INTRODUCTION
The centrality of social identity as essential to explaining human interactional processes is now a well-established fact. The assumption of socially derived identity forms the core of several prominent theories explaining the construction of individual identity, including social psychology’s social identity theory (SIT; Tajfel Reference Tajfel and Moscovici1972), sociology’s structural symbolic interactionist (or “identity”) theory (Stryker Reference Stryker1980), and developing notions in personality psychology, for example, the proposition by McLean et al. (Reference McLean, Lilgendahl, Fordham, Alpert, Marsden, Szymanowski and McAdams2018) that narrative identity draws its content from culture-based master narratives. SIT alone has proven a fertile ground for theorizing macro- and microsociological phenomena, spawning self-categorization theory (Turner et al. Reference Turner, Hogg, Oakes, Reicher and Wetherell1987), the theory of optimal distinctiveness (Brewer Reference Brewer1991), uncertainty-identity theory (Hogg Reference Hogg2007), and social identity theories of intergroup conflict (Tajfel and Turner Reference Tajfel, Turner, Austin and Worchel1979) and of leadership (Hogg Reference Hogg2001). Criticisms of the original theory have also sparked important social psychological offshoots, including system justification theory (SJT; Jost and Banaji Reference Jost and Banaji1994) and social dominance theory (Sidanius and Pratto Reference Sidanius and Pratto1999).Footnote 1
Social identity theorists posit that individuals draw self-definition from viewing themselves within larger social groups. Evaluating a group and determining that one belongs to that group is a process called categorization. One may categorize others (social categorization; Jenkins Reference Jenkins2000) or herself (self-categorization; Turner et al. Reference Turner, Hogg, Oakes, Reicher and Wetherell1987). When categorizing the self, the individual often becomes attached to the group to which she feels she belongs, and it becomes a definitional part of her identity – a process called identification (Hogg and Reid Reference Hogg and Reid2006). Identification has a direct influence on individual behavior because, like other bio-psychologically social creatures from bees to baboons, humans seek belonging in social groups which requires perceiving and adhering to group norms of behavior. Individuals perceive the corpus of norms and organize them into a mental picture of the ideal group member – a prototype – to which they compare their behavior and attempt, in most cases, to emulate (Hogg, Abrams, and Brewer Reference Hogg, Abrams and Brewer2017). Social psychologists refer to those groups to which we belong as in-groups, and those individuals and groups outside of this body as a variety of out-groups (Tajfel and Turner Reference Tajfel, Turner, Jost and Sidanius2004). We prefer our in-group to out-groups, even when we are not intimately connected to our perceived in-group (Hogg and Turner Reference Hogg and Turner1985) and even at only the slightest hint that group membership is salient in a given situation (Goette, Huffman, and Meier Reference Goette, Huffman and Meier2006). These preferences are powerfully influential across cultures, social context, and are, as researchers in the growing field of social neuroscience have shown, biologically rooted (Cikara and Van Bavel Reference Cikara and Van Bavel2014).Footnote 2
The concept of social identity, increasingly prominent throughout the field of justice studies and specifically in criminology, has been shown to be important in the explanation of criminal psychology and behavior (Boduszek et al. Reference Boduszek, Adamson, Shevlin, Hyland and Bourke2013).Footnote 3 In criminal justice, Lind and Tyler’s (Reference Lind and Tyler1988) group value model (GVM, the basic model in relational theory; Tyler and Lind Reference Tyler and Lind1992) and the subsequent group engagement model (GEM; Tyler and Blader Reference Tyler and Blader2000) provide the foundations upon which the field’s legitimacy theorization was built, a pre-eminent paradigm in explaining the relationship between governing authorities and those they govern (e.g. Tyler, Goff, and MacCoun Reference Tyler, Goff and MacCoun2015). Legitimacy is the state of being accepted as authoritative, and theories of how legal authorities attain this state are rooted in SIT (Heuer and Stroessner Reference Heuer and Stroessner2011). Recent studies have revisited Tyler’s original propositions, extending and revising the group value and group engagement models for greater and more precise explanatory power (e.g. Bradford Reference Bradford2014; Bradford et al. Reference Bradford, Hohl, Jackson and MacQueen2015). To do so, criminal justice researchers have begun to draw more directly from the broader literature on SIT (e.g. Radburn et al. Reference Radburn, Scott, Bradford and Robinson2018) to embark on building a more comprehensive social identity theory of legitimation. Nevertheless, in so doing, criminal justice scholars may find themselves stepping into debates within social identity discourses that have moved far past where the theory stood at the time Tyler’s work spawned a diverging lineage within our field.
The present study arises from the need to explore a challenge to traditional understandings of SIT, a critique that is not new to social psychology but not yet tackled in criminal justice literature. Scholars in criminal justice (e.g. Harkin Reference Harkin2015) and broader justice studies (e.g. Vainio Reference Vainio2011) have noted overlaps between Tyler’s GVM and Jost and Banaji’s (Reference Jost and Banaji1994) SJT and have suggested that these frameworks be integrated. The collaborative work of Tyler and Jost (Reference Tyler, Jost, Kruglanski and Higgins2007) might be taken to signal that integration here is possible, but there remains a central point of contention that the two did not fully address. Criminal justice scholars basing their work on the GVM aver that legitimation – the psychological process through which individuals evaluate and accept claims to legitimacy – occurs because individuals identify as part of the group the authority represents and accordingly accept norms of deference and obedience to the group’s authorities (Bottoms and Tankebe Reference Bottoms, Tankebe, du Bois-Pedain, Ulväng and Asp2017). Fair and respectful treatment is said to fortify group identification and assure legitimacy acceptance and subsequent compliance. SJT scholars, however, tie legitimation to a separate motive – the system justification motive – which operates independent from group-based concerns and, in fact, is strongest when group membership is least salient and identification weakest. There is an apparent inconsistency here, the import of which cannot be glossed over as two divergent sets of implications follow for theory, research and policy.
