This paper contributes to the literature on procedural justice by applying it to a new context—ombuds services. The ombuds, an established form of alternative dispute resolution (ADR), is an important pathway to redress outside of the national court system in many countries around the world. Despite being a well-established institution in the public sector, the ombuds has only developed in the private sector over the past 20 years. As such, the ombuds landscape is an understudied area. In particular, little is known about people's expectations of—and satisfaction with—ombuds services; in fact, before now no comparable data across different services has been available. This paper contributes toward filling this void with what is, to our knowledge, a unique dataset—a survey of recent users of ombuds services. Through this data, we are able to advance understanding of people's attitudes toward this form of ADR and explore the factors that shape acceptance of the outcomes ombuds provide. Using Tyler's procedural justice model as a lens through which to view the data, our analysis provides valuable insight not only for building and expanding the existing procedural justice literature, but also for informing policy decisions in this area.
Justice systems vary considerably and each jurisdiction has its own approach to defining how people can legally resolve disputes. This provides both challenges and opportunities for access to justice. The concept of access to justice has a number of nuances; however, on a principal level the aim is to ensure effective access to an independent dispute resolution mechanism. In this context, ADR is an additional pathway to resolve disputes without invoking formal processes of law, and the ombuds is the most established model of ADR in Europe. Indeed, the institution has become increasingly popular over recent decades, to the point where it is now a highly significant and permanent feature of legal systems in many parts of the world, with many countries having both public sector and private sector ombuds. European Union (EU) level requirements for member states to have ADR bodies in place to ensure consumer protection (ADR directiveFootnote 1 and ODR regulationFootnote 2) are a high priority. The ombuds landscape throughout EU member states presents a variety of institutional and jurisdictional arrangements, operational styles and decision-making processes (Reference CreutzfeldtCreutzfeldt 2013; Reference Hodges, Benöhr and Creutzfeldt-BandaHodges, Benöhr, and Creutzfeldt-Banda 2012; Reference Hodges and CreutzfeldtHodges and Creutzfeldt 2013), though in this paper we concentrate on just one set of services, those operating in the UK.
Despite the significance of ombuds to European constitutional and civil justice landscapes, little is known about users’ perceptions of the fairness of procedures and the significance of these perceptions for legitimacy and decision acceptance within this context. Research into these issues is particularly timely due to the current emphasis within UK government policy on saving money in the administration of justice—specifically by avoiding the courts and focusing on alternative pathways toward dispute resolution. Further, the UK government plans to merge existing public sector ombuds into a unified ombudsman model,Footnote 3 which makes this research highly relevant for both public and private sector ombuds. In this paper, we concentrate on the question of decision acceptance, exploring what motivates people to accept the decisions of ombuds.
Ombuds Services in the UK
Ombuds deal with complaints from ordinary citizens and consumers about most public bodies, and the provision of goods and services in the private sector (e.g., telecoms, financial services, energy). The underlying purpose behind the creation of ombuds services was an attempt to even out the power imbalance between the individual and the state, and between the individual and a business. The services provided by ombuds are free of charge to citizens, meaning that ombuds are accessible to individuals who could not afford a court case. Different ombuds have different mandates and, within those mandates, different powers. Crucially, they are impartial and independent from the public body or business under their jurisdiction. Public sector ombuds are funded through public funds, while private sector ombuds are funded through a combination of public funds and businesses’ membership fees. Typically, they deal with individual cases but can also undertake investigations into multiple complaints about the same problem. Ideally, ombuds aim to find solutions to complaints without having to resort to formal investigations (public sector) or recommendations (private sector).
A private sector ombuds has the power to issue binding decisions on businesses under their jurisdiction (i.e., their members). Public ombuds, conversely, have only the power to make nonbinding recommendations to the public bodies complained about. Despite this difference, it has been noted that “in essence an industry ombudsman pursues many of the same objectives as their pubic law counterpart and is subject to many of the same advantages and criticisms” (Reference StuhmckeStuhmcke 1998). Both public and private ombuds provide a dispute resolution service that is made up of different stages, and an ombuds can provide a range of ADR techniques to help the consumer and the business, or the citizen and the public body, to come to an agreement (Reference CreutzfeldtCreutzfeldt 2013). The aim of an ombuds process is to provide an accessible, fast, free, and fair process for the consumer to get redress. Most cases are resolved at an early stage through intervention by the ombuds between the individual and the business (private sector) or a government agency (public sector). As noted though, in some cases the ombuds goes beyond the negotiator or mediator role, and can issue binding decisions. Anecdotally, the ombuds report a high compliance rate with their recommendations from companies or public bodies (even when they are not binding). There are a number of advantages to the ombuds model: the consumer is provided with expert information and advice about his/her rights, the company or public body can learn from the complaints and improve their services, and as the ombuds work closely with regulators they can inform them of systemic problems. If the consumer is not happy with the outcome of the ombuds procedure, the option to take their case to a court is open (ECHR Art6). In other words, the dispute can continue in another forum.
Overall, ombuds services in the EU attract millions of claims per year by addressing those grievances that citizens have against government and state institutions and those that consumers have in relation to specific businesses (e.g., banks and energy companies). In the UK and Ireland alone, there are 34 ombuds across public and private sectors (Ombudsman Association membersFootnote 4). Their workloads vary significantly. There were 16,341 complaints and enquiries in relation to the provision of NHS health services in 2012/13 (Parliamentary and Health Service Ombudsman, 2012/13), compared with over 1 million cases and enquiries in relation to financial services (Financial Ombudsman Services, 2012/13). Table 1 provides an overview of the ombuds services studied in this paper.
