Race and Determinations of Discrimination: Vigilance, Cynicism, Skepticism, and Attitudes about Legal Mobilization in Employment Civil Rights

Abstract

What factors affect whether ordinary citizens believe that workplace decisions involving African-American employees rise to the level of discrimination? When do observers believe targets of possible race discrimination should consider mobilizing the law? We use a factorial design vignette study administered to a nationally representative sample of 2,087 ordinary people to address these questions. The “vigilance hypothesis” predicts that minorities will be more likely to perceive discrimination than whites. Our analysis partially confirms this: African Americans perceive anti-Black discrimination at higher rates than do whites and Latinos, while Latinos do not show a significant difference from whites. Where respondents believe discrimination occurred, we analyze what influences whether respondents might recommend legal mobilization. The “cynicism hypothesis” suggests that people of color may be less likely to favor using law. We find, however, that African-American and Latino respondents express more confidence in civil litigation, compared to whites. Further, African Americans express the strongest support for legal mobilization (recommending that a “friend” contact an attorney), while whites and Latinos do not differ in mobilization recommendations.

Title VII of the 1964 Civil Rights Act prohibits race discrimination in employment, protecting workers from discrimination on the basis of their race (among other characteristics) in hiring, firing, promotion, and compensation decisions. Civil litigation by individual plaintiffs is the primary enforcement mechanism for Title VII, but sociolegal scholars long have recognized that a very small proportion of potential cases enter the legal system, much less go to trial (Ewick & Silbey 1998; Felstiner et al. 1980; Galanter 1983; Nielsen et al. 2010).

A growing body of scholars have begun to examine the dynamics of identifying a problem as a legal one (Albiston 2005, 2006; Curran 1977; Engel & Munger 1996; Ewick & Silbey 1998; Sandefur 2008), the institutional and ideological forces at work in these determinations (Albiston 2010; Haltom & McCann 2004; Nielsen & Beim 2004), as well as the supports and obstacles to entering the legal system to pursue remedies (Curran 1977; Sandefur 2008). We analyze the very bottom-most levels of the “pyramid of disputes” to explore the link between perceptions of legal harm—in this case race discrimination—and the determination of whether it might be worth talking to a lawyer about the problem (Felstiner et al. 1980; Miller & Sarat 1981).

Legal cynicism,¹ a concept which refers to a sense of distrust in legal institutions and their ability to deliver fair and effective remedies (Bell 2016; Kirk & Papachristos 2011; Sampson & Bartusch 1998), may be important in this process. Research on legal cynicism principally has been concerned with perceptions of the criminal justice system, with police–citizen relations receiving the largest share of attention (Bell 2016; Carr et al. 2007; Hagan & Albonetti 1982; Hagan et al. 2005; Kirk & Matsuda 2011; Kirk & Papachristos 2011; Weitzer & Tuch 1999). Scholars have addressed the correlates of cynical views toward law and law enforcement and have examined the consequences of these attitudes for willingness to call upon legal authorities to help resolve problems. However, with few exceptions (Brooks & Jeon-Slaughter 2001; Nielsen 2004), the role of legal cynicism in shaping citizens’ willingness to pursue remedies through the civil justice system has received scant attention. Using factorial survey data on perceptions of race discrimination in the workplace, we address this gap by examining the factors that account for disjunctures between perceiving discrimination and recommending that a victim turn to law.

Although race discrimination claims represent a significant share of employment civil rights litigation, accounting for between 33 percent and 38 percent of charges filed with the U.S. Equal Employment Opportunity Commission in any given year (Equal Employment Opportunity Commission (EEOC) 2015), prior research strongly suggests that only a very small percentage of workplace race discrimination triggers a formal legal claim (Nielsen & Nelson 2005). Empirical studies of legal mobilization show that a variety of factors are important to whether ordinary people understand their problems as legal ones (McCann 1994), but we know very little about how attributions of discrimination translate into considerations of utilizing law. Even when individuals understand that their legal rights have been violated and they know who is responsible, they may not choose to pursue a claim for a variety of reasons. They may fear retaliation (Berrey et al. 2012; Ewick & Silbey 1998). They may have become accustomed to being harmed without redress due to their social location (Sarat 1990). They may lack confidence that legal actors will believe their claims or be responsive to them, or they may refuse to identify as a “victim” in the way the law requires (Bumiller 1988; Nielsen 2004; Taub & Schneider 1998). They may not be able to afford an attorney (Myrick et al. 2012).

Moreover, because civil litigation serves as the primary institutional mechanism for enforcing social ideals of racial equality in the post-civil rights era (Freeman 1978; Lucas 2008; Nielsen et al. 2010), laypeople's perceptions of legal mobilization carry implications for the legitimacy of the established system of rights protection. While prior research documents a popular narrative of excessive litigation, frivolous cases, and massive jury awards (Haltom & McCann 2004; Nielsen & Beim 2004), we know little about when people choose not to use the law. Because anti-discrimination law is meant to serve as a mechanism for combatting illegitimate status structures in the workplace, we anticipate that perceptions of discrimination and attitudes toward legal mobilization will vary in important ways by respondents’ status characteristics. Given the centrality of race in scholarship on legal cynicism, we focus primarily on race differences in perceived discrimination and support for legal mobilization.

Racial attitudes research consistently shows notable differences in how African Americans, whites, and Latinos view racial discrimination as an enduring social problem, but less is known about how these divergent views affect the system of employment civil rights. Our quantitative research examines whether support for legal mobilization matches observed differences in how race predicts perceptions of discrimination. Consistent with existing research, we find that African-American respondents perceive more anti-Black discrimination than other racial groups, although Latinos do not show significant differences compared to whites. While research on attitudes about the criminal justice system finds that African Americans and Latinos hold more negative views about criminal law, the African-American and Latino respondents in our study expressed significantly more favorable attitudes toward civil litigation compared to white respondents. We also found that African-American respondents consistently gave stronger support for legal mobilization in response to workplace discrimination, while Latinos did not differ significantly from whites.

