Procedural justice, identity and deference to the law: What shapes rule-following in a period of transition?

Abstract

The group engagement model has two core arguments. The first is that procedural justice shapes rule-following in groups, organisations and societies. The second is that the influence of procedural justice upon rule-following is mediated by changes in people's identification with groups. This study uses a sample of South Africans to test both arguments. While the procedural justice argument has already been widely tested and supported, this study extends that test to a society in rapid transition and upheaval. Further, it tests the identity mediation argument in the same context. The results support both arguments. Procedural justice shapes rule-following and that influence is mediated by identification with superordinate authority.

Keywords:

• Deference to authority
• governance
• identity
• procedural justice
• rule following

Social psychologists have a long-term interest in issues of social justice. This interest flows from the consistent finding that people's justice judgements shape their thoughts, feelings and behaviours in social settings (Tyler, Boeckmann, Smith, & Huo, 1997). Within groups, organisations, and societies justice is found to be of particular importance because it influences people's willingness to cooperate with others (Tyler & Blader, 2000). Most recently considerable evidence has accumulated suggesting that, of the various forms of justice, it is procedural justice that is most central to people's relationships to hierarchical groups (Tyler, 2000). This means that when people are reacting to the institutions and authorities that define their political and legal systems they focus on the procedures through which decisions are made and people are treated. While not insensitive to whether decisions and policies are favourable to them and/or to whether policies give people the outcomes they deserve, people nonetheless focus the bulk of their attention on how decisions are made.

Of particular importance is the argument that procedural justice influences cooperative behaviour. This analysis focuses upon one of the forms of cooperation outlined: deference to rules. All groups need to be able to bring individuals' behaviour into line with group rules and to be able to motivate their members to defer to decisions made by group leaders. Deference can be motivated either by creating credible risks of sanctioning for rule-breaking or by appealing to moral and social values. Deference is particularly valuable if it is obtained voluntarily. Consequently, the ability to encourage value-based deference is particularly important to groups. Procedural justice has been the focus of efforts to encourage such value-based deference, because procedural justice has been found to shape judgements of legitimacy, as well as deference to rules (Tyler, 2006a, 2006b).

The first goal of this study was to test the connection between procedural justice and deference to rules within a new and especially appropriate setting: a country with a history of social conflict, which is newly emerging as a unified and democratically governed society (for other such efforts see Gibson, 2002, 2004; Gibson, Caldeira, & Spence, 2005; Kluegel & Mason, 2004; Machura, 2003).

That society is South Africa. From a political perspective, South Africa typifies the problems of societies seeking to emerge from a period of autocratic government. As an ethnically diverse society, it also must deal with the problems of managing pluralism. And as a society with many economically underdeveloped regions it must deal with issues of poverty, crime, and social disorder. In the international world, where societies change their political frameworks frequently, establishing a stable and effective policy is a challenge (Gibson, 2002, 2004; Gibson et al., 2005).

Using a sample of 2,200 interviews collected in South Africa as part of a broader study of the opinions of Africans called the Afrobarometer, the present study tests the relationship between the perceived procedural justice of the political process and people's deference to the rules of the legal and political system. The analysis tests the argument that such perceived procedural justice is important because it influences people's behaviour, in this case their deference to laws.

Given that the influence of procedural justice is widely known, what is the unique contribution of this study? It is important in several ways. First, it studies how behaviour is shaped in the context of a society in conflict and transformation (i.e., during a period when political institutions were changing and when crime and political unrest were ongoing political issues). Further, this study contributes to the developing literature on procedural justice and the exercise of political authority (also see Clawson, Kegler, & Waltenberg, 2001; Farnsworth, 2003; Gangl, 2003; Hibbing & Theiss-Morse, 2002; Kershaw & Alexander, 2003; Murphy, 2004; Seligson, 2002). Most existing research has focused upon the exercise of authority in the legal and managerial arenas.

Psychology of justice

Social psychologists have identified two types of motivation that might underlie these justice influences. One reason that people might react to justice is instrumental and is linked to people's desire to receive desirable outcomes in groups. Early equity-based models of distributive fairness linked the influence of distributive fairness on cooperation to instrumental beliefs that accepting fair outcomes led to the maximum possible attainment of desirable outcomes (Walster, Walster, & Berscheid, 1978). The procedural justice model of Thibaut and Walker (1975) also linked procedural justice influences to instrumental motivations: in their case to the ability to control decisions and produce desirable outcomes. The distinctiveness of this instrumental model is supported by a variety of empirical findings, such as the demonstration by Tyler (1994) that instrumental evaluations exercise an independent influence upon affect and decision acceptance.