In the current study, I use data from a procedural justice survey conducted by the Center for Court Innovation (CCI) in two cities in the United States. Approximating measures of social categorization, criminal justice system (CJS) legitimacy, satisfaction with government, perceptions of CJS bias, and in-group cohesion, I explore the following research question: Does level of cohesion with one’s disadvantaged in-group predict (a) CJS legitimacy, (b) satisfaction with government, and (c) perceived CJS bias?
In the pages that follow, I provide a brief introduction of recent work integrating social identity concepts into explanations of CJS-related legitimacy and justice perceptions, of SJT and its differences with SIT, and of recent social psychological work responsive to tensions between the two.Footnote 4 I then present the current study, outlining the context of its data collection and explaining the measures used, the analytic strategy employed and its results. Finally, I close with a discussion of implications and limitations in this research, along with suggestions for future directions. The goal of this work is to bring criminal justice scholars more soundly into the theoretic debates taking place amongst our colleagues in social psychology (e.g. Jost et al. Reference Jost, Badaan, Gourdarzi, Hoffarth and Mogami2019; Owuamalam, Rubin, and Spears Reference Owuamalam, Rubin and Spears2019), bringing with us the unique perspective of those who study the nature of legal authority in its most palpable context: the imposition of state sanctioning power.
SOCIAL IDENTITY MODELS OF LEGITIMATION
That social identity plays a role in how individuals perceive and judge authorities is a central tenet of the GVM of procedural justice (Lind and Tyler Reference Lind and Tyler1988). According to the GVM, when a person encounters an authority, how that authority treats her or him is taken as representative of how the group that authority represents conceives of the individual’s group standing. Put differently, where individuals are treated by authority figures respectfully, this conveys that they are valued as members of their group; when they are treated disrespectfully, this reflects their diminished standing in the group.
The GVM is premised on the assumption that authority–subject encounters are intra-group, i.e. that both the authority figure and the person she or he encounters are part of the same in-group (Davis-Lipman, Tyler, and Andersen Reference Davis-Lipman, Tyler and Andersen2007). Under this assumption, the subject individual will be concerned about her or his place in the group, as SIT teaches that people are heavily guided by a desire to maintain esteem in their in-groups. Respectful treatment affirms the individual’s esteem – their value in the group – because authority figures are seen as representative of the in-group as a whole. Once affirmed, the desire to keep one’s behavior in line with prototypical group norms compels the individual to comply with the authority. The causal chain is thus: (a) self-categorization and identification as an in-group member to (b) concern about in-group standing (group identity salience) to (c) social categorization of authorities as in-group authorities to (d) affirmation of in-group status by in-group authorities’ respectful treatment to (e) motivation to maintain in-group standing by accepting its authorities’ legitimacy to (f) adherence to a group norm of compliance with its authorities.
Traditional recitations of social psychological work on social identity within criminal justice literature have focused on the GVM, and so I do here. Nonetheless, it should be noted that recent work on the intersection of social identity and legitimacy has relied instead on an alternative explanation, the GEM (e.g. Bradford et al. Reference Bradford, Hohl, Jackson and MacQueen2015). The GVM and GEM overlap substantially in their core conceptualizations, with the exception that the GEM places part (c) in the above causal chain at the start. In other words, when authorities encounter individuals, their fair and respectful treatment of the individual signals common group membership and this prompts in-group identification and the subsequent motivation to adhere to norms around legitimacy and compliance (Tyler and Blader Reference Tyler and Blader2000). Re-situating the aforementioned causal chain, the GEM posits a sequence beginning with (a) social categorization of authorities as members of some group to (b) authorities’ respectful treatment signaling their social categorization of the individual as a member of the in-group to (c) acceptance of authorities’ social categorization, self-categorizing as in-group member to (d) identification with the in-group to (e) motivation to maintain in-group standing by accepting its authorities’ legitimacy to (f) adherence to a group norm of compliance with its authorities.
These two models form the foundation of much of the work on the legitimacy of legal authorities done by criminal justice scholars and social psychologists alike. Both explain legitimation as a result of group value shown primarily through respectful and fair treatment. According to procedural justice theory, individuals judge the fairness of their treatment by the quality of decision-making and the quality of treatment they receive at the hands of authorities (Tyler Reference Tyler2007). Just treatment consists of individuals encountering an authority and being treated with dignity and respect, given an opportunity to be heard (given voice), informed of the bases for decision-making (evidence of transparency), and treated without bias (neutrality) (Tyler Reference Tyler2004). Most legitimacy scholars in the criminal justice realm have based their work on this theory, though explicit mention of its moorings in the GVM (or the GEM) is becoming less conspicuous in the literature. This is perhaps because few if any of the debates in criminal justice scholarship challenge this underlying premise.