Consciousness, Claiming, and the Bottom of the Dispute Pyramid

Sociolegal scholars long have recognized the fluidity and temporality of rights consciousness and the importance of individual experience for determining it (Ewick & Silbey 1998; Merry 2003). These seemingly idiosyncratic attitudes are influenced (or even determined) by hegemonic conceptions of the appropriate role of law in society, notions of power, and social location of parties (Bumiller 1988; Ewick & Silbey 1998; Merry 2003; Nielsen 2000; Silbey 2005). Invisible but pervasive social norms, cultural constructions, and hierarchies that shape legal consciousness systematically and variably impact people in different social locations. This section describes two empirically-tested theories of these differences: the “vigilance hypothesis,” which holds that people of color are more likely than whites to perceive discrimination because they are “on alert” (or “vigilant”) about recognizing it, and the “cynicism hypothesis” which holds that lower-status groups are less trusting of legal institutions and actors and less likely to indicate willingness to encourage a friend to contact a lawyer.

Perceiving Discrimination

Disagreement between Americans of different racial groups in perceptions of race discrimination is well documented (Bobo & Kluegel 1993; Bonilla-Silva 2003; Kluegel & Bobo 2001; Krysan 2000; Tuch & Hughes 1996). For example, Kluegel and Bobo (2001) find that while large majorities of African Americans (between 57 and 70 percent) believe African Americans experience “a lot” of discrimination in the workplace, whites generally estimate that African Americans face “some” discrimination, with substantial minorities seeing anti-Black discrimination as minimal or non-existent. Schuman and colleagues (1997) show that roughly 80 percent of white survey respondents believe that “Blacks have as good a chance as whites to get any job for which they are qualified.” When whites are asked to rank the relative significance of discrimination for African Americans across a variety of social settings, discrimination in employment ranks lowest in terms of perceived seriousness (Schuman et al. 1997). Kluegel and Bobo (2001) conclude that the perception gap “is best characterized as one between a somewhat halting recognition on the part of whites and a near consensual view among blacks that their group is the victim of prevalent job discrimination” (168).

Prior research suggests that Latinos occupy a middle position between African Americans and whites in terms of their views on race-related issues (Bonilla-Silva 2002; Bonilla-Silva et al. 2003). Latinos’ intermediate status in this “tri-racial” system is thought to be a product of objective indicators such as income and education (DeNavas-Walt & Proctor 2015; Pager et al. 2009; Sullivan et al. 2015), subjective identification (Bonilla-Silva 2004), and the historical dynamics of social assignment to the category of “non-white” (Bashi & McDaniel 1997). Despite substantial variation by subgroups, Latinos are on average similar to African Americans on measures of socioeconomic status, including income, poverty rates, and household wealth (DeNavas-Walt & Proctor 2015; Sullivan et al. 2015). Substantial proportions of Latinos identify as white, although racial self-identification varies widely by national origin, with, for example, 39 percent of Guatemalan Americans, 53 percent of Mexican Americans, and 85 percent of Cuban Americans identifying as white (Ennis et al. 2011). Yet the social experience of racialized “otherness” may sensitize Latinos to race discrimination against other minority groups. Survey evidence indicates that Latinos perceive more anti-Black discrimination than whites, but less than African Americans (Forman et al. 2002; Kluegel & Bobo 2001). Kluegel and Bobo (2001) show a clear gradient in perceived anti-Black discrimination among racial groups across the four cities they sampled, such that the distributions for “a lot” of race discrimination perceived by whites (19.4–33.1 percent), Latinos (40.3–48.9 percent), and African Americans (57–70 percent) do not overlap.

While survey evidence provides valuable context on how members of different groups view discrimination as a social problem in the abstract, it is less useful in addressing the factors that shape perceptions in particular situations that may involve discrimination. Workplace discrimination often is difficult for victims to identify, as victim-employees typically do not have access to the kind of comparative data necessary to make a confident determination that discrimination caused an adverse employment outcome. As such, potentially discriminatory situations often involve a high degree of what social psychologists call “attributional ambiguity,” or the difficulty of determining whether negative outcomes are due to discrimination or other factors such as ability and effort (Inman 2001). Examining the process of naming discrimination in situations of attributional ambiguity, social psychologists have demonstrated what is termed the “vigilance hypothesis,” which holds that perceived discrimination varies with subjects’ ascribed statuses, where members of marginalized social groups perceive more discrimination than members of higher status groups (Major & Kaiser 2005; Major et al. 2002). Researchers attribute these differential perceptions to the “vigilance” that members of marginalized groups carry toward discrimination based on past experience with status-based mistreatment (Major & Kaiser 2005). In other words, the vigilance hypothesis predicts that because people of color are more likely to have experienced mistreatment based on racial status, they are more likely to perceive race discrimination in situations that whites see as ambiguous. Research on the prevalence of workplace discrimination supports this view (Pager 2003; Smith 2002). For example, a nationwide study of 1,000 workers finds that 28 percent of African Americans report experiencing unfair treatment based on their race, compared to 6 percent of white workers (Dixon et al. 2002). Further, 22 percent of Latino respondents reported facing race-based mistreatment. Thus, among African Americans and Latinos, frequent contact with manifestations of racial hierarchy may position discrimination as a “more accessible construct” (Hirsh & Lyons 2010), making attributions to discrimination in ambiguous situations more likely (Major et al. 2002).

Legal Cynicism

Once an individual suspects that discrimination may have occurred, he or she must take the important first step of determining whether it is worth beginning to pursue a claim. The process of exercising legal rights is affected by the organizational settings in which they are applied, the nature of the competing claims made by using rights, and the different social locations of the individual rights claims (Merry 1990; Merry 2003; Nielsen 2000; Sarat 1990; Sarat & Kearns 1995; Yngvesson 1985). It is not just a product of the law, but also of social context; of how the problems are defined by court actors such as judges (Merry 1990), court clerks (Yngvesson 1988), friends and neighbors (Albiston 2005; Ewick & Silbey 1998; Nielsen 2000); and of past experiences with law and legal actors (Macaulay 1963; Merry 1990; Nielsen 2000; Sarat & Kearns 1995). Some individuals may not know the law; others may eschew the category of “victim” still others may prefer to maintain business, family, and friend relationships rather than assert legal rights; and all individuals exist in socioeconomic, race, and gender hierarchies that affect their ability and willingness to pursue legal claims.

