During New York City’s 2012 mayoral election, the NYPD’s Stop, Question, and Frisk policy- commonly referred to as Stop-&-Frisk- garnered considerable media attention and scrutiny for violating NYC residents’ civil liberties (Baker 2012). Pursuant to Criminal Procedure Law section 140.50- and legal since the early 1970’s- Stop-&-Frisk allowed NYPD officers to legally stop and detain New Yorker’s under the suspicion that they either committed, were in the act of committing or were about to commit crimes.

Although Stop-&-Frisk was ultimately deemed unconstitutional (Floyd et al. 2013), and has since diminished considerably (Goldstein 2014), its effects were ubiquitous: from 2004 through 2013, the NYPD reported a minimum of more than 300,000 Stop-&-Frisks annually with a high of 685,724 in 2011 (New York Civil Liberties Union 2014). Weekly, moreover, Stop-&-Frisk policing reached its peak in January of 2012 when more than 16,000 New Yorkers were being stopped per week (Bostock and Fessenden 2014).

Many New Yorkers were outraged by Stop-&-Frisk’s circumvention of civil liberties, and as it that were not enough, the application of Stop-&-Frisk was rife with racial bias. Reports consistently indicated that Black and Latino people comprised 85-87 % of those stopped for questioning,Footnote 1 with Black people comprising over 50 % of those stopped and Latinos comprising another 30 %; yet the same report indicated that nearly nine out of ten Stop-&-Frisked New Yorkers were found to be innocent (NYCLU 2014). Innocent people of color, then, accounted for an overwhelming majority of those targeted by Stop-&-Frisk policing.

Despite these figures, New Yorkers have consistently remained divided on their views toward Stop-&-Frisk policing; a poll of more than 1000 New York City adults revealed that 45 % of respondents endorsed Stop-&-Frisk as excessive, while 48 % of respondents reported that the policy was acceptable (7 % had no opinion; New York 2012). Support for Stop-&-Frisk was highest among White respondents, with 55 % of them endorsing the practice as acceptable and 39 % endorsing the practice as excessive (6 % had no opinion). Even among Black and Latino poll respondents- two communities that were particularly affected by Stop-&-Frisk policing- opinions remained divided. Among Black respondents, 35 % felt Stop-&-Frisk to be acceptable, while 56 % felt it to be excessive; among Hispanic respondents, a plurality felt Stop-&-Frisk to be an acceptable practice (48 % acceptable to 44 % excessive; New York 2012). Thus, it appears that support for Stop-&-Frisk tends to decrease as communities experience an increase in the likelihood of being stopped, questioned, and frisked.

Because of the tremendous negative impact Stop-&-Frisk had on Black and Latino communities, New York City, and the millions of people who have been stopped, questioned, and frisked, understanding the psychological variables that predict support or opposition for such Stop-&-Frisk policing warrants examination. This study therefore investigated how two key psychological variables that have consistently predicted prejudice contribute to one’s support or opposition to Stop-&-Frisk: Right Wing Authoritarianism (RWA) and Social Dominance Orientation (SDO).

Linking Ideologies and Prejudice

That RWA and SDO are among the most powerful predictors of prejudice has already been well-established; not only do they both correlate strongly with prejudice, but they do so independently (e.g., Altemeyer 1998; Cantal et al. 2015; McFarland 1998), as they comprise two distinct ideological attitudes- driven by different motivational needs- that give rise to sociopolitical behavior (e.g., politics) and prejudice (Duckitt and Sibley 2010).

Right-wing authoritarianism (Altemeyer 1988) is an ideological variable that assesses the degree to which people submit to established leaders and societal norms. Motivated to preserve order in a world that they perceive as dangerous and at the verge of falling apart, people with RWA leanings tend to defer to those who they view as established authorities, act aggressively toward people for whom authority figures have shown contempt, and conform to the traditions of the current state of affairs (Altemeyer 2004). Those high in RWA have repeatedly been found to be prejudiced against numerous communities, including Black Americans, women, as well as lesbian, gay, and bisexual people (Altemeyer 1998; Nicol and Rounding 2013; Rios 2013; Whitley and Lee 2000). In large adult and student samples, for example, right-wing authoritarianism positively predicted anti-Black racism- as measured by the Modern Racism Scale (McConahay 1986)- and did so more strongly than several other correlates of prejudice (Altemeyer 1998; McFarland and Adelson 1996).

