, sin, or sexual slavery characterize questions about its legalization (Weitzer 2009), prostitution laws endanger sex workers and further marginalize disadvantaged populations (Lerum and Brents 2016). Similarly, most Americans consider recreational marijuana use unproblematic (Schnabel and Sevell 2017), while the criminalization of marijuana contributes to racial disparity in the justice system and mass incarceration (Alexander 2012). Given that laws regulating morality are uniquely responsive to public opinion (Kreitzer 2015; Mooney and Schuldt 2008), understanding public preferences regarding the legal regulation of “immorality” may be key to understanding the contours of law and social control in the United States.

Laws regulating morality touch many facets of life, from sex, substance use, and gambling, to fundamental issues of life and death such as abortion or assisted suicide. Despite the variety of behaviors addressed by moral regulation laws, people who support one form of moral regulation often support others (Miller 1996; Schnabel and Sevell 2017). Moral regulation law support is also concentrated among individuals who endorse a conservative ideology, whether political or religious (e.g., Burdette et al. 2005; Ellison and McFarland 2011; Kreitzer 2015; Lykke and Cohen 2015; Sherkat and Ellison 1997). However, little research has explored why, at the individual level, conservatives are more likely than liberals to support moral regulation laws. More generally, despite the centrality of morality to moral regulation, little research has explored the role of moral cognition in shaping support for these laws (but see Sherkat and Ellison 1997). This gap in the research is especially noteworthy given advances in moral psychology suggesting that conservatives and liberals experience intuitions about different domains of morality (Graham and Haidt 2009; Haidt 2012).

The current study thus draws on moral psychological literature to propose a theoretical model to explain the association between conservatism and moral regulation law support. Briefly, conservatives may be more likely than liberals to have a “binding” moral orientation that elevates group well-being above individual well-being, and which may shape moral regulation law support both directly and indirectly. Binding morality may directly promote support for moral regulation laws, which elevate group well-being by upholding traditions, discouraging deviance, and requiring bodily and spiritual purity in accordance with social custom. Binding morality may also indirectly promote moral regulation law support by eliciting perceptions that violations of social or religious tradition are harmful to individual victims. This theoretical model is tested using three analytic samples drawn from the General Social Survey (GSS) in the years 2006–2010 (N = 3669), 1976–1994 (N = 7120), and 1988 (N = 369).

1. Conservatism and Support for Moral Regulation Laws

Broadly defined, morality policies are public policies involving conflicts about core moral issues that are salient to the general public and that lack technical complexity (Arsneault 2001; Ferraiolo 2014; Mooney and Lee 1995; Mooney and Schuldt 2008). They are distinct from nonmoral policies in that, typically, “at least one side of the debate frames the issue in terms of morality or sin, and the policies seek to validate one set of values at the expense of others” (Kreitzer 2015: 43). Morality policy is deeply embedded in American culture. As Morone (2003) argues in Hellfire Nation, defining less powerful groups as immoral and criminalizing their immorality has long allowed more powerful groups to structure and enforce social hierarchy within the melting pot of American culture. The result is a “panic-policy cycle” in which fears of a “dangerous them”—who threaten the status quo through immigration or suffrage—are stoked by the powerful, resulting in public panic and the creation of policies that regulate the behaviors of the less powerful (Morone 2003: 463), which persist until new social boundaries emerge and there is outcry that “good people face injustice” (Morone 2003: 4). Whereas Puritans used morality policy to separate good citizens from heretics and witches, and Victorian-era Americans were largely concerned with temperance and abstinence, morality policy in recent years addresses a variety of sins, as in laws regulating drugs, sex, and abortion (Morone 2003; see also Dombrink and Hillyard 2007; Hunt 1999).

The current study focuses in particular on explaining support for a subset of contemporary morality policies, referred to as moral regulation laws. These laws restrict practices that are: deemed immoral by social and/or religious tradition (which are inextricably linked in the United States; see Dombrink and Hillyard 2007); engaged in by private citizens, not organizations or institutions; and characterized by disagreement over the existence of a victim or victims (Meier and Geiss 1997). Moral regulation laws thus employ a “private behavior morality” frame, whereby “[o]pponents of gay rights, legalizing homosexuality, pornography, euthanasia, drugs, gambling, and prostitution judge these activities as sinful or unnatural and call upon the state to ban, restrict, or withhold recognition of them” (Mucciaroni 2011: 194). Moral regulation laws thus affect the private lives of citizens in many domains, including sexual deviance (e.g., prostitution or pornography), antiascetic behaviors (e.g., recreational drugs or gambling), and personal decisions about life and death (e.g., abortion or euthanasia).

Despite seemingly disparate behaviors encompassed by moral regulation laws, people who support one moral regulation law tend to support others (Miller 1996; Schnabel and Sevell 2017), and this support is ideologically divided. Although political conservatism is characterized by preferences for limited governmental involvement in economic redistribution (Stimson 2004), and religious conservatism encompasses belief in a literal interpretation of holy texts, strict adherence to traditional practices, evangelism, and belief in a punitive God (Smith 1990), political and religious conservatives have long been united in their preferences for stricter moral regulation (e.g., Burdette et al. 2005; Ellison and McFarland 2011; Kreitzer 2015; Lykke and Cohen 2015; Schwadel and Ellison 2017; Sherkat and Ellison 1997; Thomas 2013). Research at the individual level generally seeks to link political conservatives' support for moral regulation laws to cognitive structures, traits, or motives specific to political conservatism, such as perceptions of government as a “strict father” rather than a “nurturing parent” (Lakoff 2002), preferences for protection over provision (Janoff-Bulman 2009), or tolerance of change and inequality (Jost et al. 2003). Explanations for religious conservatives’ moral regulation law support, by contrast, often focus on biblical literalism (e.g., Burdette et al. 2005; Ellison and McFarland 2011; Sherkat and Ellison 1997). Relatedly, the “culture wars” perspective suggests that conservative support for moral regulation arises from the view that moral truths derive from “an external, definable, and transcendent authority” (Hunter 1992: 44), although beliefs about moral truth are thought to transcend either politics or religion.