Scholars of legal cynicism suggest that distrust of law leads harmed individuals to avoid turning to legal institutions to resolve disputes (Anderson 1999). While higher-status individuals may view law as a viable resource to enforce their rights and protect their interests, lower-status individuals may suspect that their grievances will go unheard, that harm may befall them for reporting their complaints, or that a remedy is too remote to be worth pursuing (Nielsen 2004). This perspective sees legal institutions as structured by the same social hierarchies that are revealed in the injuries that make legal recourse necessary. Subjects turn away from law when they view status hierarchy as both the cause of an injustice and an obstacle to redress, making individuals from marginalized groups reluctant to pursue civil litigation to vindicate their rights.

Definitions of legal cynicism vary in the literature. Early scholarship considers legal cynicism as a matter of normative orientation, defining the concept as “ratification of acting in ways that are ‘outside’ of law and social norms” (Sampson & Bartusch 1998: 786). Referring to criminal justice institutions, Kirk and Papachristos (2011) define legal cynicism as “a cultural frame through which people perceive the law as illegitimate, unresponsive, and ill-equipped to ensure public safety” (1190). This formulation includes elements that target both social norms (“law as illegitimate”) and institutional efficacy (“law as unresponsive” and “ill-equipped”). We employ a definition that is closer to the one Kirk and Papachristos (2011) use, understanding legal cynicism as the perception that the law and its representatives are illegitimate, unresponsive, and incapable of delivering fair and effective remedies (Nielsen 2004).

Empirical research has yielded inconsistent results on the relationship between mistrust of law and dimensions of social status. Emerging out of research on police–citizen interactions, scholars have extensively documented the existence of skeptical attitudes about law enforcement in African-American communities (Bell 2016; Carr et al. 2007; Hagan & Albonetti 1982; Hagan et al. 2005; Kirk & Papachristos 2011; Weitzer & Tuch 1999). Although African Americans express higher levels of legal cynicism compared to whites, studies produce diverse results on how these views are located along class gradients. Sampson and Bartusch (1998), for example, find that concentrated neighborhood disadvantage explains higher levels of reluctance to turn to law among African Americans. By contrast, other research shows that among African Americans, income and education are related negatively with views of the criminal justice system (Brooks 2000; Brooks & Jeon-Slaughter 2001; Hagan & Albonetti 1982; Weitzer & Tuch 1999). Further, Kirk and Papachristos's (2011) study indicates that measures of structural disadvantage do not fully account for race differences in legal cynicism, as African Americans in their sample expressed more mistrust of law than other groups, net of concentrated poverty and other structural conditions.

Latinos’ legal attitudes have received far less attention (Weitzer 2014). The existing evidence, however, points to a “racial gradient” (Hagan et al. 2005), where Latinos tend to hold more favorable views of the law than African Americans, but less favorable attitudes than whites (Carr et al. 2007; Weitzer 2014). In Kirk and Matsuda's (2011) study, neighborhoods with larger proportions of Latinos show elevated levels of legal cynicism, but still below those of predominantly African-American neighborhoods.

Research on legal cynicism has given scant attention to lay perceptions of civil justice institutions, providing limited comparative evidence on the relationship between race and mistrust of law (cf. Brooks & Jeon-Slaughter 2001; Curran 1977; Ewick & Silbey 1998; Nielsen 2000, 2004). Studying citizens’ attitudes about using law to address offensive public speech, Nielsen (2004) finds that despite frequent experience with racial harassment, African-American men expressed misgivings about using law to address this behavior. African-American respondents based this distrust of legal remedies on their lived experience, recounting contacts with the law in which they were treated unfairly or had their concerns dismissed. Among these respondents, skepticism about legal regulation was shaped by the belief that law either cannot really help or will actually be used against those it was intended to protect. In their study of views on the U.S. courts, Brooks and Jeon-Slaughter (2001) find that higher-income African Americans tend to have less favorable views than lower-income African Americans. However, their study did not include comparative analysis of racial group attitudes, leaving open the question of how the evidence for mistrust of criminal justice institutions translates to lay views of civil law.

Recent scholarship draws attention to the distinction between cynicism as a set of views about law versus real or intended use of legal institutions to solve problems (Bell 2016; Desmond et al. 2016). That is, some individuals may be highly cynical about law, yet also mobilize law at rates that are similar to or exceed those of individuals who hold more positive attitudes about the legal system. For example, although the research literature shows that African Americans hold more cynical views about criminal law, it also suggests that these views do not translate evenly to avoidance of legal authorities (Avakame et al. 1999; Baumer 2002). Bell's (2016) interview work with disadvantaged African-American mothers shows that despite their cynical views of the police and of the criminal justice system more generally, the women in her study used police notification strategically to help address a tightly defined set of problems. Carr and colleagues’ (2007) study with African-American and Latino youth found that although respondents were negatively disposed toward the police based on their lived experience, they also consistently endorsed tougher and more robust law enforcement as the solution to crime in their neighborhoods. Finally, Bobo and Thompson (2006) found substantial differences between African Americans and whites in their attitudes toward the criminal justice system, but the racial gap narrowed substantially on the question of whether respondents would report a burglary to the police: 90 percent of African Americans and 93 percent of whites indicated they would do so. The gap reappeared as to whether authorities would take the complaint seriously, with 60 percent of whites and 35 percent of African Americans believing their complaint would not be dismissed.

Other empirical research suggests a more direct relationship between attitudes about the law and willingness to use it. Studying lay cooperation with the police, Sunshine and Tyler (2003) show that perceptions of police legitimacy are the most robust predictor of calling the police to report a crime. Slocum and colleagues (2010) find that youth who were more positively disposed toward the police were also more willing to report crime, and African-American and Latino youth were significantly less willing to do so than their white counterparts.

In the area of civil law, research on the litigation process points to racial disparities that may ground skeptical views. In employment civil rights cases, white plaintiffs are more likely than their African-American counterparts to have legal representation (Myrick et al. 2012), and this difference has the important consequence of significantly increasing the likelihood that cases will survive the dismissal phase and summary judgment (Nielsen et al. 2010). Research on plaintiff demographics also indicates that white plaintiffs see higher win rates in workplace discrimination cases (Best et al. 2011). While their link to legal attitudes has not been studied systematically, these material advantages may lead to more favorable views of civil justice among whites, compared to African-Americans and Latinos.