Table 1. Overview of UK Ombuds and Their Workload in 2014
A central distinction between public and private sector ombuds is the nature of the complaints they deal with (see Table 1). This is an important aspect to take into consideration when evaluating user interaction with and acceptance of ombuds decisions. An individual who approaches an ombuds who looks after a private sector concern, such as telecoms, would seem, prima facie, to be less likely to have an emotional engagement with the issues concerned than users of public sectors services, which deal with political, governmental, and health-related issues (e.g., the overwhelming majority of UK citizens use the state-funded National Health Service as their primary—and in most cases sole—healthcare provider, and thus have direct personal and often emotional links with this institution). Typically, complaints raised to a public sector ombuds are complex, involving longstanding issues within which many parties are entangled. A further distinction is that users of a public sector ombuds may feel a sense of entitlement in relation to a service financed by taxpayers (i.e., themselves) as compared to one funded by private industry (i.e., a private sector ombuds). Lastly, the ombuds model has been an established part of the public sector's complaints machinery for many years, so people have a clearer set of expectations, while private sector ADR is now regulated through EU wide measures.
This paper is based on the dataset collected for a study: Trusting the middle—man: impact and legitimacy of ombudsmen in Europe. The research objective of this three-year project (2013–2016)Footnote 5 was to compare levels of engagement and trust in ombuds systems in France, Germany, and the UK. With that objective in mind, research questions aimed to elicit data that would show (1) what citizens expect from, and how they engage with, the ombuds complaints processes in each country; (2) the extent to which ombuds explain their decisions and thereby engage in the process.
This research will improve understanding of how the ombuds process is perceived by its users. The research objective is achieved by collecting qualitative and quantitative data in a cross-country and cross-sector comparison. This project provides important data and creates a knowledge base, which will be directly relevant to the development of policies at national and EU level, and inform multiple networks of policy making.
The project's methodology has been primarily based on collecting data by sending out surveys to users of public and private ombuds through the ombuds themselves. The total number of responses from the three countries is just over 3000. For the purpose of this paper, the UK dataset (1306 responses) is analyzed. People have various ways to access an ombuds: These include the telephone, letter/fax, email and the Internet, through a third party, and, in some cases, face-to-face. The main channels of correspondence with an ombuds, as reported by UK respondents, were email (47 percent) and phone (31 percent), with 6.5 percent choosing letter or fax. The majority of the sample respondents stated that they were very/fairly happy with the method of communication. The reported expected duration of a case was between 1 and 3 months, with the actual duration of a case longer than 3 months for two ombuds in the sample.
This paper contributes to the existing body of literature on procedural justice and expands it to the ombuds institution. Little is known about how this group of ombuds users perceives the processes by which decisions are made and what drives their decision acceptance. For clarity, in this context we are talking about the acceptance of the substantive outcome rather than acceptance of the ombuds procedure itself. Here our data allows for a comparison between public and private ombuds. Further, we take a closer look into the significance of users’ expectations and the emotional engagement with private and public ombuds. These distinctions between public and private ombuds are made to explore our hypothesis on procedural justice effects.
Procedural Justice
The theory of procedural justice seeks to explain what motivates people to cooperate and comply with authorities in group settings. The body of literature that has grown up around procedural justice originated in experimental research in social psychology, which investigated the influence of decision evaluation on outcome acceptance (Reference Thibaut and WalkerThibaut and Walker 1975). Adding concepts from sociology, economics, and political science, Reference TylerTyler (1990) produced the now widely accepted argument that people do not only comply with laws and regulations through fear of punishment or self-interested motives. Rather, the legitimacy of legal authorities also has a role to play, and those who consider the police and courts more legitimate are, for a number of reasons, more likely to obey the law (Reference JacksonJackson et al. 2012). Procedural justice theory has been tested most comprehensively in policing (Reference Bradford, Jackson and HoughBradford, Jackson, and Hough 2013; Tyler and Folger 1980; Reference Tyler and HuoTyler and Huo 2002), courts (Reference LindLind et al. 1980), and organizational settings (Reference Brockner and WiesenfeldBrockner and Wiesenfeld 1996; Reference Greenberg, Folger and PaulusGreenberg and Folger 1983; Reference HofstedeHofstede 1980; Reference Tyler and BladerTyler and Blader 2003; 2009).
The main focus of this work has been to discover what criteria people use to assess the fairness of a procedure, and why they comply willingly with authorities (and with the law) and accept their decisions (Reference TylerTyler 1988, Reference Tyler1990). A key claim is that compliance can be explained, in part, by incorporation of the values of law into people's own value systems. This process is encouraged by the experience of procedural justice at the hands of authority figures who represent and embody those laws: police officers, judges, and other legal actors. Procedural justice is demonstrated and communicated by fair, honest, and transparent decision-making, and decent, respectful interpersonal interactions that allow people a voice in the processes affecting them.
It is important to note, however, that people's values of law are formed not only through their direct experience with an authority, but also through longstanding and conditioned learning. As such, we all have a concept of a justice system, of which the courts and police are a significant part. Furthermore, we all have an understanding and acceptance of these institutions as representative of legal authority that corresponds with a set of beliefs about how they should behave (Reference Tapp and KohlbergTapp and Kohlberg 1971). Processes of legal socialization develop a set of beliefs or expectations concerning fairness, dignity, and respect, and the importance of legal authorities keeping within well-established boundaries of probity and legality (Reference Trinkner and CohnTrinkner and Cohn 2014). These beliefs or expectations condition the way the behavior of legal authorities is experienced and interpreted and may explain, in part, why people place so much emphasis on procedural justice during their interactions with legal authorities.