Although moral beliefs are implicated in several of these explanations, research on conservatives' moral regulation law support has yet to incorporate recent developments in moral psychology. This is an important oversight given that conservatives and liberals are thought to experience intuitive moral concerns about different domains [Correction added on March 12, 2021, after first online publication: a typo is removed after original publication.] (Graham and Haidt 2009; Haidt 2012; Haidt and Hersh 2001; Miles and Vaisey 2015; Silver and Silver 2017), which may have implications for understanding preferences regarding law and punishment (Silver 2017; Vaughn et al. 2019).

2. Conservatism and Binding Morality

Interest in morality among social scientists has seen an upsurge in recent years (Stets and Carter 2012), and research indicates there is utility in applying moral psychological theory to understanding ideological divides in the United States (Miles and Vaisey 2015) and individual engagement with culture more generally (Hitlin and Harkness 2017; Miles et al. 2019). One influential perspective is Moral Foundations Theory (MFT; Haidt 2007, 2012), which argues that moral judgments are both intuitive (feelings first) and pluralistic (covering multiple domains).

Like many cognitive phenomena, moral judgments are typically formed via a dual process system: initial, intuitive judgments arise spontaneously in response to stimuli and are experienced as “gut feelings,” and these initial moral intuitions motivate and inform deliberate reasoning (Haidt 2007, 2012). Along these lines, studies provide evidence of “moral dumbfounding,” where individuals rate scenarios as wrong but find it difficult to explain why, or else offer explanations that are incongruent with facts (Haidt 2012). For example, in one well-known study, respondents cited pregnancy risk as a reason that consensual incest between adult siblings was morally wrong, despite the risk being described as negligible (Haidt et al. 1993).

MFT also suggests that moral judgment is pluralistic, meaning that people experience intuitions about different domains of virtue and violation (i.e., “moral foundations”), including Care, Fairness, Authority, Loyalty, and Sanctity (Haidt 2007, 2012). The domains in which individuals experience moral intuitions are thought to be shaped by socialization and as a result, commonly endorsed “moral matrices” vary across cultures and subcultures (Graham and Haidt 2009; Haidt 2012; Haidt and Graham 2007).

Within the MFT framework, political and religious groups in the United States constitute subcultures that socialize members to experience different moral intuitions (Haidt and Hersh 2001). Conservatives emphasize “individualizing” morality centered on individuals' needs and rights (including the moral foundations Care and Fairness) as well as group-oriented or “binding” morality, which centers on respecting group hierarchies and traditions (Authority), maintaining in-group solidarity (Loyalty), and preserving bodily or spiritual purity according to group customs (Sanctity). Liberals, on the other hand, typically emphasize individualizing morality only, and do not view authority, loyalty, or sanctity as legitimate moral concerns (Graham and Haidt 2009; Haidt 2012; Miles and Vaisey 2015). Accordingly, research suggests that endorsement of binding morality, in which the needs of the collective are perceived to be paramount, marks an important difference in moral cognition among conservatives and liberals (Graham and Haidt 2009; Miles and Vaisey 2015).

Although some scholars have criticized MFT for failing to distinguish between moral foundations and conservative ideology more generally (e.g., Kugler et al. 2014), research suggests that MFT has both theoretical and empirical utility. Theoretically, moral foundations are considered “basic and generalized psychological tendencies” that “predispose individuals to certain sociopolitical beliefs” (Koleva et al. 2012: 188), and which thus provide insight into the processes underlying political preferences. Empirically, moral foundations are linked to political and religious preferences extending “above and below” a liberal-conservative continuum (Haidt et al. 2009) and explain variance in political ideology net of “core motives” of conservatism, such as tolerance of change and inequality (Yilmaz and Adil Saribay 2018).

Binding morality has also been implicated as a source of various sociolegal preferences, including views about deviance, crime, law, and punishment (Haidt and Graham 2007; Koleva et al. 2012; Rottman et al. 2014; Silver 2017; Silver and Silver 2017; Vaughn et al. 2019). More generally, research suggests that lay moral intuitions form the bedrock upon which legal codes are constructed (Robinson 2013). Accordingly, scholars argue that intuitive and pluralistic moral concerns, such as those described by MFT, may be key to understanding support for moral regulation laws (Tatalovich and Wendell 2018).

3. Binding Morality and Moral Regulation Laws

Theoretically, each of the binding moral foundations (i.e., Authority, Loyalty, and Sanctity) may be expected to shape support for moral regulation laws by eliciting intuitive perceptions of deviant behaviors as moral transgressions against the collective. The Authority foundation promotes intuitions that it is virtuous to follow traditions and maintain social hierarchies (Graham and Haidt 2009; Haidt 2012). A focus on tradition and hierarchy, in turn, may produce an absolutist worldview, wherein it is not acceptable to deviate from traditional or religious rules governing behavior (Leone et al. 2019; Miles and Vaisey 2015). Insofar as moral regulation laws require individuals to conform to traditional social and religious rules, moral concerns based in Authority (including moral absolutism) may be associated with support for moral regulation laws. The moral foundation Loyalty is thought elicit a “Durkheimian” view of the world (Graham and Haidt 2009) rather than a strictly absolutist one. Moral intuitions about the cohesion and well-being of the collective promote perceptions that there exists a shared set of moral rules—a “collective conscience”—that may be threatened or corrupted by deviant or selfish individuals (Graham and Haidt 2009; Silver and Silver 2017). Moral collectivism may, in turn, promote support for moral regulation laws, which restrict private citizens' involvement in behaviors that violate norms and threaten social cohesion. Finally, the moral foundation Sanctity may also enhance support for moral regulation laws. Sanctity promotes intuitions that it is wrong to engage in behaviors—particularly self-directed behaviors—that degrade spiritual or bodily purity. To the extent that moral regulation laws prohibit behaviors that lack personal victims but violate norms related to bodily purity (e.g., drug use) or spiritual purity (e.g., euthanasia), moral purity may also be associated with judgments that such behaviors are inherently wrong and should be regulated.