A second reason that people might cooperate is relational. In a series of models, including the group value model of procedural justice (Lind & Tyler, 1988); the relational model of authority (Tyler & Lind, 1992); and the group engagement model of cooperation (Tyler & Blader, 2000, 2003), it has been argued that fair procedures provide people with status-relevant information that helps them to create and maintain a favourable view of the self and positive feelings of self-worth and self-esteem. This argument builds upon the broader suggestion of social identity theory that an important function of groups is to help people to define their sense of self and maintain favourable views about the self (Hogg & Abrams, 1988).

Research has consistently supported the suggestion that relational motivations exercise a distinct influence on justice influences. This relational argument is supported by findings of: “value-expressive” voice effects (Lind, Kanfer, & Earley, 1990; Tyler, 1987, 1989); distinct relational influences on decision acceptance and commitment to the group (Tyler, 1994; Tyler & Blader, 2000); interactions between identity salience and the importance of procedural justice (Huo, Smith, Tyler, & Lind, 1996; Smith & Tyler, 1996; Smith, Tyler, Huo, Ortiz, & Lind, 1998; Tyler & Degoey, 1995); procedural justice influences on self-esteem (Koper, van Knippenberg, Bouhuijs, & Vermunt, 1993; Vermunt, van Knippenberg, van Knippenberg, & Blaauw, 2001); and direct links between procedural justice and people's status-related judgements (De Cremer & Tyler, 2005; Tyler, Degoey, & Smith, 1996; van Prooijen, van den Bos, & Wilke, 2002, 2004a, 2004b).

The prior psychological models of justice outlined have consistently suggested that justice is shaped by both instrumental and relational motivations, with instrumental motivations being the predominant antecedent of distributive fairness judgements, and relational motivations most strongly shaping procedural justice evaluations (Tyler, 1994). The goal of the group engagement model (Tyler & Blader, 2000, 2003) is to propose a single psychological model underlying justice influences. The group engagement model argues that both instrumental and relational justice influences are best understood through a single motivation that links both distributive and procedural justice judgements to cooperative orientations. That motivation is the impact of justice judgements on people's identities.

The argument that groups are important because they provide a fundamental basis for people's self-definitions and feelings of self-worth is central to social identity theory (Tajfel & Turner, 1979). It has also been argued more broadly within group and organisational settings (Ashforth & Mael, 1989; Ellemers, de Gilder, & Haslam, 2004; Haslam, 2004; Haslam, Postmes, & Ellemers, 2003). Research in this tradition supports the centrality of identity by showing that when people identify with groups, they make decisions and act in group-serving ways (Bourhis & Gagnon, 2001; Brewer, 1979; Diehl, 1990; Hinkle & Brown, 1990; Kelly & Kelly, 1994; Ouwerkerk, Ellemers, & de Gilder, 1999; Tajfel, Billig, Bundy, & Flament, 1971).

Beyond the general argument that identification plays an important mediating role in the case of cooperation, recent research suggests that it is superordinate identity that is central to shaping people's rule-based behaviour. Ethnic group identification, in contrast, is not linked to rule-following. This argument is important (see Huo, 2003; Tyler & Huo, 2002) because it suggests that assimilation may not be necessary, as well as that pluralism may be consistent with the effective governance of diverse societies. This study tests the hypothesis that it is superordinate identification that shapes deference to rules. It is, after all, superordinate identification that is linked to the actions of government authorities and institutions. Under other circumstances, and with other behaviours, ethnic identification might be the issue of concern. The key question is what type of behaviour is involved: behaviour linked to superordinate authorities or actions with ethnic group implications.

This study tests the identity mediation argument by examining whether the relationship between procedural justice and deference to rules is mediated by identification. It also includes indices of both superordinate identification (with South Africa) and ethnic identification. This allows us to address two questions: first, whether identity mediates the impact of procedural justice on deference to rules; and second, what level of identification is involved in producing this effect: superordinate or ethnic.

Method

The data in this study are based on Round 1 Afrobarometer interviews conducted in South Africa during the year 2000 (N = 2,200). The Afrobarometer is a survey based upon face-to-face interviews with a random sample of the residents of different African countries, one of which is South Africa. The Afrobarometer is a joint enterprise of Michigan State University; the Institute for Democracy in South Africa; and the Centre for Democracy and Development in Ghana (www.afrobarometer.org). The questions involve structured responses using scales.

The sample was 52% female, and had an average age of 42 years. Fifty-nine per cent had less than a high school degree; 26% were high school graduates; and 15% had some college or more education. The sample was 76% African; 13% White; 7% Coloured; and 4% Indian.

Dependent variables

Following the law. 

Respondents were asked about three possible types of rule-breaking behaviour that they might have engaged in. Each respondent was asked if they had ever engaged in each behaviour. The behaviours were: “claiming government benefits you are not entitled to” (7% said yes); “getting services like electricity or water without paying” (11% said yes); and “using force or violent methods, such as damaging public property to pursue grievances” (14% said yes). Because of the skewed distribution, with most respondents indicating that they had never engaged in such behaviours, the items were not strongly correlated (M r = .23).