The literature leaves open questions regarding the extent to which mistrust of law may be specific to the criminal justice system, and about the relationship between attitudes toward the law and willingness to use it to address other types of legal disputes. We address these gaps by examining legal attitudes in the area of employment civil rights litigation, providing a comparative assessment of how skepticism shapes views about turning to law to remedy perceived instances of race discrimination. We present survey respondents with vignette scenarios that describe workplace disputes in which an employee (the discrimination “target”) is denied a raise, while a similarly situated employee of a different race is granted a raise. The vigilance literature suggests that African-American and Latino respondents will perceive more discrimination against the African-American employee in these workplace disputes, compared to white respondents. To measure legal cynicism in general, we first assess views of civil legal institutions in the abstract to analyze whether skepticism towards the criminal justice system extends to other legal institutions. If so, African Americans and Latinos may express relatively negative views compared to white respondents. We then examine how perceptions of discrimination and civil litigation attitudes translate into recommendations that a target turn to law to address the perceived injury. We assess the extent to which support for one aspect of legal mobilization—operationalized here as recommending that the target contact a lawyer—differs by the race of the respondent and the target of perceived discrimination. Theories of legal cynicism would lead us to expect that after controlling for perceptions of discrimination, African Americans and Latinos would, on average, be less likely to support legal mobilization, compared to white respondents. We test the following hypotheses based on evidence from research literatures on legal cynicism and perceptions of race discrimination:

H1: Vigilance in Perceptions of Discrimination

1A. African-American respondents will perceive more discrimination against African-American employees, compared to white respondents.

1B. Latino respondents will perceive more discrimination against African-American employees, compared to white respondents.

H2: Cynical Views toward Law

2A. African-American respondents will express less favorable attitudes toward civil legal institutions, compared to white respondents.

2B. Latino respondents will express less favorable attitudes toward civil legal institutions, compared to white respondents.

H3: Mobilization Recommendations

3A. Controlling for perceived discrimination, African-American respondents will give lower support for legal mobilization than white respondents for all targets of discrimination.

3B. Controlling for perceived discrimination, Latino respondents will give lower support for legal mobilization than white respondents for all targets of discrimination.

or:

3C. Controlling for perceived discrimination, African-American respondents will give lower support for legal mobilization than white respondents for African-American targets.

3D. Controlling for perceived discrimination, Latino respondents will give lower support for legal mobilization than white respondents for African-American targets.

Data and Methods

We use a factorial survey featuring descriptions of workplace disputes to examine the relationship between race, perceptions of discrimination, legal cynicism, and support for legal mobilization. Participants came from a nationally representative sample of 2,087 people drawn by Knowledge Networks (KN), an online research and analysis firm that maintains a panel of 48,725 U.S. households for its surveys. For this project, KN recruited panel participants using both random-digit dialing and address-based recruitment methods. KN estimates that its sampling methods provide 97 percent coverage, meaning that 97 percent of the intended population would be captured under these methods. KN draws a subsample from its panel using a client's specifications for a particular study. Sixteen percent of those sampled for the KN panel were successfully recruited, and of those, 64 percent had completed the demographic profile necessary to participate. For this study, KN sampled 4,990 noninstitutionalized U.S. citizens over age 18. Of those, 2,665 individuals were screened, meaning that 53.4 percent of those sampled chose to proceed with the survey. Further, 2,222 individuals consented to participate, and we received 2,087 completed questionnaires, resulting in a 93.9 percent completion rate among those who began the survey, and an overall 41.8 percent completion rate for those recruited to participate.²

Factorial surveys combine features of traditional experiments and survey methods, pairing the strong internal validity of random assignment with the greater external validity provided by surveys (Rossi & Anderson 1982). Under the factorial survey approach, subjects read and respond to vignettes, or short statements describing a scenario of interest. Each vignette includes a set of factors (independent variables), each of which contains several possible randomly assigned levels (variable values). Typically, laboratory experiments are limited in the number of factors they can test, but factorial surveys allow researchers to examine how a relatively large number of elements shape social judgments.

The factorial design allows us to examine perceptions of discrimination in a way traditional survey methods cannot, namely by giving respondents the opportunity to evaluate scenarios, rather than discrimination in the abstract. Workplace disputes involving potential race discrimination contain many factors that may influence whether a target's grievance is perceived as discrimination and whether legal mobilization may be appropriate. Random assignment of worker characteristics and environmental factors allows us to isolate the effect of target race from other variables that may influence respondents’ judgments.

Respondents received a total of eight vignettes describing workplace disputes that concerned race discrimination, sexual harassment, and reasonable accommodation. This article analyzes only responses to the race discrimination vignettes. We employ a between-subjects, mixed-factorial design, meaning that the different dispute factors are randomly assigned to respondents, and each respondent evaluates three race vignettes containing different combinations of factors. Each vignette features a workplace dispute in which a worker (hereafter referred to as the “target”) is denied a raise by a supervisor while another equally qualified worker of a different race applied for and received the raise.³ The supervisor's race is never specified and the target's gender is specified as male for all (race) vignettes. We varied the target's race (“white” or “African American”). We also varied work history: good performance evaluations and short tenure (1 year); poor evaluations and short tenure (1 year); good evaluations and long tenure (10 years); and poor evaluations and long tenure (10 years). We also varied characteristics of the target's workplace environment. The employing organization is described as having: a history of employees making offensive remarks about racial minorities; formal human resource procedures for filing grievances; or both of these characteristics.⁴ The vignettes use different company descriptions to avoid learning and exhaustion effects (Auspurg et al. 2009). The vignette factors and their levels are displayed in Table 1.

Table 1. Vignette Factors and Levels

Each respondent received three race vignettes, and every vignette featured one randomly assigned level from each of three factors, yielding a 2 × 4 × 3 = 24 population of factorial objects. The three vignettes were not selected as a simple random sample. Instead, we used a sampling technique that maximized the likelihood that respondents received all the conditions. Specifically, for the race-of-worker condition, with two treatments (African American and white), all subjects randomly received either two African-American worker vignettes and one white worker vignette, or the reverse. For the workplace-environment condition, with three treatments, each respondent received each treatment once. For the employee-tenure condition, with four treatments, each respondent randomly received three of the treatments once and one twice. Assignment for separate conditions was done completely independently, so each combination of conditions occurs with roughly equal frequency in our data and all 24 possible vignettes are presented with roughly equal frequency.