Little research has explored the role of binding morality in shaping support for moral regulation laws among conservatives and liberals. Most relevant to the current study, Sherkat and Ellison (1997) used GSS data to assess whether measures equivalent to moral absolutism and moral collectivism (hypothesized to originate from cognitive structures conferred by biblical literalism in religious communities) were associated with conservative Protestants' acceptance of pornography (a scale combing support for pornography laws, perceptions that pornography breaks down social morals, and perceptions that pornography leads to rape). Although both absolutism and collectivism were associated with support for pornography laws, conservative Protestantism exerted an indirect effect on acceptance of pornography through moral absolutism only. While this study provides an important contribution to the literature, questions remain, especially regarding binding morality as a source of attitudes about legal regulation specifically; binding morality as a source of attitudes about regulating other “immoral” behaviors; the role of moral purity in moral regulation attitudes; and whether binding morality can account for political conservatives' moral regulation attitudes.

4. Perceived Harm and Moral Regulation Laws

In arguing that binding morality underlies support for moral regulation laws, it is important to note that public discussions about moral regulation rarely invoke concerns associated with Authority, Loyalty, or Sanctity. Rather, moral regulation laws are almost invariably framed as preventing harm to individual citizens (Burlone and Richmond 2018; Euchner et al. 2013; Ferraiolo 2014; Gandsman 2016), even in internal communications among conservatives (Thomas 2013). Pornography entices consumers to rape innocent women (Lykke and Cohen 2015). Prostitution produces human trafficking and sexual slavery on a massive scale (Weitzer 2009). Marijuana kills people in traffic accidents and encourages youth tobacco use (Ferraiolo 2014; Meier 1999). Abortion results in the murders of fetuses or “partially born” infants (Esacove 2004). Assisted suicide paves the way to widespread elder abuse (Burlone and Richmond 2018; Gandsman 2016). Moreover, as Meier (1999: 692) noted decades ago, “the worst-case scenario often dominates the policy debate,” making the most salient arguments those that involve violence perpetrated against vulnerable, innocent victims. Research also documents that harm-centered arguments in support of moral regulation laws have grown more prevalent in recent decades, while arguments associated with binding morality have declined (Euchner et al. 2013; Ferraiolo 2014; Gandsman 2016; Hunt 1999; Mucciaroni 2011; Thomas 2013).

As other scholars have noted, however, one must avoid confusing “deep-seated beliefs about public policy issues with the arguments and frames [used] in public debate” (Mucciaroni 2011: 209), for, “even while they argue in terms of situational risk,” activists' support for moral regulation often “stems from deeply held and absolute moral values” (Gandsman 2016: 579). One explanation is that harm-based discourse represents an intentional persuasive choice made by conservative politicians and other framers of the debate (Mucciaroni 2011; Thomas 2013). Given that “individualizing” moral foundations such as Care (through which harm to individuals is perceived as a transgression) are endorsed by both conservatives and liberals (Haidt 2012), and framing messages in terms of individualizing morality may allow for influence across the political aisle (Miles 2016), this is a reasonable argument.

There also exists a second (and complementary) explanation for the prevalence of harm in discussions about moral regulation, which the current study explores. Research shows an empirical relationship between evaluations of actions as morally wrong and evaluations that they are harmful, even in the absence of a clear personal victim (e.g., Rottman et al. 2014; Schein and Gray 2016). To explain the association between perceived wrongness and perceived harm, scholars in the pluralistic tradition of moral psychology have noted that Western morality tends to emphasize individual well-being, and thus, in explaining why they perceive violations of binding morality as wrong, individuals may draw upon a cultural narrative that positions harm as the ultimate moral violation (Rottman et al. 2014). That is, even if people intuitively perceive an action as wrong because it violates a social or religious tradition, they may explain their perceptions of wrongness by reasoning ways in which the violation may be harmful to individuals. Indeed, research suggests people are capable of “recruiting” harms as needed to justify support for laws criminalizing victimless transgressions, indicating that “the concept of harm may not be as cognitively stable or reliable as the legal system assumes” (Sood and Darley 2012: 1346). Thus, to the extent that conservatives are more likely than liberals to intuit that violations of social or religious norms are wrong, they may also be more likely to engage in moral reasoning that leads them to perceive those violations as harmful. Moral regulation laws, in turn, may seem necessary to protect innocent victims from the consequences of immorality.

5. The Current Study

The current study tests a two-part theoretical model regarding the role of binding morality (including concerns about absolutism, collectivism, and purity) in linking conservatism and support for moral regulation laws. First, conservatives' binding morality may influence support for moral regulation laws by eliciting intuitions that violations of social and religious norms are wrong. Second, conservatives' binding morality may promote reasoning that violations of social and religious norms are harmful to innocent victims and should be regulated by law. Using items gauging support for four types of moral regulation laws (pornography, marijuana, abortion, and euthanasia), I test the following hypotheses:

1. 1. Political and religious conservatism are positively associated with moral regulation law support.

2. 2. Binding morality is positively associated with moral regulation law support.

3. 3. Binding morality mediates the relationships between political and religious conservatism and moral regulation law support.

Using an additional set of items centered on pornography, I also test the following hypotheses regarding perceived harm and moral regulation law support:

1. 4. Political and religious conservatism are positively associated with perceived harm.

2. 5. Binding morality is positively associated with perceived harm.

3. 6. Binding morality mediates the relationships between political and religious conservatism and perceived harm.

4. 7. Perceived harm mediates the relationship between binding morality and support for laws regulating pornography.

6. Data

Data are drawn from the GSS, a nationally representative survey of noninstitutionalized adults living in the United States conducted annually or biannually by the National Opinion Research Center using a stratified multistage cluster design. Consistent with prior studies using GSS data pooled across years (Silver and Pickett 2015), all models control for the year of the interview. Data are weighted using the strategy developed by the Computer-Assisted Survey Methods Program at the University of California, Berkley (http://sda.berkeley.edu/D3/GSS12/Doc/gs12i01.htm).

7. Measures

GSS items are used to conduct three sets of analyses structured around the availability of binding morality and perceived harm measures. Items measuring moral absolutism and moral collectivism are available from 2006 to 2010 and in 1988, while items measuring moral purity are available from 1976 to 2018. Data from 2006 to 2010, which include these three binding morality measures, are used to predict support for four moral regulation laws (pornography, marijuana, abortion, and euthanasia). Available from 1976 to 1994 are items asking about various attitudes toward pornography, including its perceived harmfulness and the degree to which it weakens social values; these measures, along with the moral purity measure, are used to assess the roles of perceived harm and binding morality in predicting support for pornography laws. Finally, as all binding morality and perceived harm measures were included in 1988 GSS, data from this year is used to estimate full models replicating the previous analyses with all available measures. Measures of political and religious conservatism, as well as demographic controls, are available in all relevant years. Each measure is described below.