Identification

Superordinate identification. 

Two agree–disagree items were used (α = .48): “You want to think of your children as South African” and “You think it is desirable to create one united country”. Although the alpha was relatively low, the two items together better represented superordinate identification than either one alone.

Ethnic identification. 

Three agree–disagree items were used, each framed in terms of the respondent's ethnic group (α = .83): “You are proud to be [ ]”; “You want your children to think of themselves as [ ]”; and “Being a [ ] is a very important part of how you see yourself”.

Justice

Procedural justice. 

Six agree–disagree items were used (α = .85). Respondents were asked to compare political life to the past. Is it better or worse in terms of: “Anyone can freely say what he or she thinks”; “People can join any political organisation they choose”; “People can live without fear of being arrested by the police if they have done nothing wrong”; “Each person can freely choose to vote without feeling forced by others”; “Everyone is treated equally and fairly by the government”; and “All South Africans are equal to each other”.

A second index of the procedural justice of the prior national election was established by agree–disagree responses to the question: “On the whole, how would you rate the freeness and fairness of the last national election, held in 1999?”.

Other factors

Trust. 

Eight agree–disagree items were used to index trust (α = .82), for the question of how much of the time can the following be trusted to do what is right: “the President”; “the government”; “the local government”; “the Defence Force”; “the police”; the Courts”; “the criminal justice system”; and “the electoral commission”.

Provision of basic resources. 

Seven frequency items were used to index access to resources (α = .85). The items asked how often have you: “gone without enough to eat”; “gone without medicine or medical treatment”; “gone without a cash income”; “gone without clean water”; “gone without shelter”; “gone without home electricity” or “gone without enough fuel”.

In addition, respondents were asked to compare their material conditions to their conditions in the past. Two single-item indicators were included, each asking if the government was: “providing an adequate standard of living” and “providing access to basic necessities (food and water)”.

Law enforcement. 

Five agree–disagree items indexed enforcement (α = .82). Respondents were asked “How likely are the authorities to enforce the law” in the following situations: “A person like you committed a serious crime”; “A person like you committed a serious crime and was charged by the police”; “A person like you cheated and did not pay a tax owed the government”; “A person like you obtained services without paying” and “What if they did not pay their correct rates”.

Fighting corruption. 

Four agree–disagree items indexed corruption (α = .87). “How many officials in government are involved in corruption”; “How many people in Parliament are involved in corruption”; “How many officials in the current government are involved in corruption”; and “How many officials in local government are involved in corruption.

Extent of conventional political participation. 

Respondents were asked how often they engaged in conventional, legal, political actions. They include going to rallies, working for one's party, and voting.

Demographics. 

Gender was established by observation. Education and age were determined by self-report.

Results

Do procedural justice and/or identification shape deference to rules?

Structural equation modelling was used to test the various influences on deference to rules (Arbuckle & Wothke, 1995). The three rules considered were treated as indices of a single latent variable reflecting deference to rules. Similarly, superordinate identification; ethnic identification; and procedural justice were each treated as latent variables.

The background indicators were treated as observable variables. The study includes background variables to account for issues of outcome favourability and distributive fairness. Several indices measured whether the government delivered desirable outcomes. This indicated whether people were receiving favourable outcomes. Corruption and enforcement reflected the ability of the government to distribute outcomes fairly. The study also controlled for general political activity by asking people about their level of conventional political participation. Finally, the analysis considered several demographic variables.

Table I presents results of separate regression analyses that show the magnitude of the influence of procedural justice and identification on deference to rules. The equations shown included the effects of the background variables outlined, but the coefficients reflecting their influence were not shown. The results suggested, first, that those respondents who viewed the government as acting through fair procedures were more deferential to social rules. Further, identification also shaped deference. And, as predicted, the key issue was superordinate identification. Ethnic identification was not linked to deference to rules.

Table I. Influence of procedural justice on identification and on deference to rules

	Deference to rules Unstandardised regression coefficient (SE)		Identification Unstandardised regression coefficient (SE)
			Superordinate	Ethnic
Procedural justice	.009 (.004)*	–	.244 (.029)***	.209 (.029)***
Superordinate identification	–	.021 (.008)**	–	–
Ethnic identification	–	–.004 (.003)	–	–
Adjusted R^2	17%	21%	25%	6%
Note. *p < .05; **p < .01; ***p < .001.

An examination of the antecedents of procedural justice (Table II) suggested that procedural justice itself is based upon several factors. These included viewing authorities as trustworthy and as fighting corruption and believing that government provided basic resources. Interestingly, views about whether government was enforcing the law were not related to procedural justice, while people who were objectively materially better off viewed the government as less procedurally just.