After reading each vignette, respondents answered a series of closed-ended questions about whether they thought the target was discriminated against and what, if anything, they think the target should do about it. The survey also included manipulation check questions to ensure that participants read and understood each scenario. Our analysis eliminates respondents who answered those questions incorrectly.

We estimate OLS regressions of perceptions of discrimination and legal mobilization on vignette factors and respondent characteristics. Due to clustering of vignettes by respondent, it was necessary to use clustered robust standard errors for our regression estimates.⁵ This approach is consistent with prior uses of factorial surveys in sociolegal research (Hagan et al. 2008; Wallander 2012).

Dependent Variables

Perception of discrimination is measured by responses to the question, “In your opinion, is this discrimination?” The response categories are: “Very unlikely” (1), “Unlikely” (2), “Less likely than not” (3), “Equally likely as not” (4), “More likely than not” (5), “Likely” (6), “Very likely” (7). We use a scaled measure of legal confidence that combines a set of eight questions that test civil litigation attitudes (Hans & Lofquist 1996), where responses range from Strongly Agree (1) to Strongly Disagree (5). The statements used to construct the scale are: Most people who sue have legitimate grievances (reverse coded); There are far too many frivolous lawsuits today; People are too quick to sue rather than trying to solve disputes in some other way; The courts have made it easier to sue someone in recent years; Civil lawsuits have made this a more fair society (reverse-coded); The number of lawsuits shows that our society is breaking down; Juries do a good job determining the outcomes of lawsuits and assessing damages (reverse-coded); The money awards that juries are awarding in civil cases are too large.⁶ Support for legal mobilization is measured by responses to the question: “[The target] is your friend and asks if he should see a lawyer. Would you encourage him?” The response categories are: “Strongly Discourage” (1), “Discourage” (2), “Mildly discourage” (3), “Neutral” (4), “Mildly encourage” (5), “Encourage” (6), and “Strongly Encourage” (7).⁷

Independent Variables

The OLS specifications use vignette factors (target race, work history, and environment) as independent variables predicting perception of discrimination and the level of support for legal mobilization. Respondent race is measured by self-identification into one out of five categories: white, non-Hispanic; African American, non-Hispanic; Hispanic; other, non-Hispanic; and two or more races, non-Hispanic. We only present results for whites, African Americans, and Latinos, although the regression analyses include all five groups. The other respondent characteristics measured are gender, age, education level, family income, and self-reported political ideology.⁸ We also use perception of discrimination and legal confidence as independent variables in the final models of support for legal mobilization. Table 2 displays the categories for the control variables.

Table 2. Descriptive Statistics

Results

Table 2 presents (weighted) descriptive statistics and shows mean values of discrimination perception, legal confidence, and support for mobilization.⁹ On average, subjects interpreted the vignette scenarios as race discrimination. The sample mean across all scenarios and all subjects was 5.19 for perceived discrimination, notably higher than the middle category (4) which would indicate that the vignettes were “equally likely as not” to be discrimination. This perception did not translate neatly into opinions on legal mobilization, which registered a lower mean (4.92) than the perception variable. However, like perceived discrimination, the sample mean for legal mobilization is well above the value of the middle category (4), meaning that subjects generally supported contacting an attorney as a response to the workplace dispute. The lower average for legal mobilization recommendations compared to perceived discrimination is consistent with prior research that demonstrates a pyramid of disputes, where perceived injuries outnumber formal claims of discrimination (Nielsen & Nelson 2005). While we do not expect the legal mobilization average to perfectly match perceived discrimination, we are interested in variation in the perception-mobilization differential by respondent and dispute characteristics.

On average, white and Latino respondents recorded 5.18 on perceiving discrimination, whereas African Americans scored an average of 5.38. While each group shows a decline from their mean perceived discrimination to their mean legal mobilization attitude, African-American respondents view legal mobilization most favorably (5.29). Whites, by contrast, hold the lowest opinion of legal mobilization (4.84), and Latinos score in between white and African-American respondents (4.98). African Americans and Latinos both express higher legal confidence than whites.

Women on average perceive more discrimination than do men, although this gap narrows somewhat on the question of legal mobilization. In general, groups with an education level at “some college” and below perceive the most discrimination (5.17 to 5.40), while means decrease at higher levels of education, reaching their lowest score with those who hold a professional degree or doctorate. The pattern roughly translates to mobilization attitudes, with highly educated groups showing a distinctly negative view of legal mobilization. Legal confidence is greatest at the lowest and highest levels of education. A similar pattern appears for income.

Table 2 also reveals differences along ideological lines. Political moderates (5.28) see more discrimination than liberals (5.19), and conservatives see the lowest level of discrimination (5.11). These averages pool vignettes featuring African-American and white targets, likely suppressing the liberal average. Liberals tend to favor legal mobilization (5.03), while conservatives are least inclined to suggest that targets consult with an attorney (4.78). Liberals also express higher legal confidence than moderates, while conservatives score lower.

We devote the remainder of our analyses to examining the nature of the relationship between perceiving discrimination and recommending mobilization, and how it is shaped by interactions between the race of the respondent and the target. We first test for race differences in the determinants of legal confidence, the set of measures that target respondents’ views on the legal system in the abstract. We then examine whether some groups record a larger drop-off between perceiving discrimination and recommending that the victim turn to law, relative to other groups. Finally, we present two sets of regression estimations that test the influence of target and respondent race on perceiving discrimination and support for mobilizing law to address it. In sum, our analyses partially validate the vigilance perspective and contradict the cynicism and mobilization hypotheses, at least in the case of employment discrimination.

In Table 3, we examine whether legal confidence differs by respondent race, regressing standardized legal confidence on respondent race, and other sociodemographic characteristics. Although the legal cynicism hypothesis predicted that African Americans and Latinos would express less favorable views of civil litigation, the results contradict this expectation. African Americans hold considerably higher confidence in civil litigation than whites, while Latinos also show a positive and significant difference, albeit one with a smaller magnitude. Political liberalism also is associated with higher legal confidence, while conservatism and higher incomes are associated with lower confidence.

Table 3. Determinants of Legal Confidence

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

Notes: Reference categories: Race: White; Political Ideology: Moderate.