7.1 Support for Moral Regulation Laws

7.1.1 Pornography

Although the filming and sale of pornography is currently legal in all U.S. states, obscenity laws were the subject of rigorous debate in the 1970s and 1980s (Weitzer 2009), and as of 2018 only a minority (43 percent) of Americans believed that pornography is morally acceptable (Gallup 2018a; 2018b). Support for Pornography Regulation is measured using a single item: “Which of these statements comes closest to your feelings about pornography laws?” Respondents who answered “There should be laws against the distribution of pornography whatever the age” are coded “1” (support regulation), whereas respondents who answered “There should be laws against the distribution of pornography to persons under 18” or “There should be no laws forbidding the distribution of pornography” are coded as “0” (oppose regulation).¹

7.1.2 Marijuana

Unlike pornography, recreational marijuana use remains illegal in most U.S. states, although a growing number of states (including Alaska, California, Colorado, Illinois, Maine, Massachusetts, Michigan, Nevada, Oregon, Vermont, and Washington) have legalized recreational use. Public support for marijuana legalization has increased over the past several decades (e.g., Schnabel and Sevell 2017), and as of 2019, only 32 percent of Americans agreed that marijuana should be illegal (Pew Research Center 2019a, 2019b). The item for Support for Marijuana Regulation asks respondents, “Do you think the use of marijuana should be made legal or not?” Responses are coded 1 = “Not legal,” (support regulation) and 0 = “Legal” (oppose regulation).

7.1.3 Abortion

Although abortion has been legal in all U.S. states since the landmark 1973 Roe v. Wade Supreme Court case, passionate debates regarding government restrictions on abortion continue and as of 2019, 38 percent of Americans agreed that abortion should be illegal in all or most cases (Pew Research Center 2019a; 2019b). The item measuring Support for Abortion Regulation asks whether respondents “think it should be possible for a pregnant woman to obtain a legal abortion if the woman wants it for any reason.” Responses are coded 1 = “No” (support regulation) and 0 = “Yes” (oppose regulation).

7.1.4 Euthanasia

Assisted suicide refers to ending one's life using substances prescribed by a physician, while voluntary euthanasia describes the practice of ending one's life with aid from a physician. Currently, assisted suicide is illegal in all but six states (Hawaii, California, Oregon, Montana, Vermont and Washington) and the District of Columbia, and voluntary euthanasia is not legal in any state. As of 2018, however, only 27 percent of Americans reported that they thought voluntary euthanasia for terminally ill individuals should be illegal (Gallup 2018a, 2018b). Support for Euthanasia Regulation is measured using the question, “When a person has a disease that cannot be cured, do you think doctors should be allowed by law to end the patient's life by some painless means if the patient and his family request it?” Responses are coded 1 = “No” (support regulation) and 0 = “Yes” (oppose regulation).

7.2 Binding Morality and Perceived Harm

This study draws on several binding morality and perceived harm measures.² Two binding morality measures (moral absolutism and moral collectivism) are available in 1988 and from 2006 to 2010. Moral Absolutism is measured using two items: “Morality is a personal matter and society should not force everyone to follow one standard” and “Right and wrong are not usually a simple matter of black and white; there are many shades of gray.” Both are measured on five-point scales (coded 1 = Strongly agree to 5 = Strongly disagree) and averaged to form a measure (r = 0.201).³ Moral Collectivism is measured using a single item, that is, ., “Immoral actions by one person can corrupt society in general” (1 = Strongly disagree to 5 = Strongly agree), indicating a perception that the collective is threatened by immorality or deviance.

A third binding morality measure, Moral Purity, is available from 1976 to 2018. Moral intuitions related to Sanctity are strongly linked to views about the moral acceptability of deviant sexual practices (Graham et al. 2011; Haidt 2012; Haidt and Hersh 2001). Thus, as a proxy for moral purity, I use items assessing the degree to which respondents perceive nontraditional forms of sex to be wrong. Specifically, respondents were asked to judge the moral acceptability of “sexual relations between two adults of the same sex” as well as “a married person having sexual relations with someone other than his or her husband or wife” (coded 1 = Not wrong at all to 4 = Always wrong; r = 0.283).⁴ These items were averaged to form a scale.

A fourth binding morality measure and a measure of perceived harm are available from 1976 to 1994. One item captures concern about pornography as a threat to collective morality and is analogous to the item assessing general moral collectivism in the first part of the analysis: specifically, respondents were asked whether they thought that “Sexual materials lead to breakdown of morals.” Thus, pornography-specific Moral Collectivism was coded as Yes = 1 and No = 0. Respondents were also asked whether “Sexual materials lead people to commit rape,” reflecting a view that pornography harms innocent victims. This item was used to measure Perceived Harm and was coded Yes = 1, No = 0. Note that the pornography-specific perceived harm and moral collectivism items are correlated strongly at r = 0.58 in the 1976–1994 analytic sample and r = 0.62 in the 1988 analytic sample. Implications are addressed in the discussion.

7.3 Ideology and Demographics

Political Conservatism was measured using a single item asking respondents to self-identify their political preference on a seven-point scale (1 = Extremely liberal to 7 = Extremely conservative). Religious Conservatism was measured on a three-point scale (1 = Religious liberal, 2 = Religious moderate, 3 = Religious fundamentalist); see Smith (1990).⁵ Note that, as expected, both conservatism measures are positively correlated with each of the binding morality measures (see Tables A1, A2, A3 in Appendix A); however, the correlations are weak to moderate in size (ranging from 0.12 to 0.35), suggesting that discriminant validity is not an issue for these measures. Additionally, as in research using validated moral foundations scales (e.g., ), supplemental analyses (not shown) indicated that binding morality scores predicted the strength of conservative beliefs among self-identified moderates and conservatives, and among self-identified conservatives only, providing confidence in the validity of the measures.