On the other hand, there is lively debate about the precise definition of legitimacy, its constituent parts, antecedents, correlates and products. Tyler has defined legitimacy as “the property that a rule or an authority has when others feel obligated to voluntarily defer to that rule or authority” (Tyler Reference Tyler2006:311), comprised also of individuals’ trust in legal authorities (Tyler and Huo Reference Tyler and Huo2002). In this telling, the obligation felt to obey and trust forms the constituent core of legitimacy, with compliance being the observable product. Subsequent work, however, characterizes trust as a distinct but highly related correlate of legitimacy (Hamm, Trinkner, and Carr Reference Hamm, Trinkner and Carr2017) and argues that the obligation to obey is a product of legitimacy, not its heart (Bottoms and Tankebe Reference Bottoms and Tankebe2012). Procedural justice is, under most conceptualizations, seen as an antecedent of legitimacy, but Bottoms and Tankebe (Reference Bottoms and Tankebe2012) argue that it, along with perceptions of authorities’ lawfulness, effectiveness and justness in the distribution of goods, is constitutive of legitimacy. More recently, scholars have asserted that legitimacy is best defined as the normative alignment between authorities and their subjects (Jackson et al. Reference Jackson, Bradford, Hough, Myhill, Quinton and Tyler2012; Tyler and Jackson Reference Tyler and Jackson2014). Still, settlement on any one of these definitions does not impact an overarching premise that legitimacy, whatever it is, is a product of the social identification process.
There is much debate surrounding the antecedents of legitimacy. Study after study has demonstrated the predominance of procedural justice concerns in determining legitimacy judgments (Wolfe et al. Reference Wolfe, Nix, Kaminski and Rojek2016). However, studies of legitimacy using samples outside the “Western world” have found that effectiveness can be equally, if not more, important (Tankebe Reference Tankebe2008, Reference Tankebe2009, Reference Tankebe2013; Tankebe, Reisig, and Wang Reference Tankebe, Reisig and Wang2016). There is a panoply of “justices” said to be important to legitimacy assessments, while moral alignment and benevolent motive have also been implicated in the evaluative process (Jackson and Sunshine Reference Jackson and Sunshine2007). Nevertheless, whatever the aspects of treatment important for legitimacy judgments (or even whether outcomes outweigh treatment; Blader Reference Blader2007), legitimation occurs because an individual believes that she or he is a valued member of the group, and subject to its normative demands.
SOCIAL IDENTITY MODELS AND THEIR CHALLENGERS
The more substantive challenges to the GVM have come mainly from work outside of criminal justice. One significant critique has come from the work of Stähl and colleagues, who have shown the inadequacy of the traditional GVM in addressing what happens when one does not self-categorize as, nor identify with, the group that the authority represents (Stähl, Van Pooijen, and Vermunt Reference Stähl, Van Pooijen and Vermunt2004; Stähl, Vermunt, and Ellemers Reference Stähl, Vermunt and Ellemers2006, Reference Stähl, Vermunt and Ellemers2008). This is an important situation in the context of criminal justice as significant proportions of the governed do not see themselves represented by legal authorities (Bell Reference Bell2017; La Vigne, Fontaine, and Dwivedi Reference La Vigne, Fontaine and Dwivedi2017). One might expect that individuals who do not value group membership because they socially categorize legal authorities as out-group would not be swayed by a need to be incorporated into the in-group. Stähl’s work (e.g. Stähl et al. Reference Stähl, Van Pooijen and Vermunt2004) shows this to be true, and SIT provides a ready answer for this seeming complication. While the number of studies are somewhat limited, it turns out that those who encounter out-group authorities still have identity group concerns at play. In this case, individuals seek affirmation of out-group authorities’ respect for the individuals’ in-groups (conveying the respect of the group that authorities represent; Heuer and Stroessner Reference Heuer and Stroessner2011), as well as satisfaction with the legitimacy of the decision-making process and its outcomes (Stähl et al. Reference Stähl, Van Pooijen and Vermunt2004). Therefore, it appears that the group value and group engagement models of legitimation may stand, though as part of a more comprehensive multi-value model (see Heuer and Stroessner Reference Heuer and Stroessner2011).
Critiques that the GVM neglects to explain out-group evaluations of authorities are not a challenge to the SIT upon which the model is founded. Interestingly, an alternative theory that has gained more traction within criminal justice circles (e.g. Harkin Reference Harkin2015) as a candidate for integrating and expanding social identity models of legitimation is one more critical of the propositions of SIT, one that suggests not extension but instead diminution of SIT in legitimacy theory. SJT, at its simplest, argues that individuals’ social behavior is driven not only by the two motivational bases acknowledged in SIT literature (personal and group interests) but also by a distinct motivation to uphold system interests (Jost and Banaji Reference Jost and Banaji1994). System justification is the tendency of people to accept and defend the legitimacy of the status quo social order (Jost and Hunyady Reference Jost and Hunyady2002). This is hypothesized to aid in their making peace with their place within that social order. Accepting the assumptions of just world theory (Lerner Reference Lerner1980), if the social order is fair and just, people’s place within it is fair and just, and they do not have to contend with the discomfort that comes with living in an unjust social state without power to change it.
SJT was devised as an answer to the frequent observation that individuals often act counter to the predictions of SIT. Instead of acting in the interest of their identity groups or themselves, people often accept, comply with, and even defend systems designed to their disadvantage. More pointedly, at an intergroup level, individual members of disadvantaged social groups often show out-group favoritism instead of bias towards their own. This phenomenon can be seen in disadvantaged groups even where they explicitly espouse in-group centrism. A quintessential example lies in studies that find American Blacks who embrace Afrocentrism nonetheless show evidence of implicit bias against other Blacks (Jost, Banaji, and Nosek Reference Jost, Banaji and Nosek2004; also Banaji, Greenwald, and Rosier Reference Banaji, Greenwald and Rosier1997, referenced in Jost et al. Reference Jost, Banaji and Nosek2004) or, placed in rapid decision situations, perceive other Blacks as more threatening than members of other races (Correll et al. Reference Correll, Park, Judd and Wittenbrink2002).