Table II. Antecedents of procedural justice

	Procedural justice (H = just)
	Unstandardised regression coefficient (SE)
Trust (H = trustworthy)	.256 (.027)***
Corruption (H = fighting corruption)	.060 (.022)**
Effectively managing problems (H = have basic resources)	−.127 (.019)***
Providing adequate standard of living (H = yes)	.229 (.016)***
Providing necessities (H = yes)	.134 (.016)***
Enforcing laws (H = yes)	.011 (.020)
Gender (H = male)	.008 (.027)
Age (H = old)	−.002 (.001)
Education (H = highly educated)	−.023 (.008)**
Adjusted R^2	42%
Note. p < .05; **p < .01; ***p < .001.

Does identification mediate the influence of procedural justice on deference?

Structural equation modelling was used to test the mediation hypothesis. As before in the analysis of deference, identification and procedural justice were latent variables, while background characteristics were operationalised as observed indicators. The results are shown in Figure 1. Overall, the model fit the data well (Comparative Fit Index [CFI] = .85). The results supported the mediation model. Procedural justice significantly influenced superordinate identification (β = .41, p < .001) and identification shaped deference (β = .22, p < .05). The direct relationship, however, between procedural justice and deference, which was significant in the regression equations shown in Table I, disappeared.

Figure 1. Does identification mediate between procedural justice and deference?

A comparison of the influence of superordinate and ethnic identification suggested that, consistent with past research, it was superordinate identification that was important in shaping people's orientation toward social rules. Those people who had stronger superordinate identification were more willing to defer to the rules. Ethnic identification did not shape people's reactions to social rules. This was consistent with the argument that when the behaviours in question were linked to state actors, it was identification with the state that was crucial to behaviour.

Discussion

The present findings support the suggestion of prior research in the procedural justice tradition that governing via fair procedures is important because it encourages people to accept the rules and policies decided upon by political and legal authorities. This study supports that finding with political authorities and in the context of a society in transition: South Africa.

This analysis focuses on several types of behaviour: breaking legal rules concerning payment for services and eligibility for benefits and participating in political actions that break the law. These behaviours were assessed via self-report and are skewed, with most people saying that they never break the law nor engage in unconventional political actions that violate the law. Hence it is especially striking that the study finds a clear connection between the procedural justice of government activities and law-related behaviour. In both cases those people who evaluate the processes of government as fairer are more supportive of the State and State regulations as expressed by following the law.

In addition to shaping their behaviour, procedural justice also influences people's identification with South Africa, as well as with their ethnic group. Both of these effects are strong, supporting the argument of the group engagement model that procedural justice gives people identity-relevant information. In this study only superordinate identification was then influential in shaping behaviour, but the behaviour in question was related to the government and governmental authorities.

The findings further support the arguments of the group engagement model concerning identity mediation. The influence of procedural justice did not impact directly upon people's behaviour. Rather, as predicted, it shapes people's identification both with South Africa and with their own ethnic group. Of these two forms of identification it is superordinate identification with South Africa that is central to rule-following. Hence, identity mediates the influence of procedural justice on rule-oriented behaviour.

Conclusion

The goal of the group engagement model is to provide a single theoretical framework within which the influence of procedural justice on cooperative orientations toward groups can be understood. The model argues that the key mediating construct is identity. The present results support the argument that identity is a key antecedent of cooperative orientation, and that procedural justice influences cooperative orientation to the degree that they impact upon people's identities.

Cooperation has been widely recognised as being important within all types of groups. Communities depend upon the cooperation of their members, political and legal authorities have difficulty maintaining social order without the help of community members; and work organisations are less effective and efficient without willing worker cooperation (Tyler & Blader, 2000). Such cooperation is, however, often difficult to obtain because it is frequently not in the self-interest of group members to voluntarily cooperate.

Cooperation for non-instrumental reasons such as identification matters a great deal during times of crisis or transition. When the government is in the midst of a difficult period it cannot promise rewards, nor can it credibly threaten punishment for rule-breaking. At such a time the ability of government to be effective depends upon having alternative means for motivating people to comply with rules. This analysis suggests that identification is one such means. Hence, the focus on a transitional society suggests a further value of procedural justice, its ability to motivate rule-following by motivating identification with the government.

The importance of identity in shaping cooperation points to a way to overcome the obstacles to obtaining desirable cooperation in groups. If people merge their sense of self with the group, they think of issues that matter to group success as being linked to their own psychological status and wellbeing. Hence, they care about issues that influence the group for identity reasons. Together with the finding that procedural justice shapes identity, these findings point to the direction that groups ought to take to motivate their members. They ought to create rules and procedures that are viewed by members as just, and make decisions that are viewed as fair.