Two further concepts provide important “bridging” mechanisms between procedural justice and compliance, decision acceptance, and other “outcomes” of procedural justice. First, the legitimacy of a legal authority is enhanced when it acts in a procedurally just manner; when people hold a legal authority legitimate they are motivated to follow the norms and laws it mandates and enforces, and to abide by its decisions (Reference JacksonJackson et al. 2012; Reference TylerTyler 1990). Second, when an authority figure (e.g., a police officer) treats a subordinate (i.e., ordinary citizen) in a procedurally fair manner, they encourage in that subordinate a feeling of shared group membership and sense of affiliation with the superordinate group the authority represents. This in turn promotes pro-social behavior such as cooperation and compliance (Reference BradfordBradford 2014; Reference Murphy, Bradford and JacksonMurphy, Bradford, and Jackson 2016).
Procedural Justice Research in the Dispute Resolution Setting
Much procedural justice research has focused on translating the procedural justice model developed by Tyler and colleagues to different social contexts, regulatory regimes, and countries. In the context of dispute resolution, Reference Thibaut and WalkerThibaut and Walker (1975) noted long ago that litigants’ satisfaction with decisions is influenced by their judgments about the fairness of the dispute resolution process. This has been widely supported by subsequent studies in different settings (Reference Lind and EarleyLind and Earley 1992; Reference Lind and TylerLind and Tyler 1988): on legal trial procedures (Reference LindLind et al. 1980), nontrial procedures (Reference Casper, Tyler and FisherCasper, Tyler and Fisher 1988), in mediation (Reference LindLind et al. 1990), and in organizational (Reference Brockner and WiesenfeldBrockner and Wiesenfeld 1996; Reference Greenberg and TylerGreenberg and Tyler 1987), political (Reference Tyler and CaineTyler and Caine 1981), interpersonal (Barrett-Howard), and educational settings (Reference Tyler and CaineTyler and Caine 1981). There is a widespread suggestion that those affected by the decisions of third parties in both formal and informal contexts react to the procedural justice of the decision-making process at least as much, and often more, than they react to the decision itself (Reference Lind and TylerLind and Tyler 1988; Reference TylerTyler 1996).
Moreover, despite the fact that ADR procedures are substantially different from police and court procedures (i.e., inquisitorial by nature; aiming for a settlement as outcome rather than a judgment; not bound by formal rules; and more flexible and informal than many criminal justice procedures (Reference Bercovitch and HoustonBercovitch and Houston 1985)), research has shown that ADR “disputants [who] have feelings of control and fairness, perceive the procedures and solutions to have greater legitimacy, and are more likely to comply with the terms of the conflict resolution decision” (Reference WelshWelsh 2002).
When trying to understand people's perceptions of an ombuds procedure then, fairness and decision acceptance are likely to be significant indicators. People are responsive to evaluations of the fairness of procedures, even if authorities do not provide the outcomes they hoped for (Reference Tyler and HuoTyler and Huo 2002), and there is empirical evidence that people will accept an adverse outcome if the procedure was experienced as fair (Reference Lind and TylerLind and Tyler 1988). This, at least in part, is because the experience of fair process enhances the legitimacy of the authority involved (Reference Tyler and BladerTyler 2003). Following Tyler's model, we therefore assume in this paper that procedural justice will prove important and be applicable to the context dispute resolution settings that involve ombuds. If people perceive ombuds procedures to be fair, this will build the legitimacy of the institution and motivate decision acceptance.
Several questions arise from this basic proposition, however. First, will the association between procedural justice, legitimacy, and outcomes such as decision acceptance be as strong in this context as it often seems to be in relation to the police and courts? The procedural justice model focuses on noninstrumental motivators of compliance and decision acceptance, and suggests that procedural justice is more important than outcome favorability for improving perceptions of legitimacy. A key claim in terms of decision acceptance is that people care, on average, more about the way a decision was reached than about its substantive content, and, moreover, that they are more likely to accept even unfavorable decisions if they feel they were reached in a procedurally fair way. As noted above, the extent to which people have, due to processes of legal socialization, internalized values of fairness may be a key factor in explaining these patterns—people expect legal authorities to behave fairly and are intensely attuned to behaviors that communicate fairness (or unfairness).
To turn to the subject of this paper, however, this understanding of the nature and role of the legal system, and the processes of socialization that inculcate it, may not apply to an ombuds, even though the model has been a longstanding part of justice systems around the world. There is generally a low level of awareness and use of the ombuds system (even in the public sector)—people do not necessarily know who ombuds are or what they do, and they may lack a well-developed sense of how ombuds should behave or the outcomes they can work to secure. Although our empirical data has little to say one way or the other in relation to this claim, we speculate that this is a likely result of a lack of familiarity with the ombuds model.