Clustered robust standard errors in parentheses.

The finding of higher legal confidence held by African Americans and Latinos compared to whites indicates that these groups on average may hold views of the legal system that are more amenable to mobilizing law in response to workplace disputes. Although the race differences for abstract measures of legal confidence are suggestive, we are interested in how these ideas are applied when respondents consider situations of potential discrimination. Does higher legal confidence among African Americans and Latinos translate into recommendations to use law? Table 4 presents adjusted Wald tests that examine differences in mean perceptions of discrimination and legal mobilization attitudes by target and respondent race. Taken together, the vigilance and legal cynicism hypotheses predict that African-American and Latino respondents should perceive high levels of discrimination but express reluctance about the possibility of victims using law to address these disputes. Column 1 displays differences in discrimination perceptions between respondent racial groups for each target race. Thus, column 1 shows that in comparison to non-whites (5.46), whites perceive significantly less discrimination against African-American targets (5.22). African-American respondents, by contrast, perceive more anti-Black discrimination than do other racial groups (5.78 versus 5.23). For both the pooled and race-specific perceptions of discrimination, Latino respondents do not differ significantly from the other racial groups in the sample.

Table 4. Perception of Discrimination and Support for Legal Mobilization by Target and Respondent Race

^*p < 0.05, ^**p < 0.01, ^***p < 0.001 (Two-tailed test).

Notes: Standard deviations in parentheses.

¹ Discrimination and mobilization are each collapsed into three-category variables, where discrimination is unlikely, equal, or likely; and mobilization is discouraged, neutral, or encouraged.

On legal mobilization attitudes in column 2, whites show a notable decline in their support for legal mobilization relative to their group mean on the perceived discrimination scale, and they register a low and significant difference from non-white groups. When considering the situation of an African-American target, white respondents are also less inclined than people of color to recommend legal mobilization. The same holds true for white targets: again, white respondents are less willing than non-white respondents to recommend legal mobilization for a white target.

The mobilization hypotheses suggest that although African Americans and Latinos will see higher levels of discrimination (compared to whites) against African Americans in the workplace, these perceptions will not translate into support for legal mobilization. While African-American respondents do show lower support for legal mobilization relative to perceptions of discrimination, this decline is not unique. Against the mobilization hypothesis, column 2 shows that African-American subjects strongly recommend that African-American targets turn to law in response to workplace disputes. Moreover, compared to other racial groups, African-American subjects do not show a significantly different opinion of legal mobilization for white targets. Latinos do not differ significantly from the other groups in their recommendations to mobilize law.

Table 4 also displays average delta scores, which represent a group's legal mobilization recommendation average subtracted from its perceived discrimination average. Larger delta scores represent greater divergence between a group's average level of perceived discrimination and its average legal mobilization recommendation. Significant differences in delta scores appear between white (0.33) and non-white (0.11) respondents, meaning that whites show a larger overall decrease from perceived discrimination to mobilization in comparison to non-whites. Conditioning for target race, whites differ from people of color only for white targets (0.36 versus 0.01).

African Americans (0.09) register a lower average delta than other groups (0.30), meaning that African-American respondents record less of a decrease between perceiving discrimination and recommending legal mobilization. Adjusting for target race, African Americans (0.01) also show a smaller decrease for white targets in comparison to other racial groups (0.31). If evidence for the mobilization hypothesis exists, we should find it in the African-American target row, where a distinctly large decrease from perception to mobilization would indicate reluctance to pursue legal redress despite heightened consciousness of discrimination. However, African Americans do not significantly diverge in their average discrimination-mobilization delta for African-American targets. That is, African-American respondents do not express greater reluctance about legal mobilization than do other groups. While both the pooled and race-specific delta scores for Latinos are below the sample average, these do not differ significantly from the other group scores.

The delta scores in column 4 are calculated based on re-categorized discrimination and mobilization variables, where each measure is collapsed into three categories. The purpose is to check whether the significant delta differences shown in column 3 merely reflect degrees of magnitude, or whether they show a substantively meaningful change from perception to mobilization. For example, a respondent who rates the likelihood of discrimination as 6 and mobilization at 5 has simply registered a change in magnitude, but a respondent who moves from perceiving discrimination to discouraging the target from seeing a lawyer has made a shift that is more consequential. The significant differences in delta scores that we see in column 3 are also present with the recoded measures. This suggests that there is meaningful movement from perception to mobilization in the negative direction, although the patterns of group-based differences that we observe in column 3 are not disturbed.

In Table 5 we present OLS regressions to examine the extent to which the differences in discrimination perception shown in Table 4 are reducible to other observed respondent characteristics. The first model regress perceived discrimination on respondent sociodemographic and attitudinal variables, as well as vignette factors.¹⁰

Table 5. Perceptions of Discrimination

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

Notes: Reference categories: Race: White; Political Ideology: Moderate.

Clustered robust standard errors in parentheses.

We find a significant positive effect of the African-American target condition on respondents’ perception of discrimination. We also find a large, positive effect for favorable performance evaluations, indicating that respondents are more inclined to see discrimination when model workers are treated unfairly. The results from the first estimation are notable for the absence of significant respondent race differences.

We add interactions between respondent characteristics and vignette conditions in the second model. The African-American respondent by African-American target interaction tests the vigilance hypothesis, which holds that in comparison to white respondents, African Americans will perceive more discrimination against African-American targets. The significant positive interaction confirms this expectation. We do not find, however, that Latinos perceive significantly more discrimination against African-American targets. We also estimated interactions between other respondent sociodemographic variables and target race, but, with the exception of a positive interaction for self-reported political liberalism and African-American targets, none were statistically significant. As such, we do not present the interactions here.

Table 6 presents our final tests of the mobilization hypotheses, displaying OLS regressions of legal mobilization attitudes on vignette factors and respondent characteristics. Because a subject's judgment about the appropriateness of legal mobilization will likely be informed by whether or not the subject sees the workplace dispute as constituting discrimination, the OLS specifications in Table 6 include discrimination perception as an independent variable. We also use standardized legal confidence as an explanatory variable.

Table 6. Support for Legal Mobilization

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

Notes: Reference categories: Race: White; Political Ideology: Moderate.