Whereas people high in RWA fear the collapse of traditions and the status quo, people high in SDO already view the current state of affairs as a “dog-eat-dog world,” reject equality as a virtue worth striving for, and need no other justification for denigrating minorities beyond that they perceive them as weak (Altemeyer 1998). Accordingly, social dominance orientation (Pratto et al. 1994)- another ideological variable that reliably predicts prejudicial attitudes- refers to “the extent to which one desires that one’s in-group dominate and be superior to out-groups (p. 742).” Those high in SDO may use hierarchy-legitimizing myths, such as negative stereotypes (e.g. “Men are smarter than women”), as a means of justifying prejudicial beliefs and discriminatory practices (e.g. “Women should not be allowed to be president”). Research suggests, moreover, that SDO may be the only variable to predict prejudice as well or better than RWA (McFarland and Adelson 1996); in particular, SDO performed slightly better than RWA at predicting anti-Black attitudes (Altemeyer 1998; McFarland and Adelson 1996; Whitley 1999).

Indeed, even when economic-political conservatism is controlled for, SDO negatively relates to one’s support for gay rights, women’s rights, social welfare programs, and ameliorative racial policies (Pratto et al. 1994). Additionally, high levels of SDO relate to negative affect, stereotyping, and negative attitudes toward the equality enhancement of Black Americans (Whitley 1999). Similarly, items from a shortened SDO-scale (e.g., ‘It’s probably a good thing that certain groups are at the top and other groups are at the bottom’) were related to negative feelings about Latinos (Kteily et al. 2011). Together, RWA and SDO predict nearly 50 % of the variability in prejudice (Altemeyer 1998), and have been shown to influence a wide range of social, political, and intergroup phenomena. RWA and SDO work in concert to predict prejudice so powerfully because, as we will discuss next, they may form the motivational and ideological bases of prejudice and discriminatory behavior.

Causal Models of Prejudice

In exploring the role of RWA and SDO in predicting prejudice, some theorists propose that RWA and SDO are not simply associated with prejudice, but that they cause it. The dual process motivational model of ideological attitudes (Duckitt 2001), for example, implies that prejudice and intergroup attitudes result from stable, dispositional ideologies (i.e., RWA and SDO), which are also anteceded by both personality and worldviews in a causal model. A similar line of work situated RWA and SDO in a theoretical model where they also caused generalized prejudice and were preceded by dimensions of the Big Five personality factors (Ekehammar et al. 2004). Collectively, these models therefore suggest that RWA and SDO stem in part from people’s personality and/or social worldviews and that, in turn, they engender prejudicial behavior. Although the link between ideologies and prejudice has been examined extensively, research on the role of RWA and SDO on social outcomes like politics has been relatively scarce (Duckitt and Sibley 2010), and the dearth of research linking ideologies like RWA and SDO to both prejudice and social policies like Stop-&-Frisk further situates the present approach as a worthwhile endeavor.

In summary, recent research suggests that attitudes toward the NYPD’s Stop-&-Frisk program are considerably variable, and we propose that people’s attitudes toward Stop-&-Frisk may be predicted by important social-ideological variables that work in concert with and are related to (but distinct from) prejudice. Moreover, we argue in particular that RWA and SDO predict prejudice, which in turn predicts people’s attitudes toward the NYPD’s Stop-&-Frisk policy.

Overview of the Present Research

The aim of the present research is to examine the role of authoritarianism, preferences for inequality, and prejudice as psychological variables that predict support for or opposition to the NYPD’s Stop-&-Frisk policy. Specifically, this research aims to elucidate the role of RWA, SDO, and prejudice in predicting support for the NYPD’s Stop-&-Frisk policy. Consistent with prior casual models of prejudice, moreover, we hypothesize that both RWA and SDO will indirectly predict support for Stop-&-Frisk through prejudice; that is to say, prejudice will mediate the relationship between (1) RWA and support for Stop-&-Frisk, and (2) SDO and support for Stop-&-Frisk.

Method

Sample and Procedure

One hundred forty-eight undergraduate students at a multicultural university in New York City completed questionnaires for partial fulfillment of course requirements. Participants’ ages ranged from 17 to 54 years (M = 20.84, SD = 5.47), and 102 of participants were female. The sample, moreover, was ethnically diverse, with 51 participants reporting their ethnicity as African American or Black, 29 reporting their ethnicity as Asian, 20 reporting their ethnicity as Caucasian or White, 17 reporting their ethnicity as Latino/a or Hispanic, eight reporting their ethnicity as Middle Eastern, 17 reporting their ethnicity as multi-racial, and six reporting “Other.”