This raises questions as to where ombuds are situated within people's value systems, and whether norms of procedural justice figure as prominently in people's evaluations of ombuds behavior as they do in, for example, their evaluations of the behavior of police officers. On a more fundamental level, the type of procedure under consideration may have an impact on people's perceptions of its fairness, and on which aspects of the process figure most prominently in their assessments of it (Reference Leung and LindLeung and Lind 1986; Reference Lind and EarleyLind and Earley 1992). It may simply be that people conceive of their interactions with ombuds in a different way to their interactions with more established legal authorities, such as police officers and judges, who, not coincidently, wield far greater symbolic and literal power over them. We might speculate, then, that the modal interaction between members of the public and an ombuds service will be more transactional in nature than that between a citizen and a police officer or court official. Many people who contact ombuds services wish simply to get a refund from a goods or service provider they feel has not lived up to expectations (see Table 2 below). Given this, might the legitimacy of the ombuds have a different set of preconditions, given that people may approach the ombuds not for justice but for a purely instrument instrumental return (“getting their money back”). Studies within the procedural justice literature regularly contrast instrumental against affective or relational “drivers” of legitimacy, decision acceptance and other outcomes, and most conclude that affective, process-based concerns are, on average, more important. However, will this be the case in the current context where many contacts with ombuds will have been generated directly by the experience of financial loss? Might people in this context be more concerned with the outcome they receive than with the process by which it was determined?
Table 2. Variables Used in the Analysis
Second, recall that on most accounts the reason why procedural justice is so important to people is that it communicates shared group membership. Police officers, judges, and other legal authorities (and also employers) represent social categories that many people find important, and by their behavior these authority figures can communicate messages either of inclusion, status, and belonging, or of exclusion and denigration, within these social categories. Since most people desire positive messages concerning status and belonging—which help provide for and reinforce their sense of self—they are highly attentive to the identity-relevant aspects of others’ behavior, and particularly if the “other” is a figure of authority. Indeed, it is partly because procedural justice strengthens the social bonds between authorities and subordinates that it is thought to be so closely linked to legitimacy, cooperation, decision acceptance and compliance (e.g., Reference BradfordBradford 2014; Reference Tyler and BladerTyler and Blader 2000). The question arises, therefore, as to whether ombuds represent a social category with which people feel an association, and which they feel is important. If ombuds do not have this representational quality then the association between procedural justice and legitimacy, and the outcomes of that association (such as decision acceptance) may be weaker than is commonly found elsewhere. A further complication is that people may draw a distinction between the public and private services. It does not seem unreasonable to suggest, for example, that public ombuds services might be more closely linked to the state and thus superordinate identities of the kind also associated with the police and courts. Private ombuds, however, may lack such associations since, for example, they are not funded by taxation, in which case the fairness of their activity may be less identity-relevant to users of their services.
Third, by definition all those using ombuds services felt strongly enough about their situation to make contact in order to seek some kind of restitution, whether financial, an apology, or the opportunity to have their voice heard in relation to their complaint. Although, as noted above, people's emotional engagement with a complaint brought to a public ombuds is often far higher than a private one, users of both public and private ombuds are likely to enter into the interaction expecting a particular outcome to transpire. However, these expectations concern a situation and procedure about which the individual knows relatively little (due to the general lack of knowledge of ombuds services). By contrast many people who have contact with the police and even courts—for example witnesses, those contacted for information more generally, those interacting with police during crowd events or simply during their day-to-day lives—may have less at stake in the outcome of the encounter or procedure, may well not have initiated it themselves, and may not expect a particular outcome. Others, of course—notably victims, those making various forms of complaint, and offenders—have much at stake and may expect a particular outcome. What all will have, however, is a strong set of expectations about how they should be treated by legal authorities, derived from the process of legal socialization outlined above.
That people using ombuds may have a clear outcome expectation but an unclear process expectation has two potential implications. On the one hand, outcome expectation may moderate the association between procedural justice and outcome acceptance. Most obviously, those who simply want their money back may place less weight on issues of procedural justice (i.e., care less about the quality of the process than its outcome); conversely, those who wished for an apology or the opportunity to have their voice heard may place more weight on issues of procedural justice, precisely because these give them voice.Footnote 6 There is evidence in support of the latter possibility; for example, Reference van den Bos, Vermunt and Wilkevan den Bos, Vermunt, and Wilke (1997) found that people who expected to have their voice heard at the beginning of a process placed more emphasis on this aspect of procedural justice than those who did not expect to have voice in the process.Footnote 7
On the other hand, it may be that in the ombuds context procedural justice and outcome favorability are closely intertwined. Much research in other contexts suggests that perceptions of procedural justice tend to predict perceptions of outcomes (Reference Lind and TylerLind and Tyler 1988). That is, people are inclined to believe their outcome was fair and/or favorable if they feel the process by which it was arrived at was fair. This process may be particularly important in the context of ombuds; indeed, for those using the service to “have their say” process and outcome may collapse into one another (since “fair” and “favorable” outcomes are likely to be largely indistinguishable, at least from the viewpoint of participants). Yet, equally, the unfamiliarity of the ombuds service may mean that outcome favorability predicts perceptions of procedural justice to a greater extent than in other contexts. When people are unsure about how a process will function they may use the perceived fairness of the outcome to assess the fairness of the procedure (Reference van den Bos, Vermunt and Wilkevan den Bos, Vermunt, and Wilke 1997). In other words, there may be a particularly strong bidirectional relationship between process and outcome fairness in the context of ombuds services that would imply that perceptions of procedural justice do not only predict perceptions of the outcome, but that perceptions of the procedure are also shaped by perceptions of the outcome achieved.