I also include controls relevant to understanding support for moral regulation laws.Haidt 2007; 2012⁶ Race is coded 1 = White and 0 = Non-White. Sex is coded 1 = Female and 0 = Male. Age is measured in years from age 18 until age 89, at which point all respondents 89 or older are included in a single category. Marital status is coded 1 = Married and 0 = Unmarried. The presence of children under age 18 in the respondent's household is coded 1 = Children in household and 0 = No children in household. Education is coded 1 = less than high school; 2 = high school degree; 3 = junior college degree; 4 = college degree; and 5 = graduate degree. Household income is measured using the GSS variable “realinc,” which provides household income in dollars standardized to a 1986 inflation rate (divided by 1000). City residence is measured as living in a city of at least 50,000 residents (1 = Urban residence, 0 = Nonurban residence). Residence in the South is coded 1 = Southern residence, 0 = Non-southern residence.

8. Analysis Plan

As noted previously, the analysis proceeds in three stages. The first set of models uses data from 2006 to 2010 to estimate logistic regression models predicting support for moral regulation laws (pornography, marijuana, abortion, and euthanasia). The use of four outcome variables, which differ in the behaviors involved, legal status, and level of public support, provides evidence regarding the applicability of the theoretical model to laws addressing different types of behaviors. Base models include political and religious conservatism and controls, and additional models introduce binding morality measures (i.e., moral absolutism, moral collectivism, and moral purity).

The second set of models uses data from 1976 to 1994 to examine the role of perceived harm in support for pornography laws. I use logistic regression models to predict perceived harm from political and religious conservatism and controls, then introduce binding morality measures (i.e., porn-specific moral collectivism and moral purity). Next, I predict support for pornography regulation laws from conservatism and binding morality measures and assess whether perceptions of harm mediate the effects of binding morality on support for pornography laws. This section includes a replication of the 2006–2010 models predicting support for pornography laws from conservatism and binding morality. Just as the use of multiple dependent variables is helpful in assessing the robustness of the results to different outcomes, the 1976–1994 analyses provide a test of the robustness of the results to modeling choices (i.e., different morality variables) and to the historical context in which the data were collected.

The third analysis uses data from 1988, the only year in which all binding morality and perceived harm measures (i.e., moral absolutism, general and porn-specific moral collectivism, and moral purity) are available. Here, I estimate full models predicting perceived harm and support for pornography laws. Although the power is reduced compared to the first two analyses, these models provide two additional benefits: they allow for an assessment of the relative contributions of all binding morality measures to perceived harm and moral regulation law support; and they show the effects of the absolutism and collectivism measures used in the 2006–2010 analysis in an earlier decade.

Where appropriate, I use the Karlson-Holm-Breen (KHB) method to test for mediation.⁷ This user-written Stata program provides an estimate of the difference in coefficients in nested logistic regression models independent of the rescaling effects that occur when a mediator variable has independent effects on a dependent variable (see Karlson et al. 2012; Kohler et al. 2011). Specifically, the KHB method adjusts for rescaling by including residuals from a regression of the outcome on all covariates (estimated as an intermediate step) as a variable in the initial “crude” model. Evaluation research suggests KHB is an appropriate method for addressing bias from rescaling in logistic regression models (Kohler et al. 2011). The program also provides estimates of the disaggregated effects of multiple mediators of multiple independent variables, net of controls.

The analysis employs three analytic samples: cases with complete data on the relevant variables from 2006 to 2010 (N = 3669); cases with complete data on the relevant variables from 1976 to 1994 (N = 7120); and cases with complete data on the relevant variables in 1988 (N = 369).⁸ Descriptive statistics for each analytic sample appear in Table 1. Bivariate correlations for all variables in each analytic sample are provided in Appendix A (Tables A1, A2, A3). Variance inflation factors for all variables are below the standard cutoff of 2 (the largest is 1.6), indicating that multicollinearity is not an issue in any model.

Table 1 Descriptive Statistics

Note: Standard deviations are omitted for dummy variables.

^a Split ballot item with smaller N.

9. Supplemental Analyses

Three supplemental analyses provide additional insight into the findings. First, to assess whether the effects of the morality measures differ when included separately in the models, I estimate supplemental analyses with the binding morality measures disaggregated (see Appendix B). Second, to assess the sensitivity of the findings to time period and to examine whether the effects hold in recent years, I examine the effects of moral purity across all years in which it is available (i.e., 1976–2018) and in the most recent survey years only (2012–2018) (see Appendix C). Third, to examine whether the findings are specific to moral regulation laws rather than conservative policy preferences more generally, I examine the effects of binding morality on support for nonmoral (i.e., economic) policy preferences (see Appendix D).

10. Results

10.1 Binding Morality and Support for Moral Regulation Laws (2006–2010)

Table 2 presents the results of logistic regression analyses predicting support for laws regulating pornography (models 1 and 2), marijuana (models 3 and 4), abortion (5 and 6), and euthanasia (models 7 and 8) from 2006 to 2010. As expected in the base models (1, 3, 5, and 7), both political and religious conservatism are associated with support for moral regulation laws. Indeed, political and religious conservatism are the only variables that consistently predict support for all four moral regulation laws. The next set of models (2, 4, 6, and 8) introduces the binding morality measures. The strongest and most consistent effects are for moral purity, which range from standardized effect sizes of b = 0.461 (marijuana) to b = 0.554 (euthanasia). Additionally, moral absolutism predicts support for each moral regulation law, but with modest standardized effect sizes ranging from b = 0.235 (abortion) to b = 0.326 (euthanasia). By contrast, moral collectivism is significantly associated with support for abortion regulation only, and the standardized effect is smaller (b = 0.098).

Table 2 Logistic Regression Models Predicting Support for Moral Regulation Laws (2006–2010)

Abbreviations: DV = dependent variable; LSE = linearized standard Error.

Note: 2006 is excluded as a reference category for “year.”

^a Standardized variable.

^*p < 0.05; ^**p < 0.01; ^***p < 0.001.