Predictor Variables

Right-Wing Authoritarianism

RWA (Altemeyer 1988) assesses the degree to which people support the established authorities in their society (e.g., political and/or religious leaders). The 20-item RWA scale includes items such as, “It is always better to trust the judgment of the proper authorities in government and religion than to listen to the noisy rabble-rousers in our society who are trying to create doubt in people’s minds,” and “Homosexuals and feminists should be praised for being brave enough to defy traditional family values (reverse-scored)” on 9-point bipolar scales with the anchors of 1 (very strongly disagree) and 9 (very strongly agree). Scores on these items were summed to create a reliable index of RWA (α = 0.87).

Social Dominance Orientation

The 16-item SDO (Pratto et al. 1994) measure used in the present study includes items such as “Some groups of people are just more worthy than others,” “Sometimes other groups must be kept in their place,” and “Group equality should be our ideal (reverse-scored).” Participants responded on 7-point scales with the anchors of 1 (strongly disagree) and 7 (strongly agree), and scores on these items were averaged to create a reliable index of social dominance (α = 0.89).

Prejudice

We measured prejudice via the quick discrimination index (QDI; Ponterotto et al. 1995)— a 30-item measure of participants’ overall sensitivity, awareness, and receptivity to cultural diversity and gender, and it includes items like, “I really think affirmative action programs on college campuses constitute reverse discrimination” (reverse scored), “Overall, I think racial minorities in America complain too much about racial discrimination” (reverse scored), “My friendship network is very racially mixed,” and “I think White people’s racism toward racial minority groups still constitutes a major problem in America.”

In measuring prejudice across diverse samples, the QDI is advantageous as a self-report measure of generalized prejudice for two important reasons. First, unlike some earlier measures of prejudice, the QDI is able to detect some subtle aspects of prejudice (i.e., compared to other self-report measures that assess people’s overt attitudes toward specific racial groups; Burkard et al. 2002). Second, for researchers interested in examining the role of generalized prejudice in diverse samples, the QDI was designed for use across various racial and ethnic groups (Biernat and Crandall 1999; Ponterotto et al. 1995). QDI items range from 1 (strongly disagree) to 5 (strongly agree), and the sum of all items produced a reliable measure of prejudice (α = 0.77), scored such that higher scores represented having more prejudice.

Outcome Variable

Attitudes Toward the NYPD’s Stop-&-Frisk Policy

Of particular interest to the present study were participants’ attitudes toward the New York City Police Department’s Stop, Question and Frisk policy, frequently referred to as “Stop-&-Frisk.” We assessed participants’ attitudes toward Stop-&-Frisk with a single-item: “Do you oppose, favor, or neither oppose nor favor the NYPD’s Stop-&-Frisk policy?” Participants responded to the item by endorsing one of the following three options: 1 (oppose), 2 (neither oppose nor favor) and 3 (favor).

Results

We hypothesized an indirect effect of prejudice in two separate relationships: (1) the relationship between RWA and attitudes towards the NYPD’s Stop-&-Frisk Policy, and (2) the relationship between SDO and attitudes towards the NYPD’s Stop-&-Frisk Policy. We evaluated these models using Preacher and Hayes (2008) bootstrapping procedure for indirect effects (SPSS macro), controlling for gender and age.

The bootstrapping procedure reduces the total effect of a predictor variable on an outcome variable into a specific indirect effect and a direct effect. An indirect effect refers to the causal effect of the predictor variable on the outcome variable that can be accounted for by a mediating variable. The direct effect, moreover, indicates the causal effect of the predictor variable on the outcome variable, controlling for the mediator. The procedure establishes mediation when the total indirect effect is significant, which implies that a direct effect is significantly smaller than the total effect (Preacher and Hayes 2008).

As a nonparametric procedure, bootstrapping generates 5000 resamples from the original sample data to estimate confidence intervals for an effect, and this procedure is preferable to other mediation analyses for three important reasons. First, the bootstrapping procedure can detect indirect effects without relying on assumptions about the normality of the data. Second, bootstrapping reduces the likelihood of Type I error by limiting the number of inferential tests required to establish mediation. Third, bootstrapping is powerful enough to test for mediation in smaller samples. The bootstrapping procedure yields a 95 % confidence interval, and researchers may claim that an indirect effect is present when zero is outside the lower and upper bound of the confidence interval.