Jost et al. (Reference Jost, Banaji and Nosek2004) provide a comprehensive overview of 20 central hypotheses coming from SJT and the literature supporting it at that time (see Jost Reference Jost2019 and Osborne, Sengupta, and Sibley Reference Osborne, Sengupta and Sibley2019 for more recent reviews). Within SJT, system justification is a synonym of legitimation; however, the construct is painted in a much different light. Akin more to Reference SuchmanSuchman’s (1995) notion of cognitive legitimacy, legitimation from a SJT perspective involves the presence of an operating social order (a system) that manages the impressions of itself held by its subjects through defining standards of legitimacy, socializing the governed to accept those standards, promoting its (the system’s) adherence to those standards, and relying on human desire for certainty to lead subjects to take for granted its legitimacy, dulling the desire for subjects to question it (Jost et al. Reference Jost, Becker, Osborne and Badaan2017). Adapting the causal chains above, SJT explains legitimation as beginning with (a) self-categorization as a system member and identification with the existing social structure to (b) concern about its fairness and legitimacy to (c) rationalizing its fairness and legitimacy to (d) interpreting authorities’ treatment as legitimate to (e) motivation to maintain system standing by adherence to a social norm of compliance with system authorities. SJT was conceived as a macro explanation of individuals’ relationships with social systems, and so was not originally constructed to explain outcomes of individual authority figure-to-subject interactions, but its premises are easily fitted to that purpose.
In his work with Jost, Tyler at least accepts that legitimation as a psychosocial process can be manipulated by system actors to engender a sense of legitimacy, though the authors do not specify where, if anywhere, group-based processes fit into that equation (Tyler and Jost Reference Tyler, Jost, Kruglanski and Higgins2007). To be clear, SJT, while originally positioned as a challenge to SIT, need not necessarily be mutually exclusive of it. SJT scholars accept that individuals are, in fact, driven by group interests and that it is possible for group interest motives to overcome and guide behavior over system justification motives. Indeed then, SJT may be offered as a fuller understanding of individuals’ relationships with social and governmental systems, of which group value and group engagement models are just a piece.
Recent work, though, rebuts SJT’s challenge in favor of another explanation of findings seeming to suggest a tendency towards system justification. Chuma Owuamalam and colleagues propose a social identity model of system attitudes (SIMSA), which holds that system justification is, in fact, an identity management strategy by which disadvantaged group members do justify disadvantageous systems, but only under certain circumstances unspecified and masked in the findings of SJT studies (Owuamalam, Rubin, and Issmer Reference Owuamalam, Rubin and Issmer2016a; Owuamalam, Rubin, and Spears Reference Owuamalam, Rubin and Spears2016b, Reference Owuamalam, Rubin and Spears2018a, Reference Owuamalam, Rubin and Spears2018b, Reference Owuamalam, Rubin and Spears2019; Owuamalam et al. Reference Owuamalam, Rubin, Spears and Weerabangsa2017). In short, whereas SJT proposes system justification as evidence of a motive independent of group interest, Owuamalam argues that system justification is marshalled in the service of group interest.
The SIMSA advances three SIT-based explanations of system justification (Owuamalam et al. Reference Owuamalam, Rubin and Spears2019). A social reality explanation suggests that what appears to be system justification may simply be an acknowledgement of social reality as individuals view it. This need not be interpreted as a proactive motive to justify systems nor necessarily indicate less identification with and pride in one’s disadvantaged in-group. In fact, it may be accompanied by the hope that one’s in-group will advance in the social hierarchy in the future and this hope itself is motivation to believe in (to hope for) a fair system that allows mobility (Owuamalam et al. Reference Owuamalam, Rubin and Issmer2016a). Social reality may also lead disadvantaged group members to suppress openly expressing in-group favoritism to members of the dominant out-group (e.g. to which most researchers likely belong) in order to avoid imperiling the advancement of the group by inciting fear and retaliation by dominant group members (Owuamalam et al. Reference Owuamalam, Rubin and Issmer2016a).
The SIMSA puts forward a second SIT explanation of apparent system justification: what might be being observed is superordinate in-group favoritism (Owuamalam et al. Reference Owuamalam, Rubin and Spears2019). For example, Black Americans, by many measures, are disadvantaged as a group in the United States (Williams Reference Williams2018). However, Black American culture is distinctly “American,” detached, as it were, from its African moorings and centered now on a unique Black American experience from which arises a significant skepticism towards governing systems and the status quo social arrangement (Blount-Hill and St. John Reference Blount-Hill and St. John2017; Haider-Markel and Joslyn Reference Haider-Markel and Joslyn2017). Nonetheless, Black culture draws heavily from the mainstream American culture of which it is a part (Nunnally and Carter Reference Nunnally and Carter2012). Thus, while accepting issues in the American system, one might nonetheless expect support for the basic structure of the United States society among Black Americans, even if it is less enthusiastic than that among their White counterparts and accompanied by caveats on needing to resolve issues of race. As a third group-interest explanation for system justification, Owuamalam et al. (Reference Owuamalam, Rubin and Spears2019) argue that hope for the future advancement of one’s disadvantaged in-group, as mentioned above, provides a strong motivation to find the current system flexible and unstable enough – or, even better, fair enough – to allow the group to advance itself, albeit with much effort.