Why Does Procedural Justice Motivate Outcome Acceptance—The Role of Legitimacy
Tyler differentiates between instrumental and normative models of cooperation, decision-acceptance and compliance with the law (Reference TylerTyler 2011). This paper takes both models into consideration. The instrumental model is premised on the idea that individuals respond to legal and administrative processes in a self-interested way, such that decision acceptance (for example) is motivated by reward (in our case a monetary procedural outcome). The normative model of process-based regulation is premised on concepts of personal morality and legitimacy.Footnote 8 According to this approach (most commonly described in the context of policing), the fairness of procedures that people experience have both immediate and long-term effects on their behaviors and perceptions. In particular, evaluations of the legitimacy of an institution are shaped by perceptions of the fairness of its procedures—more specifically, the quality of interpersonal treatment and the quality of decision making. Tyler has persuasively argued that when people feel that an institution is legitimate they are more likely to accept its decisions not because they stand to gain from them but because they think that accepting the decisions of legitimate authorities is the right thing to do. Following this argument, legitimacy is won and lost partly through the experience of procedural justice and injustice, and this process can have important implications for outcomes such as decision acceptance. Moreover, while instrumental concerns, for example about the effectiveness of an institution, are also likely to be important in shaping legitimacy and various outcomes, primacy is usually accorded to procedural justice.
If people feel that the institution is legitimate, then, they are more inclined to feel an obligation toward it and a responsibility to cooperate with it. They are also more likely to accept its decisions, not least because these are experienced as “right and proper” when handed down by a legitimate authority (Reference Tyler and RasinskiTyler and Rasinski 1991; Reference Tyler and BladerTyler 2003; but see also Reference Gibson, Caldeira and SpenceGibson, Caldeira, and Spence 2003). This suggests that legitimacy as well as procedural justice needs to be taken into account in any study examining people's responses to the decisions handed down by legal and quasilegal authorities. There is, however, significant debate in criminological (and particularly policing) literature on how legitimacy should be conceptualized and measured (Reference Bottoms and TankebeBottoms and Tankebe 2012; Reference GauGau 2011; Reference Jackson, Gau, Shockley, Neal, Pytlik Zillig and BornsteinJackson and Gau 2015; Reference Riesing, Bratton and GertzRiesing et al. 2007; Reference TankebeTankebe 2013; Reference Tyler and JacksonTyler and Jackson 2013). There are two important strands to this debate.
The first strand concerns the conceptual content of legitimacy. While most current work concurs that legitimacy is a multi-dimensional construct, there is disagreement as to what these different dimensions are. Most conceptualizations start with perceived duty to obey, envisioning that an authority can be considered legitimate when those subordinate to it willingly obey and defer to it, where such obedience is experienced by individuals as stemming from a sense of moral duty toward the authority (Reference JacksonJackson et al. 2012; Reference Tyler and JacksonTyler and Jackson 2014). To this “core” tends to be added either (a) institutional trust (such that legitimacy can be considered to be present when and to the extent that people believe an authority wields its power appropriately and effectively); and/or (b) “normative alignment,” that is, people's sense that the authority operates according to a moral framework they themselves share. On this account, legitimacy is granted or withheld on the basis of judgments people make about the alignment between their own norms and values and those of the authority concerned (see Reference Jackson, Gau, Shockley, Neal, Pytlik Zillig and BornsteinJackson and Gau 2015, for a review of these issues).
By contrast Reference TankebeTankebe (2013) has recently argued that duty to obey is an outcome of legitimacy rather than constitutive of it, and that the legitimacy of the police, at least, is comprised of public perceptions of procedural fairness, distributive fairness (or outcome justice), effectiveness, and lawfulness (or legality). This latter approach, however, can be criticized for conflating the pre-conditions of legitimacy—of which all four elements just listed are important components—with legitimacy itself, undermining conceptual clarity, and diminishing the explanatory power of the concept (Reference Hough, Jackson, Bradford, Tankebe and LieblingHough, Jackson, and Bradford 2013).
The second strand of debate revolves around issues of measurement. While research has conceptualized and measured procedural justice and legitimacy as separate constructs, some recent studies have pointed out that the scales and measures often used can suffer from poor convergent and particularly discriminant validity (Reference GauGau 2011, Reference Gau2014; Reference Jackson, Bornstein and TomkinsJackson 2015; Reference Jackson, Gau, Shockley, Neal, Pytlik Zillig and BornsteinJackson and Gau 2015; Reference Riesing, Bratton and GertzRiesing et al. 2007). These issues seem most acute when institutional trust is included in legitimacy scales and, briefly, it seems that procedural justice and institutional trust are often so highly correlated with each other it can make little sense to think of them as separate “things” (distinct psychological constructs). To the extent that this is true, and when measures of institutional trust are included in legitimacy scales, it becomes difficult to claim procedural justice “causes” legitimacy, since both seem to be part of the same construct. This compromises an important link in the chain leading from procedural justice to legitimacy and on, for example, to decision acceptance.
In this paper, we conceptualize legitimacy via a composite of this recent work within criminology. Specifically, we follow others in drawing on the work of Reference BeethamDavid Beetham (1991) and Reference JacksonJackson et al. (2012) to suggest that, in the context of ombuds services, legitimacy can be considered as having three distinct components. We suggest that an ombuds can be considered legitimate when users of the service feel (1) a moral obligation to obey the decisions of this authority; (2) that it is acting in a lawful manner; and (3) that it operates according to a normative framework they themselves share. This conceptualization, therefore, taps into the idea that institutional legitimacy is the right to rule and the recognition by the ruled of that right (Reference Bottoms and TankebeBottoms and Tankebe 2012; Reference CoicaudCoicaud 2002; Reference TylerTyler 2006). Legitimacy can be said to be present in a relationship of governance when: (1) The “governed” offer their willing consent to defer to the authority concerned, where such consent is grounded on (2) the authority's conformity to standards of legality (acting according to the law) and (3) on a degree of “normative alignment” between power holders and the governed, reflecting shared moral values.