Right-Wing Authoritarianism

Although the present data found no support for a direct relationship between RWA and support for Stop-&-Frisk, establishing direct effects is no longer a prerequisite for examining the role of potential intervening variables (Hayes 2009). As can be seen in Fig. 1, the bootstrap procedure revealed an indirect effect of prejudice in the total effect of RWA on attitudes toward Stop-&-Frisk, controlling for age and sex. RWA positively predicted prejudice (a = .38, p = .00); a one standard deviation unit increase in RWA yields a .38 standard deviation increase in prejudice. Being more prejudiced, moreover, positively predicted support for Stop-&-Frisk (b = .20, p = .03), such that each standard deviation increment in prejudice increased support for Stop-&-Frisk by .20 standard deviation units. That is to say, those who were more prejudiced were also more in favor of the Stop-&-Frisk policy. The indirect effect of RWA on support for Stop-&-Frisk through prejudice (i.e., the ab path) was significant, with a point estimate for prejudice of .076 (and a bias-corrected and accelerated 95 % CI of .0055 to .1813). These results suggest that prejudice mediated the relationship between RWA and support for Stop-&-Frisk.

Fig. 1
figure 1

Indirect effect of Prejudice on RWA and Support for Stop-&-Frisk. Values = standardized coefficients (SE) for indirect effects, c’ = direct effect in the full model. **p < .00, *p < .05

Full size image

Social Dominance Orientation

The pattern of results reported for social dominance orientation resemble those reported for RWA. The bootstrap procedure found an indirect effect of prejudice in the total effect of SDO on attitudes towards Stop-&-Frisk, controlling for sex and age (see Fig. 2). SDO positively predicted prejudice (a = .40, p = .00), such that a one standard deviation unit increase in SDO produced a .40 standard deviation increase in prejudice. Being more prejudiced again predicted attitudes towards Stop-&-Frisk (b = .25, p = .01); those who were more prejudiced also tended to favor the Stop-&-Frisk policy. The indirect effect of SDO on attitudes towards Stop-&-Frisk through prejudice (i.e., the ab path) was again significant, with a point estimate for prejudice of .10 (and a bias-corrected and accelerated 95 % CI of .03 to .20). Thus, these analyses suggest that— like the results reported for RWA— prejudice mediated the relationship between SDO and support for Stop-&-Frisk.

Fig. 2
figure 2

Indirect effect of Prejudice on SDO and Support for Stop-&-Frisk. Values = standardized coefficients (SE) for indirect effects, c’ = direct effect in the full model. **p < .00, *p < .05

Full size image

Discussion

In summary, we found an indirect relationship between the ideological attitudes of RWA and SDO and support for Stop-&-Frisk through prejudice. In other words, although RWA and SDO were not directly correlated with support for Stop-&-Frisk, they were indirectly related to it via the degree to which people were prejudiced. To our knowledge, this was the first psychological study to investigate support for the New York City Police Department’s controversial policy.

Given the impact of Stop-&-Frisk on the lives of people of color in New York City, and previous research showing that police officers report higher levels of SDO (Sidanius et al. 1994), these findings further implicate the importance of social-psychological variables in public policy research. Specifically, understanding the role of variables like RWA and SDO in those who choose to adopt social roles that put them into positions of power over others, in those who experience abuses of power, and in third party reactions to justice and injustice can reveal how people come to view policies like Stop-&-Frisk as fair or unfair.

Limitations

While this study discovered a relationship between ideological attitudes and support for Stop-&-Frisk, its findings are limited in two major ways. First, the sample that this study examined was not representative, and more research should be conducted to ascertain the generalizability of these effects. Second, we investigated only the effects of participants’ ideological attitudes on their support for the policy without investigating the influence of other possible factors, such as primary news source and perceived risk of crime.

Although the present results were not obtained from a representative sample, they nonetheless depart from the common WEIRD (Western, Educated, Industrialized, Rich, and Democratic; Henrich et al. 2010) samples used in many psychological studies in some key ways. Specifically, our sample was drawn from a multicultural, urban university in New York City, and was comprised predominantly of people who were from the ages and minority backgrounds most affected by the Stop-&-Frisk policy. While the data from this sample provided some insight into the influence of ideological attitude on support for Stop-&-Frisk, it offered relatively less insight into the attitudes of those most likely to be affected by the policy– young men of color. To more fully examine the relationship between ideological attitudes and support for Stop-&-Frisk would require a sample with more men in general, and more men of color in particular.