THE CURRENT STUDY
The propositions of the SIMSA provide multiple points of divergence from SJT ripe for empirical inquiry. The data at hand provide the ability to explore just one. SJT would suggest that identification with the disadvantaged social group would have a negative relationship with system justification. That is, individuals with lower levels of identification with their disadvantaged in-group should be more likely to hold system-justifying views. On this, Owuamalam et al. (Reference Owuamalam, Rubin and Issmer2016a, Reference Owuamalam, Rubin, Spears and Weerabangsa2017) strongly differ. According to a SIMSA perspective, those individuals with a stronger identification with their disadvantaged in-group will be more likely to justify the system so long as they have hope to advance their group’s position in it. As such, I explore below the following: Does level of cohesion with one’s disadvantaged in-group predict (a) legitimacy of the CJS, (b) satisfaction with government, and (c) perceived CJS bias?
Owuamalam and his colleagues have not yet tested the propositions of the SIMSA on a United States sample, so this study advances their body of work by doing so. Furthermore, the Black American experience of disadvantage and systemic discrimination in nearly every aspect of social life presents a much different context for an exploration of the SIMSA than did previous Malaysian samples (either the Malaysian Chinese who are disadvantaged politically but economically advantaged or Malay Malaysians in the reverse position), or those drawn from Australian or German populations. As Owuamalam’s studies have focused exclusively on student samples, the current study, albeit exploratory, constitutes the first exploration of the SIMSA using a non-student population of which I am aware. In fact, though Jost (Reference Jost2019) has disputed Owuamalam’s conclusions on SJT on theoretic grounds, this study is the first exploration of the SIMSA that I am aware of after those done by Owuamalam and colleagues themselves, let alone within the field of criminal justice and on a population of individuals with the marginalizing experience of criminal justice involvement.
In the present statistical study, I use data collected as part of a study of procedural justice theory by the CCI, a criminal justice-focused non-profit organization that pilots and evaluates justice-related policy and programming. CCI administered 807 surveys to respondents in Newark, New Jersey (399), and Cleveland, Ohio (408), from approximately May until December 2016 (Swaner et al. Reference Swaner, Ramdath, Martinez, Hahn and Walker2018). The study sought respondents with current or previous experience with a criminal justice case, and recruited these from county and municipal courthouses, community organizations serving post-conviction populations, and in several outdoor public spaces where individuals with justice system involvement might congregate.Footnote 5 CCI used a convenience sample for its study, following other studies of criminal justice perceptions (e.g. Allen Reference Allen2016). Convenience sampling does not produce generalizable findings, though the still exploratory nature of the question I pose make analyses of these data meaningful in suggesting novel avenues of research. Over the course of the study, the team tracked sample composition to assure enough variance on important demographic variables such as age and gender.
Unfortunate disparity in the racial composition of those subjected to criminal prosecution in the United States ensured a final sample that was heavily skewed towards Black Americans – approximately 73 % of the sample was Black with no Latino heritage.Footnote 6 For my analyses, I choose to focus exclusively on Black respondents, reducing the data analyzed here to 590 respondents. This choice was made for two reasons: First, the questions above concern the social identity implications for system justification by members of disadvantaged groups, a category to which Black individuals sadly belong (Blount-Hill and St. John Reference Blount-Hill and St. John2017; Haider-Markel and Joslyn Reference Haider-Markel and Joslyn2017; Williams Reference Williams2018). Second, even if I had desired, the great overrepresentation of non-Latino Black respondents in the overall sample limits drawing any meaningful conclusions about racial differences, while excluding non-Black respondents ensures conclusions about Black responses are not skewed by responses from those of other races.
The survey contained 109 items using five-point Likert scales, including on attitudes towards the law, the overall justice system, and justice agencies, such as the local courts. At the end of each survey, the interviewer collected demographic information from the respondent, such as respondent gender, age, race, language proficiency, nationality, education level, marital status, and whether the respondent had full-time employment and/or jail experience. Table 1 shows the demographics of the sample, which did not change much from those of the overall sample. The survey was designed to take 15 to 20 minutes to complete. At its conclusion, respondents earned a gift card of five dollars for Dunkin’ Donuts, a worldwide pastry and coffee franchise.
Table 1. Sample Demographics
PRINCIPAL MEASURES
Since the survey instrument used for CCI’s study was not originally designed for my research question, a detailed discussion of the measures used for analyses is necessary to fully inform interpretation of the results. The research question posed requires measurement of two perceptual constructs: (1) system justification and (2) social cohesion. Furthermore, because the original researchers did not intent to measure system justification and, therefore, did not include direct measures of the construct, I use three measures of the justification process identified in the propositions of SJT: respondents’ scored responses on (a) perceived CJS legitimacy, (b) satisfaction with the government, and (c) perceived CJS bias. I explain the measures used for these constructs in turn.
System Justification
I measure system justification using three factor-derived constructs, including perceptions of (a) CJS legitimacy, (b) satisfaction with government, and (c) system bias and cynicism towards the CJS. SJT suggests that legitimacy is a product – or “consequence” of system justification, operationalized by SJT scholars as “trust and approval of the government” (Jost and Hunyady Reference Jost and Hunyady2002). This definition overlaps with Jackson and colleagues’ (Reference Jackson, Bradford, Hough, Myhill, Quinton and Tyler2012) recent conceptualizations of legitimacy as normative alignment, i.e. “the perceived appropriateness of an institution” (Huq, Jackson, and Trickner Reference Huq, Jackson and Trinkner2017:1106). CCI’s survey instrument included two measures of respondents’ belief in general norms regarding the legitimacy of the law, which included item 20, “Laws are intended to protect people,” and item 21, “People should obey the law even if it goes against what they think is right.” Next, the survey included beliefs about the legitimacy of the police, specifically: item 51, “You would call the police if you were in trouble,” item 52, “The police generally have the same sense of right and wrong as you do,” item 53, “If a violent crime were to occur near to where you live, you can trust the police to arrive quickly at the scene,” item 54, “The police are usually trying to protect and look out for people,” and item 55, “The police are generally respectful.” Respondents were asked about the legitimacy of the municipal court system, namely, if their local judges were “fair in their decisions,” item 76, or “out of touch with what’s going on in their communities,” item 77. Finally, respondents were asked about the legitimacy of the correctional system, with items 94, “If a fight broke out in the facility, the correctional staff would arrive quickly and break it up,” item 95, “The correctional staff are too quick to use force against inmates,” item 96, “The correctional staff are usually trying to protect and look out for inmates,” and item 97, “The correctional staff are generally respectful.” These items were measured using a five-point Likert scale, with 5 being “strongly agree” and 1 being “strongly disagree.”