KHB analysis shows that binding morality partially mediates the relationship between conservatism and moral regulation law support. Taken together, the binding morality measures significantly (p < 0.001 for all) mediate the effects of political conservatism on support for moral regulation laws, reducing the effect by 58 percent for pornography (indirect effect b = 0.192); 38 percent for marijuana (b = 0.158); 32 percent for abortion (b = 0.183); and 52 percent for euthanasia (b = 0.219). The binding morality measures also significantly (p < 0.001) mediate the effects of religious conservatism, reducing the effect by 46 percent for pornography (b = 0.162); 65 percent for marijuana (b = 0.146); 39 percent for abortion (b = 0.133); and 40 percent for euthanasia (b = 0.134). Additionally, across all models, moral purity accounts for between 69 percent (pornography) and 71 percent (euthanasia) of the indirect effect for political conservatism, and between 75 percent (euthanasia) and 80 percent (pornography) of the indirect effect for religious conservatism. Moral absolutism also contributes to the mediation of both forms of conservatism, accounting for between 23 percent (abortion) and 28 percent (euthanasia) of the indirect effect for political conservatism, and between 13 percent (pornography) and 24 percent (euthanasia) of the indirect effect for religious conservatism. By contrast, moral collectivism accounts for between 1 percent (euthanasia) and 8 percent (pornography) of the indirect effect for political conservatism, and between 1 percent (euthanasia) and 5 percent (abortion) of the indirect effect for religious conservatism.

Taken together, these results are consistent with the hypotheses regarding binding morality and support for moral regulation laws: binding morality is associated with moral regulation law support, and binding morality measures partially explain the relationships between political and religious conservatism and moral regulation law support. Moreover, the effects of binding morality are strikingly similar across models predicting the regulation of pornography, marijuana, abortion, and euthanasia. The consistency of effects is especially notable for moral purity, which is operationalized using items asking about sexual practices, but which has similar effects on support for laws addressing both sexual and nonsexual behaviors.

10.2 Perceived Harm and Support for Pornography Regulation Laws (1976–1994)

The next set of models examines relationships among conservatism, binding morality, perceived harm, and support for laws regulating pornography from 1976 to 1994. Table 3 presents logistic regression models predicting the perceived harmfulness of pornography (models 1 and 2), as well as models predicting support for laws regulating pornography (models 3–5).

Table 3 Logistic Regression Models Predicting Perceived Harm and Support for Moral Regulation Laws (1976–1994)

Abbreviations: DV = dependent variable; LSE = linearized standard error.

Note: All models include dummy variables (not shown) for the years 1980, 1984, 1987, 1988, 1989, 1990, 1991, 1993, and 1994, with 1976 excluded as a reference category.

^a Standardized variable.

^*p < 0.05; ^**p < 0.01; ^***p < 0.001.

Model 1 indicates that, as expected, political and religious conservatism are both significantly and positively associated with perceived harm. Model 2 shows that binding morality also predicts perceived harm as expected. In these models, porn-specific moral collectivism has a larger standardized effect (b = 1.093) than moral purity (b = 0.310), although both are significant predictors of perceived harm. KHB analysis indicates that binding morality significantly (p < 0.001) mediates the effects of conservatism on perceived harm, reducing the coefficient by 59 percent for political conservatism (indirect effect b = 0.220) and 97 percent for religious conservatism (b = 0.180). Approximately 73 percent of the indirect effects for both forms of conservatism derive from moral collectivism, while 26 percent derive from moral purity.

Models 3 and 4 replicate the first part of the analysis using the 1976–1994 measures and data. Model 3 shows that political and religious conservatism both predict support for pornography regulation, with effect sizes comparable to those in the 2006–2010 models. In model 4, binding morality is again associated with support for pornography laws. Although the standardized effect for moral purity (b = 0.511) is similar to the effect for purity in the previous models, pornography-specific moral collectivism has a much larger effect (standardized b = 0.754) than did general moral collectivism. As before, the binding morality measures significantly (p < 0.001) mediate relationships between conservatism and support for pornography regulation (indirect effect b = 0.210 for political; b = 0.171 for religious). For both forms of conservatism, porn-specific moral collectivism accounts for approximately 53 percent of the indirect effect, while purity accounts for approximately 47 percent.

Model 5 introduces perceived harm. As anticipated, perceived harm is significantly associated with support for pornography regulation, and significantly (p < 0.001) mediates the relationship between binding morality and support for pornography regulation: the effect for moral collectivism is reduced by 28 percent (indirect effect b = 0.441) and the effect for moral purity is reduced by 9 percent (b = 0.047).

Overall, these results are consistent with the hypotheses regarding the role of perceived harm in moral regulation law support. That is, binding morality appears to be associated with perceived harm, which in turn provides a pathway by which binding morality promotes moral regulation law support. However, both binding morality measures also continued to exert direct effects on moral regulation law support net of perceived harm, suggesting that other moral concerns remain relevant even after accounting for harm.

10.3 Full Models (1988)

The final set of analyses draws on data from 1988 to replicate the previous analyses using all binding morality measures: moral absolutism, general and porn-specific moral collectivism, and moral purity. Table 4 presents logistic regression models predicting perceived harm (models 1 and 2) and support for pornography regulation (models 3–5). In model 1, neither political nor religious conservatism is significantly associated with perceived harm, although the effect sizes are similar to those in the previous analyses. In model 2, both moral collectivism measures are strongly and significantly (p < 0.001) associated with perceived harm (standardized b = 1.635 for porn-specific moral collectivism; b = 0.563 for general moral collectivism). Neither moral absolutism nor moral collectivism has a significant effect on perceived harm. KHB analysis indicates that binding morality does not significantly mediate the effects of conservatism on perceived harm, although the indirect effects are again comparable to those in the previous analyses (b = 0.194 for political; b = 0.270 for religious, with 102–103 percent reduction in coefficient sizes for both).

Table 4 Full Logistic Regression Models Predicting Perceived Harm and Support for Moral Regulation Laws (1988)

Abbreviations: DV = dependent variable; LSE = linearized standard error.

Note: All models include dummy variables (not shown) for the years 1980, 1984, 1987, 1988, 1989, 1990, 1991, 1993, and 1994, with 1976 excluded as a reference category.

^a Standardized variable.

^*p < 0.05; ^**p < 0.01; ^***p < 0.001.