Future Directions

Despite these shortcomings, our investigation of the relationship between ideological attitudes and support for Stop-&-Frisk complements the rich literature associated with the dual process motivational model of ideological attitudes (Duckitt 2001; Duckitt et al. 2002), and tests of the dual process motivational model on social policies like Stop-&-Frisk or other phenomena involving police-community relations seem like natural follow-ups to the present research. One benefit of such a test would be that it could identify the divergent predispositions and motivational bases that explain the intervening role of RWA and SDO on prejudice and social policies detrimental to communities of color. If, for example- and in accordance with the dual process model- RWA is predicated on the motivational goal of societal order and security, which is further predicated on possessing a view of one’s social world as dangerous, then campaigns to change restrictive social policies could address those specific concerns. Similarly, if one’s level of SDO is based on the motivational goal of power and superiority over others (which is anteceded by a “social Darwinist” worldview [Duckitt 2006]) then efforts to change police-community relations would be more effective to the degree that they were tailored to persuade people who possess those worldviews and motivational goals. Additionally, studies examining the influence of the dual process model, RWA and SDO in major political, social, or justice-related events— such as the police shootings of unarmed teenager Michael Brown in Ferguson, Missouri (Bosman et al. 2014)— could provide insight into the factors that predict the public’s reaction to justice-related issues.

Future research should also consider the interplay between RWA, SDO, and additional factors that could influence someone’s opinion of social policies, such as primary news source or fear of crime. Examining the effects of local news viewing on prejudice and racial misconceptions, Gilliam and Iyengar (2000) found that local news emphasis on violent crime, when coupled with racial cues, led White participants to support more punitive action and report more negative attitudes toward Black people than Black participants did. Thus, Gilliam and Iyengar’s findings suggest that news source can influence not only perceived risk of crime, but also attitudes towards minorities, which in the present study mediates the relationship between ideological attitudes and support for Stop-&-Frisk.

In addition to the type and degree of media exposure, future research on reactions to social policy should consider variables that facilitate or inhibit fear of crime, including neighborhood disorder, types of homes (single-family v. apartment buildings), neighborhood integration, and responsiveness by authorities (McGarrell et al. 1997). Based on their analysis of facilitators and inhibitors of fear of crime, McGarrell and colleagues suggested that responsiveness by authorities may reduce fear of crime in poorer, high crime inner-city neighborhoods. If policies like Stop-&-Frisk tend to reduce fear of crime, then perhaps residents of high-crime areas would view policies like Stop-&-Frisk more favorably, and questions like these would be amenable to testing in future research on social policies that involve police-community interactions.

Investigating individual’s perception of the police’s use of deadly force, Perkins and Bourgeois (2006) found a negative correlation between SDO and perceived misuse of deadly force. Recent research, moreover, found that people who were high in RWA were trusting of officers in hypothetical situations even after they killed an unarmed terrorism suspect (White et al. 2008). These studies provide a foundation for future examinations of people’s reactions to police shootings, and they outline a framework for investigating people’s support for Stop-&-Frisk- or similar social policies nationally- following police searches that reveal or fail to reveal incriminating evidence.

In keeping with the idea that RWA and SDO might warrant more investigation within the context of constructing or maintaining social policies in general, and within policy-community interactions in particular, previous research on the shooter bias (Correll et al. 2002) has employed task-specific video games to investigate how quickly people choose to shoot or not shoot targets that are variously armed or unarmed and White or Black. This work brings into question the effects of ideological attitudes involved in the processes leading up to, for example, a police officer’s decision to stop a citizen for frisking and questioning. Furthermore, later research by Correll and colleagues (Correll et al. 2011) found a moderating effect of dangerous environments in explaining the shooter bias, showing that shooter bias disappeared when participants had to decide to shoot targets in environments considered to be dangerous. More specifically, the authors attributed the decrease in bias to the increased shooting of White targets, relative to environments not considered dangerous. The perceived danger of an environment may therefore explain additional variability in people’s support for Stop-&-Frisk, and this seems especially plausible given that people with high RWA tend to perceive their world as dangerous.

In conclusion, our research provides evidence that prejudice explained the relationship between both RWA and SDO and their respective support for Stop-&-Frisk in a sample drawn from a large multicultural urban university in New York City. Continuing to research the influence of ideological attitudes on attitudes toward policy is important to understanding the strong reactions that some members of the public have in the face of significant justice issues.