I conducted an exploratory factor analysis, a statistical test in which patterns of movement in a dataset are examined to determine whether an underlying construct (a latent factor) not captured by direct measurement in the data nonetheless can be predicted by measuring its unseen influence on dependent items that were directly measured. Factor analysis revealed the presence of one factor (λ = 5.24) upon which the two general law legitimacy items, every police legitimacy item, one of the two judicial legitimacy items (item 76) and three of the four correctional legitimacy items (all but item 95) loaded at above .4. Table 2 presents the findings of this analysis. I used Stata’s “predict” function to predict factor scores for each observation in the data, creating values for a variable I call CJS legitimacy.
Table 2. Factor Analysis of Consequences of System Justification (n = 590)
Jost and Hunyady (Reference Jost and Hunyady2002:263) also note that positive or negative effects follow as consequences of system justification, operationalized as “self-report ratings of (a) happiness, satisfaction, contentment, and general positive affect.” I have previously argued that satisfaction with the job performance of an authority is a component of its perceived legitimacy, in line with work in the political sciences (Gibson and Nelson Reference Gibson and Nelson2014) and with Tankebe’s incorporation of “effectiveness” as a measure of satisfaction with an authority’s service provision (Tankebe Reference Tankebe2008, Reference Tankebe2009). However, in keeping with SJT’s differentiation of the two – in line with recent practice within legal legitimacy scholarship – here I treat satisfaction as an empirically separate construct and another consequence of system justification. CCI’s survey instrument includes measures of satisfaction with government, using a similar Likert scale, to rate prosecutors (item 25), defense attorneys (item 26), judges (item 27), the court system (item 28), people who run the jail (item 29), the mayor (item 30), the sanitation department (item 31), the fire department (item 32), and the public schools (item 33). I subjected these to a separate factor analysis and every item, save item 26, loaded onto a single factor (λ = 3.26).Footnote 7 Table 2 shows the results of this analysis. I again used the “predict” function to create factor scores for a variable I call satisfaction with government.
Finally, prominent SJT scholars suggest “fairness of the economic system” as a standard operationalization of perceived legitimacy, along with normative alignment. CCI’s instrument included item 22, “Bankers, lawyers, and politicians get away with breaking the law every day,” and item 23, “The criminal justice system is racist,” along with item 78, “African Americans get treated worse by the courts than other groups,” item 79, “Latinos get treated worse by the courts than other groups,” item 80, “Those who do not speak English get treated worse by the courts than other groups,” and item 81, “Poor people get treated worse by the courts than other groups.” These measure perceptions of system bias, which is related to legitimacy (e.g. the neutrality concept in procedural justice), in this case, judicial legitimacy. These items, however, might also be, together, measures of legal cynicism, distrust in the motives of the legal system and belief that it is not a reliable resource and may, in fact, be proactively oppressive (Kirk and Matsuda Reference Kirk and Matsuda2011). While related to legitimacy, legal cynicism is generally conceptualized as a distinct psychosocial phenomenon (Nivette et al. Reference Nivette, Eisner, Malti and Ribeaud2015). The bias items were entered in the same factor analyses as the legitimacy items (as seen in Table 2) and items 78 through 81 loaded onto a separate factor (λ = 1.79), the second of two which met the Kaiser criterion for eigenvalue significance (Fabrigar et al. Reference Fabrigar, Wegener, MacCallum and Strahan1999). I used the “predict” function to estimate factor scores for the construct, though it was still unclear whether it should be considered a measure of cynicism or legitimacy. In the following analysis, I include it as a separate variable called perceptions of CJS system bias.
Importantly, Jost and Hunyady (Reference Jost and Hunyady2002) propose these consequences as almost inevitable, like how legitimation is a process distinct from its product – legitimacy – but inevitably linked to it. Therefore, measuring system justification by its consequences serves, as measures of legitimacy do, to identify the operation of the underlying psychological production process.
Social Cohesion Within Respondents’ Neighborhood In-Groups
Both Newark and Cleveland are over 50 % Black in terms of population demographics, and CCI recruitment took place in areas with strong concentrations of Black residents. The neighborhoods within each city where residents resided also revealed that these were areas with heavy Black concentrations. Consequently, when CCI asked respondents to reflect on the social cohesion of their neighborhood, those places reflected Black communities – and, unfortunately, those very communities disadvantaged by current status quo social arrangements. The CCI survey instrument included several items measuring collective efficacy and social cohesion. SIT literature is clear that social group identification does not necessarily lead to social cohesion (Hogg and Turner Reference Hogg and Turner1985), but that group identification is an important part of social cohesion (Huddy Reference Huddy, Sears, Huddy and Jervis2003). Because the present dataset does not include social group identification as a measure, I derived a factor variable – called cohesion – to use in its stead. Because social identification can occur where social cohesion does not, but social cohesion necessarily indicates a degree of group identification, the use of social cohesion in these analyses is a limited, but conservative, estimate of group identification. My use of perceived social cohesion within Black respondents’ Black neighborhood in-groups means that significant findings using the social cohesion variable likely underestimate the influence of group identification on the dependent variables of interest.