Hypothesis
The discussion above can be distilled into a set of hypotheses to be tested against data from a survey of recent users of ombuds services. We focus on the issue of decision acceptance, and ask what motivates people to accept the decision made by the ombuds in their case. Are procedural concerns about the fairness of the ombuds’ actions uppermost in people's minds, as the procedural justice model would suggest, or are users of these services more concerned with the instrumental outcome they receive? Hypotheses 1 and 2 relate to the “standard” procedural justice model (see e.g., Reference Tyler and HuoTyler and Huo 2002: 82 ff), while the remainder relate to issues raised by applying this model to a new context.
Hypothesis 1 is that perceptions of procedural justice are associated with decision acceptance. Those who experience the ombuds as more procedurally fair will be more likely to accept the decision reached in their case.
Hypothesis 2 is that part of the reason why procedural justice promotes decision acceptance is that it enhances the legitimacy of the ombuds.
Hypothesis 3 is that outcome favorability shapes decision acceptance—those who receive a favorable outcome will be more likely to accept the decision than those who do not.
Hypothesis 4 posits that the importance of procedural justice as a predictor of decision acceptance varies according to what people expect from the process. For example, procedural fairness will be more important to those who wanted to have their voice heard (hypothesis 4a), and less important to those who wanted reward or refund (hypothesis 4b).
Hypothesis 5 is that procedural justice will be more strongly associated with decision acceptance among users of public ombuds services. Because public ombuds more clearly represent a superordinate social category to which most people feel some sort of affiliation (the state), the identity-relevant aspects of their behavior should be more salient to users of their services.
Finally, Hypothesis 6 is that in this particular context the association between perceptions of procedural justice and process outcome will be stronger than is usually the case in more overtly criminal justice contexts. Respondents who receive a favorable outcome will be more likely to think the process fair, and vice versa.
Methods and Data
The data for this study are taken from a larger dataset of just under 3000 recent users of ombuds, in several sectors, across three countries. We designed a survey and structured it according to shared stages in a consumer's complaint journey through an ombuds system: initial contact, the procedure, the outcome, the overall experience. We included measures of procedural justice and legitimacy through which we could elicit data about outcome acceptance.
The focus of this paper, the UK dataset, is made up of responses from recent users of two public sector ombuds (PHSO,Footnote 9 LGOFootnote 10) and three consumer ombuds (OS,Footnote 11 LeO,Footnote 12 FOSFootnote 13). The results of 1712 surveys provide the data for the analysis (representing a response rate of 17 percent). Unresolved cases, and those from a smaller service that returned very few responses (four overall) were excluded from the analysis, resulting in a final sample size of 1306. The sample used for this paper consists of 67 percent male respondents and 33 percent female, the average age of respondents was 59 years old. Some 41 percent were employed, 34 percent retired and 65 percent had a degree.
The people in the survey made judgments about procedural justice and other factors after they had been exposed to the ombuds procedure. Outcome acceptance and predictors of procedural justice are studied in the public and private ombuds context.
Constructs and Measures
Response Variable
The primary response variable represented respondents’ willingness to accept the outcome of their case. A single item measure was used, taken from a survey question that asked “Were you willing to accept the outcome?” Responses were on a five-category, Likert-type scale, ranging from “Very willing” to “Very unwilling.” Since we are concerned with what predicts acceptance of the outcomes provided by the ombuds services, responses on this item were collapsed to generate a new binary variable, outcome acceptance, coded one if the respondent was “very” or “fairly willing” to accept the outcome and zero if they were “neither willing or unwilling,” “fairly unwilling,” or “very unwilling.” What is at stake here, then, is the subjective state of mind of the respondent – when asked, did they feel that they were willing to accept the decision reached?
Explanatory Variables
Explanatory variables included four scales generated and validated via Confirmatory Factor Analysis (CFA) in the statistical package Mplus 7.11. These represented, first, the extent to which the respondent contacted the ombuds service because they wanted fairness from the process. This construct was measured by survey items that tapped into whether respondents decided to complain to the ombuds because they wanted an apology, someone to listen to them and/or to be treated with dignity and respect (all used a five category Likert-type response scale).
The second scale represented respondents’ perceptions of the procedural fairness of the ombud's staff with whom they first had contact (this will be an important explanatory variable when we come to address hypothesis 6). Items used here covered issues such as helpfulness, understanding, respect and neutrality; all used binary yes/no response categories.
The third scale measured respondents’ perceptions of the procedural justice of the ombuds service as a whole, with items tapping into issues of voice, neutrality, respect and trustworthiness, all on the same five category Likert-type response scale.
The fourth scale represented respondents’ assessment of the legitimacy of the ombuds service, with three items covering perceived duty to comply with the ombuds decision, normative alignment, and legality (all, again, using five category Likert-type response scales).
Full question wordings from the items used in these measures, and results from the CFA modeling, are shown in Table 3. Three different models are shown—note that full information likelihood estimation was used in each case, meaning that cases with some missing data were included in the analysis. Indicators were set to categorical throughout.
Table 3. Confirmatory Factor Analysis: Latent Constructs and Measures
The shaded areas in the columns indicate where items have been allowed to load onto one factor.