The final set of models predicts support for pornography laws in 1988. Model 3 shows that religious, but not political, conservatism is associated with pornography law support in this year. In model 4, moral purity is associated with support for pornography laws (b = 0.687), as is porn-specific moral collectivism (b = 0.620). Neither general moral collectivism nor moral absolutism is associated with moral regulation law support in this year. KHB analysis shows that binding morality mediates the effects of religious conservatism on moral regulation law support (indirect effect b = 0.246, p = 0.012), and although the coefficient for political conservatism is not significant in the base model, there is a significant indirect effect for political conservatism (b = 0.229, p = 0.020). In model 5, perceived harm is again associated with support for pornography regulation. KHB indicates that only porn-specific moral collectivism exerts an indirect effect on moral regulation law support via perceived harm (indirect effect b = 0.673, p = 0.001).

These results largely corroborate the findings from the first two analyses. Binding morality is associated with support for moral regulation laws and mediates the effects of conservatism. Binding morality, especially moral collectivism, also exerts indirect effects on moral regulation law support via perceived harm. However, moral collectivism appears to be more strongly related to moral regulation law support in the 1988 models than in the 2006–2010 models, while the opposite is true for moral absolutism.

11. Supplemental Analyses

11.1 Disaggregated Effects of Absolutism, Collectivism, and Purity

The first set of supplemental analyses addresses whether the effects of the binding morality measures on moral regulation law support and perceived harm differ when the measures are included in separate models (see Appendix B). Table B1 indicates that from 2006 to 2010, the effects of purity and absolutism on support for pornography, marijuana, abortion, and euthanasia laws are substantively similar to those in the main analysis in their size, valence, and significance. Unlike in the main analysis, collectivism is also a significant predictor of moral regulation law support when included in the models separately, although the effect sizes are relatively small. Table B2 shows that in predicting perceived harm and pornography law support in the 1976–1994 and 1988 analytic samples, the effects of the binding morality measures are generally comparable to the main results in their size, valence, and significance, with two exceptions in 1988: purity is significantly associated with perceived harm, while absolutism is significantly associated with pornography law support. Thus, it appears that while some binding morality effects are masked by the inclusion of other measures, the overall pattern of results from the disaggregated models is similar to the main analyses.

11.2 Purity Effects 1976–2018

The second set of supplemental analyses assesses the effects of purity (the only binding morality measure available after 2010) on moral regulation law support in the full range of years in which the purity measure is available (1976–2018) and in recent years only (2012–2018) (see Appendix C). Table C1 indicates that the standardized effects of purity in the 1976–2018 sample and in the 2012–2018 sample are similar in size, valence, and significance to those in the main analysis as well to the effects in the purity-only models shown in Appendix B. Additional analysis including survey year as a linear variable (not shown) indicated that purity significantly interacted with year in a negative direction for marijuana (b = −0.004, p = 0.019), in a positive direction for euthanasia (b = 0.006, p = 0.012), and did not significantly interact with year for pornography (b = −0.002, p = 0.375) or abortion (b = 0.001, p = 0.460). Thus, it appears that the effects of purity are relatively similar in all years examined, and although there have been changes in the effects of purity on moral regulation law support, those effects have not been consistent across different types of moral regulation.

11.3 Binding Morality and Nonmoral Policy Support

The final set of supplemental analyses examines whether binding morality measures are associated with preferences for nonmoral (i.e., economic) policy, specifically, government size and economic redistribution (see Appendix D). Views about the size of federal government are measured using the GSS variable helpnot (coded 1 = “Government should do more” to 5 = “Government should do less”). Views about economic redistribution are measured using the GSS variable eqwlth (coded 1 = “Government should reduce income differences” to 7 = “Government should take no action”). These models use data from the years 2006–2010, in which the three key binding morality measures are available. Table D1 shows that political (but not religious) conservatism predicts both outcomes, which is consistent with the assertion that economic policy preferences are central to political ideology (Stimson 2004). Neither purity nor collectivism is associated with economic policy support, and although absolutism is associated with preferences for smaller government and less economic redistribution, the standardized effect sizes are relatively small. The latter finding is consistent with the notion that although they are not typically included under the umbrella of morality policies (Mooney and Schuldt 2008), economic policies do invoke moral considerations (Ryan 2014). Overall, however, it appears that the results obtained for economic policy preferences are substantively different from the results obtained for moral regulation law support.

12. Discussion

This study sought to provide evidence regarding two questions: first, whether conservative support for moral regulation laws derives from binding morality; and second, whether binding morality also contributes to support for moral regulation laws by eliciting perceptions of immoral behavior as harmful to individual victims. Overall, the results supported the proposed theoretical model. Binding morality fully or partially mediated the relationships between conservatism and moral regulation law support, and also explained conservatives' perceptions of immorality as harmful, which were in turn associated with greater moral regulation law support. For the most part, the effects were similar for political and religious conservatism, including the extent to which binding morality measures mediated the effects of conservatism on each outcome. Moreover, with only a few exceptions, the findings were similar across different types of moral regulation laws, model specifications, and years.

Taken together, the results have broad implications for understanding political and religious divides in support for moral regulation. Whereas conservatives' support for moral regulation appears to derive in part from a moral orientation that privileges collective needs over individual needs, research suggests that liberals consider individuals' needs to be paramount (Graham and Haidt 2009). Thus, whereas conservatives may perceive a need for moral regulation laws, liberals may perceive an opposing imperative to relax moral regulation laws that have negative consequences for individual citizens. Indeed, moral concerns about individuals' rights and well-being commonly form the basis for criticisms of moral regulation laws (Morone 2003; Schnabel and Sevell 2017; Verbakel and Jaspers 2010), as in the introduction to this study. Given that conservatives' and liberals' perspectives may be rooted in divergent moral concerns, which each side accepts as truth without conscious reflection, political debates regarding moral regulation may be contentious or unproductive (Haidt 2012; Haidt and Graham 2007). However, researchers have suggested that moral perspective taking may present a road to understanding, and potentially bridging, ideological divides (Haidt and Graham 2007).