To derive the social cohesion metric, I conducted factor analysis of 10 items. Item 6 asked, “How well do you know the people in your neighborhood?” and was measured on a five-point Likert scale and reverse coded such that “not at all” was assigned a value of 1 and “very well” a value of 5. Item 7 was a four-point Likert scale that asked, “How involved are you in neighborhood events/activities?” and was coded such that “not at all” received a value of 1 and “heavily” received a value of 4. The remaining nine items were measured on a five-point Likert scale and requested respondents to state their level of agreement (1 = strongly disagree, 5 = strongly agree) with statements preceded by “In my neighborhood …” These included item 8, “People treat each other with respect,” item 9, “People look out for each other (e.g. when someone is sick, take care of each other’s pets),” item 10, “People try to look out for kids and teenagers (e.g. take care of each other’s kids, intervene if the youth were in trouble),” item 11, “People would help a resident who has been hurt (e.g. mugged, robbed, jumped),” item 12, “People feel safe being outside,” item 13, “If there were a local school closing, people would organize to try to keep it open,” item 14, “If there were a shooting nearby, people would try to raise awareness and give support to neighbors,” and item 16, “If people need help, they go to informal resources such as local clergy/pastors, family, friends, or unelected leaders.” Three of the items (8, 10, and 11) loaded onto the single factor that met the Kaiser cut-off criterion (λ = 1.55). Using the “predict” function, I estimated scores for this latent factor and, for ease of interpretation, adjusted the scale by rounding it to the nearest whole number and shifting scores along the x-axis by three units. This produced a variable called cohesion. (Results from this analysis are available upon request.)
Van der Eijk and Rose (Reference Van der Eijk and Rose2015) suggest caution both in using the Kaiser criterion (despite its being the most popular way of validating factors) and the potential for over-dimensionalization in the use of Likert-scaled items. They show that Likert scale items can be used appropriately in place of continuous variables (harkening to Lord Reference Lord1953), but that more sophisticated methods of validation must be used before definitive conclusions can be drawn, including subsequent confirmatory factor analyses. Here, I set aside further factor analyses for another day, as the exploratory nature of my questions and the secondary nature of my data make me skeptical of pursuing any more definitive factor findings than I have presented here. However, future studies, using data more fitted to purpose, should pick up here where I have left off.
Two points of caution: first, the chronological and temporal proximity of the items could have created some degree of acquiescence bias, or “yea-saying” (Savalei and Falk Reference Savalei and Falk2014). The survey positioned the social cohesion/collective efficacy questions consecutively, all measuring levels of agreement using the same scale and all worded to make strong agreement (“5”) a generally positive response and strong disagreement (“1”) a negative one. This may have made responses to one item more a residual measure of preceding responses. However, the results from item-by-item Pearson’s correlation analyses seemingly disconfirm this suspicion. While the collective efficacy items significantly co-vary (α = .66), the specific values of single items are not strongly correlated with those of any other single items.Footnote 8
ANALYSIS AND RESULTS
I used Stata statistical software to perform a multivariable ordinary least squares (OLS) linear regression, a test of the degree to which unit change in one or more (independent) variables correlate with changes in an outcome (dependent) variable. Recall, the research question was, “Does level of cohesion with one’s disadvantaged in-group predict (a) CJS legitimacy, (b) satisfaction with government, and (c) perceived CJS bias?” This required three separate models positioning either (a) CJS legitimacy, (b) satisfaction with government, or (c) perceptions of CJS bias as the dependent variable. The results of these models provide support for both further exploration of the SIMSA within criminal justice research and the applicability of Owuamalam’s SIMSA approach outside of Oceanian student samples.
Using a measure of social cohesion among residents of disadvantaged Black neighborhoods, the results presented in Table 3 show that increased perceptions of social cohesion within one’s neighborhood in-group significantly predicted more positive perceptions of CJS legitimacy (β = .2182, p < .001), along with respondent age (β = .2453, p < .001) and whether the respondent resided in Cleveland rather than Newark (β = .0927, p = .047), R 2 = .1207, F (10, 430) = 5.90, p < .001. These findings support Owuamalam’s findings (Owuamalam et al. Reference Owuamalam, Rubin and Issmer2016a, Reference Owuamalam, Rubin, Spears and Weerabangsa2017) that social identification with, and concern for the interests of disadvantaged groups nonetheless tend to increase the likelihood that one will engage in system justification, not decrease it as SJT originally suggested. Again, this was the only variable amongst demographic controls that significantly predicted satisfaction (β = .2777, p < .001), R 2 = .0865, F (10, 549) = 5.20, p < .001. Social cohesion within Black communities increased (not decreased as SJT would suggest) satisfaction with the CJS. Social cohesion did not significantly predict perceptions of CJS bias (β = –.0023, p = .961), though controlling for it revealed that, the rest being the same, age was a significant predictor of perceptions of CJS bias (β = –.112, p = .023), R 2 = .0435, F (10, 430) = 1.96, p = .0366.
Table 3. Regression Analysis Predicting Consequences of System Justification (Including Social Cohesion) (n = 590)
CJS, criminal justice system; SE, standard error.