Model 1 in Table 3 shows results from a four-factor solution with no cross-loadings. The approximate fit statistics indicate a good fit to the data (Reference Hu and BentlerHu and Bentler 1999). There were, however, two issues with this model. First, the latent variables representing overall perceptions of procedural justice and legitimacy were very strongly correlated (r = 0.95), suggesting poor discriminant validity—it seems that the survey respondents did not really draw a distinction between the procedural justice and legitimacy of the ombuds service. Accordingly, Model 2 in Table 3 shows results from a three-factor solution (again with no cross-loadings), where all the ombuds’ procedural justice and legitimacy items were allowed to load onto the same factor (see the shaded section of the table). The approximate fit statistics again indicate a good fit to the data, although the change in X 2 values indicate a somewhat less well fitting model when compared with model 1. As such, it difficult to claim that the procedural justice and legitimacy of ombuds are separate constructs, in this dataset at least, and in any case they are so highly correlated it makes little sense to proceed while treating them as different from one another. To retain conceptual and analytic clarity we proceed simply with the measure of procedural justice, and we return to this issue in the conclusion.
The second issue with Model 1 in Table 3 is with the measures of staff and ombuds procedural justice, which were also highly correlated (r = 0.82). Model 3 in Table 3, therefore, shows another three-factor solution, where this time all the procedural justice items were allowed to load onto one factor. In this case, the resulting model is clearly a worse fit to the data than Model 1—the X 2 change is very large and the RMSEA value is 0.1, well above the standard cut-off point of 0.06. We, therefore, proceed with the two distinct measures of procedural justice, and factor scores were extracted from Model 1 and saved for further analysis.Footnote 14
The remaining explanatory variables were all single item measures. First was a measure of outcome favorability, derived from an item that asked “Was the outcome of your case in your favor?” to which there were four possible responses: “yes,” “no,” “partially,” and “not sure” (recall that cases where no outcome had been reached—or where respondents were “not sure” of the outcome—were excluded from the analysis). Outcome favorability was therefore represented in the model by a pair of dummy variables representing positive and negative responses, with “partial” responses as the reference category. Second, to complement the scale representing the extent to which respondents sought apology and restoration from the ombuds, a further dummy variable covered the extent to which they wanted “a financial award and/or my money back.” Responses to this item—wanted reward—were on a 5-point scale (1 most important to 5 least important); responses of one or two were coded as one on the new dummy variable, with response in the remaining three categories comprising the reference category.
Control Variables
We also included a number of control measures in the models (all as dummy variables). These represent measures of potential importance to decision acceptance which may confound some of the explanatory variables described above, but for which we did not formulate any particular hypotheses. First, a further measure of respondents’ views of the process with which they were dealt was an item that assessed the extent to which they felt had control over the outcome—process control has been found to be an important correlate of procedural justice (e.g., Reference Tyler and BladerTyler and Blader 2000, chapter 8). Two dummy variables were used, the first coded one if the respondent felt they had control over the outcome and zero otherwise, the second coded one if the respondent did not feel they had any control, and zero otherwise (the reference category was those who were unsure in relation to this issue).
Two further variables covered aspects of respondents’ views of the outcome of their case, each entered into the models as pairs of dummy variables representing positive and negative responses with neutral/uncertain responses as the reference category. First was a measure of outcome expectation, derived from a survey item that asked “Was the outcome what you expected?” with responses on a 4-point scale (“exactly’; “close”; “not sure what I expected”; “not at all”). Second was a measure of distributive justice, derived from an item that asked “Do you think that others with a similar complaint to yours receive the same outcome from the ombuds?”; responses were on a 5-point Likert-type scale.
Finally, dummy variables for each of the individual ombuds services in the sample were also included in the models. We, therefore, estimated fixed effects models that allowed us to partial out the statistical effects of any unobserved variables, operating at the level of the different ombuds services, which might influence the outcomes of interest.
Results
Descriptive Statistics
Table 2 shows that, overall, just over half (51 percent) of respondents were willing to accept the outcome of their case, with 39 percent unwilling (10 percent respondents said they were neither willing nor unwilling to accept the outcome—i.e., were at the mid-point of the underlying 5-point scale).
Descriptive statistics for the explanatory variables are also shown in Table 2. Notably, outcome favorability was almost exactly split between those who felt they had obtained a favorable outcome and those who did not (the remainder were unsure). Similarly, while 50 percent of respondents obtained the outcome they expected, 40 percent did not (with the remainder again unsure). Other variables were more skewed—perceptions of distributive justice were positive, those of process control much less so. Obtaining a financial reward was important for just less than half of respondents (note that this variable was negatively correlated with want an apology/dignity—most respondents tended to want one or the other but not both, although there were exceptions).
Table 2 also shows descriptive statistics for the three scales used in the analysis. Note all are mean centered, and there is little suggestion of significant skew in any.
Multivariate Analysis
Since the response variable was a dichotomous measure we used binary logistic regression modeling to address hypotheses 1–5.Footnote 15 Recall that the measure of outcome acceptance was coded such that one equaled self-assessed willingness to accept: a positive regression coefficient therefore indicates that an increase in the relevant explanatory variable was, controlling for the other variables in the model, associated with a greater probability of being willing to accept the outcome determined by the ombuds. A negative regression coefficient, by contrast, indicates that an increase in the explanatory variable was associated a lower probability of being willing to accept the outcome.