Clustered robust standard errors in parentheses.

In the first specification, we omit perceived discrimination and find significant positive effects for the African-American target condition, positive performance evaluations, and longer work history. Net of these factors, African-American respondents also express considerably stronger support for legal mobilization, compared to white respondents. As expected, the second model shows that perceptions of discrimination play an important role in legal mobilization recommendations, such that adjusting for perceived discrimination increases the proportion of variance explained from 5 percent to 55 percent. Perceptions of discrimination also moderate the effect of the African American target condition, which drops from statistical significance in the second model. A sizeable share of the difference between African-American and white respondents also is due to perceptions of discrimination. As the results in Table 5 confirmed, African Americans see more discrimination against African-American workers compared to white respondents, but African-American and white respondents do not differ significantly on their perceptions of white-targeted discrimination. Thus, whites are less enthusiastic about legal mobilization in part because they see less anti-Black discrimination than African Americans, but this does not explain all of the difference between African-American and white respondents.

The coefficient of legal confidence is also in the expected direction: respondents with higher levels of legal confidence are more willing to recommend legal mobilization for the target worker. Both of the work history factors play an independent role in shaping legal mobilization recommendations in the second model, as respondents appear to find mobilization more appropriate for employees with long tenure and favorable performance evaluations.

Two scenarios could validate the mobilization hypotheses. First, after controlling for perceptions of discrimination, legal confidence, dispute factors and sociodemographic characteristics, we may find that African-American and Latino respondents give lower support for legal mobilization than white respondents (Hypotheses 3A and 3B). The results in the second specification give evidence against this expectation, as African-American respondents provide stronger legal mobilization recommendations compared to white respondents, net of controls. Latino respondents, however, do not differ significantly from whites in their support for legal mobilization. Second, we could find that, compared to white respondents, African Americans and Latinos give lower mobilization recommendations for African-American targets (Hypotheses 3C and 3D). We test this possibility by including respondent race by target race interactions in the third mobilization model. We do not find evidence of a significant African-American respondent by African-American target interaction, indicating that African-American support for legal mobilization is not conditional on the race of the target.¹¹ Against Hypothesis 3D, we also fail to find a significant Latino by African-American target interaction.

Discussion

Our analytic strategy addressed questions of race differences in perceptions of discrimination, confidence in the civil litigation system, and support for legal mobilization in response to potential race discrimination. Drawing from recent literature on the extent and nature of the relationship between attitudes about law and willingness to mobilize it, we used separate analyses for legal cynicism and support for legal mobilization, testing the former with a scaled attitudinal measure, and the latter with respondents’ recommendations that the discrimination target contact a lawyer.

The vigilance hypothesis (H1) anticipated that African-American and Latino respondents would perceive more discrimination against African-American workers in workplace dispute scenarios, compared to white respondents. Consistent with prior work on perceiving discrimination, our results partially confirmed the vigilance hypothesis, as we demonstrated a strong positive interaction between African-American respondents and African-American targets, net of other dispute factors and sociodemographic characteristics. Notably, we were not able to identify positive interactions for Latinos and African-American targets. That the vigilance effect of respondent race was limited to African Americans considering disputes involving African-American targets raises the possibility that vigilance is specific to the shared status characteristics of respondent and target. Because our design includes two values for race, white and African American, it does not allow for a full test of the specificity of vigilance through race-by-race interactions, where, for example, Latino respondents may perceive more discrimination against Latino targets, as compared to other groups.

While we were able to demonstrate that perceptions do vary by respondent and target race, the inability of key dispute factors and our set of respondent sociodemographic variables to account for a larger share of variation represent an opportunity for further research into the determinants of perceived discrimination. The importance of studying these determinants is underscored by the fact that perceptions of discrimination themselves represent the single largest predictor of mobilization recommendations in our models, accounting for nearly 50 percent of variation on this outcome, while the net effect of all other independent variables amounts to an r-squared statistic only slightly higher than that of the full perceptions model.

The legal cynicism hypotheses (2A and 2B) anticipated that African American and Latino respondents would give less favorable assessments of the civil litigation system, compared to white respondents. Against this expectation, we find that African-American respondents expressed the most favorable views of civil litigation, and Latinos also expressed views that were significantly more favorable than those of white respondents.

Having shown that African-American respondents on average both perceive more discrimination against African-American workers than other groups and express more positive views on civil litigation in the abstract, we tested whether these associations would extend to legal mobilization recommendations for targets of race discrimination. The first two mobilization hypotheses held that, controlling for perceived discrimination, African-American and Latino respondents, respectively, would give lower legal mobilization recommendations for all targets, in comparison to white respondents. The second set of mobilization hypotheses predicted that, controlling for perceived discrimination, African Americans and Latinos would give lower mobilization recommendations for African-American targets, in comparison to white respondents. Although all respondent racial groups recorded lower scores for mobilization than for perceived discrimination, African Americans showed the most favorable attitudes towards legal mobilization in terms of raw average alone. By stratifying our sample by respondent and target race, we showed that African Americans on average provided stronger legal mobilization recommendations for African-American targets, compared to other respondents. We also constructed delta scores to test for significant differences in the average change between groups’ perceived discrimination and legal mobilization scores. The mobilization hypothesis held that African Americans would show a larger gap between perceived discrimination and mobilization than other groups. We demonstrated, however, that African Americans recorded less of a drop-off between perceived discrimination and mobilization for white targets, compared to all other respondent groups. Further, African Americans showed no significant delta difference from other groups for disputes involving African-American targets.

Our full test of the legal mobilization hypotheses regressed mobilization recommendations on dispute factors and respondent characteristics. These analyses provided evidence against the mobilization hypothesis as it applies to African Americans and Latinos. While Latinos showed no significant difference from whites, African-American respondents expressed stronger legal mobilization recommendations than other groups, net of perceived discrimination and sociodemographic factors. Moreover, legal mobilization recommendations by African Americans were not specific to African-American discrimination targets, as a significant positive coefficient remained after adjusting for both the target's race and respondent-by-target race interactions.