* P < .05, *** P < .001.
DISCUSSION AND CONCLUSION
That social identity is of central influence in critical areas of study within the justice disciplines is now little questioned, from criminal psychology (Boduszek et al. Reference Boduszek, Adamson, Shevlin, Hyland and Bourke2013) to psychological responses to legal authorities (Lind and Tyler Reference Lind and Tyler1988). At the same time, it is also clear that theories of identity adapted to criminal justice studies from social psychology and other fields have not always kept pace with developments in their fields of origin nor have contributed to ongoing debates within those disciplines (Radburn et al. Reference Radburn, Scott, Bradford and Robinson2018). SIT, developed by Tajfel (Reference Tajfel and Moscovici1972) and interpreted by Tyler and others in specifying the GVM (Lind and Tyler Reference Lind and Tyler1988) and GEM (Tyler and Blader Reference Tyler and Blader2000), is now strongly embedded in criminal justice scholars’ understanding of legitimacy and justice perceptions and is beginning to be more explicitly so (e.g. Bradford Reference Bradford2014). Yet this embrace has come seemingly without accounting for ongoing challenges to the underlying theory arising from the field that birthed it.
Criminologists have begun to explore applications of SJT in explaining legitimacy evaluations (e.g. Harkin Reference Harkin2015). Incorporating more social psychological theory into the criminal justice mainstream invites more opportunity for broad acceptance without attention to ongoing issues in need of resolution, a fraught path that should be avoided. The present study highlights one such debate and invites other criminal justice scholars to join in addressing it. According to Owuamalam et al. (Reference Owuamalam, Rubin and Spears2019), Jost and Banaji’s (Reference Jost and Banaji1994) original arguments that SJT was a necessary antidote to fatal oversights in SIT were both overblown and inaccurate. In a growing body of retorts, Owuamalam presents an expansion of SIT that accounts for the conundrums that Jost points out and does so, he avers, better than SJT (Owuamalam et al. Reference Owuamalam, Rubin and Issmer2016a, Reference Owuamalam, Rubin and Spears2016b, Reference Owuamalam, Rubin, Spears and Weerabangsa2017, Reference Owuamalam, Rubin and Spears2018a, Reference Owuamalam, Rubin and Spears2018b, Reference Owuamalam, Rubin and Spears2019). While criminal justice scholars have not been privy to this debate, its outcome is consequential, for if proponents of SJT are correct to sound the alarm on overreliance on SIT explanations of social behavior, cornerstone theories of legitimation in criminal justice (i.e. the GVM and GEM) must be reconsidered. If, on the other hand, Owuamalam’s work holds promise, criminal justice researchers should both continue in their confidence of the GVM’s validity and remain skeptical of claims that SJT is a necessary supplement.
These data are, of course, limited in their ability to bring closure to this debate. In this exploratory study using secondary data, I had to make several theoretical and rational choices about the meaning of my measures that would be much more cleanly done had I designed an original study. The findings presented here offer, in my view, the justification to go further and do just that. The sample was not randomly selected, making it impossible to generalize these findings to a general population. This study also took place across two cities, Cleveland and Newark and thus, while controlled for, there may be important, more specific differences in citywide perceptions due to features of the location not adequately captured. A fourth limitation is one that is common to survey and quantitative work: validity. It is impossible to know for sure that constructs adequately operationalize concepts being studied and that is true here as elsewhere. In the end, the limitations of this study are not too tall to invalidate its findings but are notable enough to encourage further study of any results rendered. I encourage this.
The present study finds support for a continued look at the SIMSA. Owuamalam has been very clear about positioning the SIMSA in opposition to SJT on the matter of whether social identification with a disadvantaged group leads to greater (as held by SIMSA) or lesser (as posited by SJT) system justification. The present study finds that sociopsychological connectedness with one’s disadvantaged in-group, as measured by social cohesiveness, seems to increase positive perceptions of CJS legitimacy and satisfaction with the CJS. This finding suggests that a greater centering of group interests leads to more system justification, not less as SJT scholars had supposed. Using the SIMSA’s retelling, SIT is, in fact, suitable to the task of explaining why disadvantaged groups sometimes support disadvantageous systems.
Acknowledgements
Giving first and full thanks to Inspiration, I also thank the Center for Court Innovation for the use of its data, and specifically Josy Hahn, Cassandra Ramdath, and Rachel Swaner for their roles in facilitating the research that made this study possible. I thank the International Society of Criminology and the Education for Justice Initiative of the United Nations Office on Drugs and Crime, as well as the Academy of Criminal Justice Studies, for their awards and recognition of this work, along with the organizers and participants in the 19th World Congress of Criminology. I thank Zavan A. Blount-Hill for his assistance in the completion of this work, Satish G. Nathan for allowing its completion during our Alaska vacation and Chuma Owuamalam for his encouragement to continue in this research.
Kwan-Lamar Blount-Hill, J.D., is a doctoral candidate in criminal justice at the City University of New York Graduate Center, John Jay College of Criminal Justice. He is a senior research manager for the New York City Mayor’s Office of Criminal Justice (MOCJ), a former police officer and former public safety officer (firefighter), licensed attorney, and has served as researcher and lecturer for several institutions. His research centers on institutional legitimacy and public perceptions of justice, specifically as these are influenced by social psychological processes such as identity formation and role performance. The current work was one of ten selected as winners of the Young Scholars’ Research Paper Competition held by the International Society of Criminology in collaboration with the Education for Justice Initiative, United Nations Office on Drugs and Crime, at the 19th World Congress of Criminology in 2019.