Understanding the sources of moral regulation law support may also help to explain the adoption of moral regulation laws over time or across cultures. As various scholars have noted, the past several decades have been marked by an increasing emphasis on individual needs and autonomy in place of binding moral concerns (Pinker 2011; Schnabel and Sevell 2017; Silver 2017; Stets and Carter 2012). As binding moral values lose their hold in contemporary Western society, the public may become more tolerant of “immoral” behaviors (Euchner et al. 2013; Mucciaroni 2011; Schnabel and Sevell 2017) and less tolerant of laws restricting harmless yet “immoral” behaviors (Euchner et al. 2013; Pinker 2011). A similar logic suggests that cross-cultural variation in moral foundations may help to explain differing punitive and regulatory regimes across countries or regions (Silver 2017).

The current study also provides insight into “the apparent puzzle of why people whose activism appears to be motivated by religious beliefs make arguments without mention of religion and why people whose opposition to the issue stems from absolute moral beliefs make arguments in a vocabulary of situational ethical risk” (Gandsman 2016: 579), often with a focus on worst case scenarios of individual harm that outstrip empirical evidence (Meier 1999; Weitzer 2009). That is, insofar as judgments about harm are influenced by moral intuitions, they may be “plastic,” serving as justifications for criminalization preferences rather than as sources of those preferences (Sood and Darley 2012). It is also important to note that, although the current study focused on binding morality among conservatives, liberals also experience moral intuitions and may recruit harms to justify their policy preferences. For example, scholars point to perceived harm as the basis for the unlikely union between evangelical Christians and radical feminists in support of laws criminalizing sex work (Brents et al. 2017; Weitzer 2009).

Whereas the overall pattern of results was similar across models, it is important to note where the findings were most consistent and where they diverged. Across all models and time periods, the binding morality measure that consistently predicted support for moral regulation laws was purity. Given the use of sex-related items as a basis for the purity measure, one of which (homosexual sex) has become more accepted by the public over the past several decades (Schnabel and Sevell 2017), it is somewhat surprising that the effect sizes for purity were similar for laws regulating sexual and nonsexual behavior in different decades. However, the findings are consistent with research suggesting that moral concerns about the moral foundation purity drive various social policy attitudes (Koleva et al. 2012), as well as research showing that moral purity concerns are associated with desires to regulate “immorality” across sexual and nonsexual domains (Anderson and Masicampo 2017). By contrast, the effects of moral absolutism and moral collectivism on support for moral regulation laws were more variable. Absolutism was a significant predictor of moral regulation law support from 2006 to 2010 but not in 1988, whereas the opposite pattern was observed for collectivism (though the disaggregated models in Appendix B do show effects for collectivism from 2006 to 2010 and absolutism in 1988).

On the other hand, moral collectivism had the largest and most consistent effects on perceived harm. A possible explanation is that because it may be difficult to articulate dread at the prospect of social dissolution, people who value moral collectivism may be especially likely to reason that immoral behaviors that feel threatening must be harmful. Interestingly, this explanation echoes research on “moral panics” (Cohen 1972), described by Jenkins (1998: 8–9) as “begin[ning] with an event or condition that represents a serious challenge to accepted values” and which is later connected “with other conditions that [activists] believe to be harmful or threatening.” Future research might explore the role of binding morality in the development of moral panics, as well as in policy responses, particularly given that social definitions of good and evil (and resulting panics) are often responses to shifting social boundaries (Morone 2003). Research might also consider whether the effects of collectivism on perceived harm remain powerful in more recent years, given the decline the effect of collectivism on moral regulation law support from 1988 to 2006–2010.

The results for perceived harm also suggest that it may be useful to further consider differences among political and religious conservatives. Whereas binding morality mediated the effects of both forms of conservatism on moral regulation law support to a similar decree, in the models predicting perceived harm, binding morality more strongly mediated religious conservatism than political conservatism, particularly in the 1976–1994 models. The greater importance of binding morality in predicting attitudes among religious conservatives is consistent with prior research examining punitive attitudes (Silver and Silver 2017) and suggests that future research might consider how different forms of conservatism condition the effects of binding morality on perceived harm and support for moral regulation.

This study has several limitations. All binding morality measures were limited to a few items that lack the reliability and validity of existing moral foundations measures (e.g., Graham et al. 2011). Relatedly, the items measuring moral purity included questions about the moral acceptability of sexual behaviors (i.e., homosexual sex and extramarital sex) that may themselves be perceived as policy issues or sources of harm. It was also not possible to assess whether concerns about binding morality were rooted in intuition, as is suggested by the theory, although this is a limitation of all survey-based morality measures (Miles et al. 2019). Future work might address these issues by (a) using validated measures of moral intuitions, and/or (b) employing methods aimed directly at measuring respondents' intuitions, such as implicit association tests or questions posed while respondents are under cognitive load (see Miles et al. 2019). Relatedly, researchers have critiqued moral foundations theory for failing to distinguish between binding morality and conservative ideology, or for failing to appropriately describe the relationships between those constructs (e.g., Hatemi et al. 2019; Kugler et al. 2014). Future research might thus also explore whether the effects hold when other forms of conservative ideology (e.g., right-wing authoritarianism) are included in the models, or whether an alternate theoretical model would be appropriate.

Regarding perceived harm, a limitation of the single-item measure used in this study is that rape is but one potentially harmful consequence of pornography, and negative outcomes such as family conflict and mental illness have also been attributed to pornography use. Additionally, this item had a strong correlation with the measure of porn-specific moral collectivism, suggesting conceptual overlap (which makes sense given that both measures address negative consequences of pornography). Future research might thus replicate the current study using measures of perceived harm with better content and discriminant validity.

A final limitation is that much of the data are old: the most recent of the main analyses use data from 9 to 13 years ago, whereas the older data range from 26 to 43 years old. Given that attitudes about moral regulation have relaxed in the last decade alone (Schnabel and Sevell 2017), and the correlates of moral regulation law support differed across the years measured, it is possible that not all results would generalize to the current decade.

Despite these limitations, this study contributes to the literature by examining binding morality and perceived harm as sources of support for moral regulation laws. Overall, the results indicate that binding morality divides liberals and conservatives on questions about the governmental regulation of morality and suggests that conservatives' support for moral regulation laws may be driven in part by moral concerns about protecting social rules, customs, and cohesion, or by perceptions of “immoral” behaviors as harmful.