Our findings extend the research literature on legal cynicism, which has focused largely on race differences in attitudes toward criminal justice institutions. While prior research has demonstrated that African Americans, and to a lesser degree Latinos, hold negative views of the criminal justice system compared to whites, our findings suggest that these orientations do not generalize to views of other legal institutions—including the civil litigation system. This contrast, we suggest, is due the different roles that the criminal justice and civil litigation systems have played historically in American racial structures. It may be that our results are specific to the area of anti-discrimination law because of unique historical associations that link equal employment opportunity to the broader project of civil rights activism. Title VII of the 1964 Civil Rights Act, one of the centerpiece reforms of the civil rights era, was designed specifically to assist targets of race discrimination in employment. By contrast, although criminal law is purportedly a resource open to everyone, African Americans and Latinos believe (with ample evidence) that the criminal law represents a critical support for established racial structures (Alexander 2010; Bobo & Thompson 2006; Garland 2001; Omi & Winant 1994). In this study the absence of evidence for cynicism may be based on the view that employment civil rights litigation is a distinctive and important institutional protection against the effects of racial hierarchy, whereas criminal law historically has served as a key mechanism in racial hierarchy's enforcement. Although scholars have documented the shortcomings of the civil litigation model's potential to effect social change (Berrey et al. 2017; Crenshaw 1988; Freeman 1978; Nielsen et al. 2010), lawsuits remain one of the few institutionalized resources for contesting illegitimate hierarchies in a variety of settings.

While we find race differences that contradict the legal cynicism hypothesis, our findings are ultimately consistent with that scholarship when we theorize skepticism as a function of social location. To the extent the civil litigation offers one possible institutional resource to address established social hierarchies of race, views on civil litigation may vary inversely with social status. The measure of legal confidence that we use in this study targets several key features relevant to collective evaluations of civil litigation, ranging from perceptions of whether plaintiffs’ grievances largely are legitimate or frivolous, to whether lawsuits contribute to making American society fairer. Our results show that while Latinos and especially African Americans tend to hold much more positive views, whites and higher-income respondents give the least favorable answers to these questions about legitimate grievances and social fairness. That is, we find that respondents who are structurally positioned to navigate the litigation process more easily and achieve more favorable outcomes also tend to express the lowest confidence in civil legal institutions (see e.g., Berrey et al. 2017; Ewick & Silbey 1998; Myrick et al. 2012). Although social scientists have shown exhaustively how dimensions of social status shape outcomes in both the criminal and civil justice systems (see e.g., Cole 1999; Galanter 1974), we suggest that the status valences particular to these institutions contribute to contrasting perceptions by lay evaluators. Thus, a more consistent status picture emerges if we consider the results through the lens of civil litigation's redistributive potential, where respondents from relatively high status groups may view lawsuits as litigiousness run amok—precisely because these claims often pit plaintiffs from marginalized social groups against higher-status defendants. By contrast, for marginalized racial groups, this status framework leads to more positive views about civil litigation. The results add texture to the concept of legal cynicism, suggesting that the mistrust of law that scholars identify in the context of the criminal justice system does not generalize to other legal institutions, where law's redistributive potential has a more powerful symbolic presence.

A more expansive set of racial categories would improve this research by capturing something more akin to the social experience of race (Gomez 2012). Previous research has shown the importance of intersectionality for understanding race dynamics in the workplace and in the courts using qualitative and narrative methods (Berrey et al. 2012, 2017; Bumiller 1988). That research, while important, often is met with questions about perceptions of discrimination and whether the subjects selected for interviews (even if randomly drawn) systematically differ from the broad mass of ordinary citizens in how they perceived discrimination. This argues for an experiment designed to recognize the importance of isolating race effects. As such, we limited the design to a comparison of scenarios that propose possible discrimination using an African-American/white dichotomy. We urge the reader to use this study as only one piece of information about the complexity of the construction of race, discrimination, and mobilization in the workplace and in courts.

Conclusion

Our findings extend theories of legal cynicism and test their applicability in the civil justice context. Law is ambiguous but ordinary citizens and legal professionals alike must put it into practice every day. Only a strict legal formalist would argue that the same facts applied to the same law will result in the same decision regardless of the judge or jury making that determination. And yet, consistency is one of the most important principles of the rule of law. Intricate legal opinions, detailed jury instructions, and carefully worded statutes all are written to reduce ambiguity and uncertainty. Despite that, ordinary citizens decidedly disagree about what constitutes discrimination and whether or not it may be worth turning to the law.

Unlike jurors, the respondents in this study were given no instructions about what constitutes discrimination itself and we told them nothing about what legal redress might be available to the hypothetical target they were counseling about seeing a lawyer. However, the fundamental assessment of whether or not discrimination occurred is what drives all future plaintiffs to lawyers’ offices, the EEOC, and the courts, making research about initial gut feelings and expectations of law an important area of scholarship. Partially supporting the vigilance hypothesis, this article shows that the race of the person making a determination about whether or not race discrimination occurred has an influence on that determination. Consistent with prior research, we find that, compared to whites, African Americans perceive more anti-Black discrimination. However, we do not find that Latinos’ perceptions of anti-Black discrimination differ significantly from those of white evaluators.

Our results provide evidence against the legal cynicism hypothesis as it pertains to African Americans. Further, these results reveal that after a determination of discrimination has been made, African-American respondents provide stronger recommendations for legal mobilization for both African-American and white targets, compared to their white counterparts. While our evidence contradicts the legal cynicism hypothesis as it applies to Latinos, we also find that recommendations to mobilize law do not differ significantly between Latino and white respondents.

As a policy matter, these findings have implications for understanding one of the largest portions of the federal civil docket. With a significantly lower than average win-rate for plaintiffs, understanding what people think constitutes discrimination could lead to better education about the legal definition of discrimination, which now is largely ruled by consultants in the business arena who systematically inflate the risk of discrimination to employers. Those mandatory workplace trainings may or may not be effective in countering workplace discrimination, but they may have another effect in increasing ordinary people's expectations about the ability of law to guarantee a discrimination-free workplace. If those ordinary citizens then turn to law only to lose when they believed they had been targeted by discrimination, the legitimacy of law itself is eroded. Finally, antidiscrimination laws are litigious policies (Burke 2002). To achieve our shared social goal of workplaces free of discrimination, we rely on ordinary people to bring lawsuits. How laypeople operationalize the concept of discrimination and assess the wisdom of turning to law to remedy it is central to how antidiscrimination law operates in the workplace.