Results from four binary logistic models predicting outcome acceptance are shown in Table 4. Model 1 contains main effects only. We find, first, a very strong association between the procedural fairness of the ombuds and willingness to accept her/his decision (ß = 1.66, p < 0.01). A clearer idea of the strength of this association can be gained by calculating fitted probabilities from the model.Footnote 16 For example, holding all other variables at their mean the probability of a respondent perceiving the mean level of procedural justice being willing to accept the outcome was 0.58. For a respondent perceiving a low level of procedural justice, one standard deviation below the mean, the probability of being very willing to accept the outcome was just 0.17; but for a respondent perceiving a high level of procedural justice, one standard deviation above the mean, the probability of being very willing to accept the outcome was 0.91.
Table 4. Results from Binary Logistic Regression Models Predicting Outcome Acceptance
+ p < 0.1; *p < 0.05; **p < 0.01.
Second, perceptions of the nature of the outcome were associated with acceptance of it. Most notably, outcome favorability was strongly and independently associated with acceptance: those who received a favorable outcome were more willing to accept it (ß = 1.8, p < 0.01), while those who received an unfavorable outcome were less willing (ß = −0.86, p < 0.01). Fitted probabilities can again be used to illustrate the nature this relationship. Conditioning on the other variables in the model (again holding them at their mean), the probability of someone who obtained the outcome they wanted being willing to accept it was 0.85. By contrast, the probability of being willing to accept the outcome for someone who did not get the outcome they wanted was 0.28. Equally, though, and independent of their favorability, outcomes that were not expected were less likely to be accepted than those that were.
Third, respondents who felt they had control over the outcome were no more likely to accept it, and those who felt they did not have control were no less likely to accept, than those who were unsure about this element of their experience. This finding corresponds with results from many other procedural justice studies (e.g., Lind et al. 1997; Reference Tyler and BladerTyler and Blader 2000; Reference Tyler and HuoTyler and Huo 2002), which have found control issues to have no independent statistical effect on outcome acceptance once procedural justice is taken into account.
There was no association between perceptions of distributive justice and outcome acceptance once the other variables in the model were taken into account, nor were those who wanted fairness, or financial compensation, more or less likely to accept the outcome. Finally, conditioning on the other variables in the model, there was no consistent association between the type of ombuds service involved (public/private) and decision acceptance. Respondents were less likely to be willing to accept the decisions of some ombuds compared with others, but both public and private services fell into this category.
Models 2 and 3 in Table 5 tested the interactions between reasons for complaining to the ombuds and procedural justice. Recall that we hypothesized procedural justice would be a more important predictor of outcome acceptance for those who wanted fairness from the outset, while it would be less important for those who wanted a cash reward. We find support for the former idea but not the latter. Model 2 shows that the interaction between procedural justice and “wanted fairness” (reward*PJ) was positive and significant, while the reward*PJ coefficient in Model 3 was not significant. It seems that procedural justice may have been more important for those who wanted fairness from the process, but no less important a predictor of outcome acceptance for those who wanted financial compensation.
Table 5. Results from Linear Regression Models Predicting Perceptions of Ombuds Procedural Justice (High = More)
Model 4 tested the hypothesis that procedural fairness would be a more important predictor of outcome acceptance for users of public ombuds. However, we find little evidence to support this idea. Compared with the reference category (Private 1), users of the public ombuds services included in the model did not seem to place any more emphasis on procedural justice, and there was no consistent link between the nature of the ombuds service involved (public/private) and the importance of procedural justice to its users.
When is the Procedure Experienced as Fair?
Thus far we have treated the procedural justice of the ombuds and the outcomes they are able to deliver as separate aspects or components of people's experiences. This is of course a key claim of the procedural justice model—that process and outcome are separate factors, and that a fair process can offset an unfavorable outcome. Yet, as discussed above, there are reasons to suggest that in the context of ombuds services the line between these two factors may be considerably more blurred than in the case of, for example, the courts or policing (Reference Tyler and HuoTyler and Huo 2002). We therefore estimated a second set of regression models, this time with ombuds procedural justice as the response variable—since this was an interval/ratio level measure we switched to linear regression. Explanatory variables were the same set of process and outcome concerns used above and, in addition, the measure of staff procedural justice. Results are shown in Table 4.
Model 5 in Table 5 contains just the dummies representing the various services and outcome favorability. The latter was very strongly associated with perceptions of procedural justice—those who received a favorable outcome tended to have more favorable views of ombuds procedural justice (ß = 0.41, p < 0.01), while those who received an unfavorable outcome tended to have less favorable views (ß = −0.45, p < 0.01). Note the large R 2 value for this model (0.49); nearly half the variation in perceptions of ombuds procedural justice can be explained by knowing just the ombuds involved and whether the outcome was favorable or not—and most of this was down to outcome favorability. In a model excluding the ombuds dummies (not shown) R 2 = 0.45.
Model 6 adds the other process and outcome measures. Conditioning on the other variables in the model, those who experienced little or no control over the outcome (ß = −0.34, p < 01) and who received an outcome they were not expecting (ß = −0.24, p < 01) tended to see the ombuds as less procedurally fair. However, there was no independent association between distributive justice and ombuds procedural justice. Note that in this model the coefficients for outcome favorability are much reduced in size—some of the association between this measure and procedural justice was explained by process control and expectation.
Finally, model 7 in Table 5 adds the measure of staff procedural justice. Conditioning on the other variables in the model, there was a strong association between the perceived procedural fairness of particular ombuds staff and the perceived procedural fairness of the service as a whole (ß = 0.57, p < 0.01); note also that the R 2 value jumps to 0.76. Equally, though, even taking into account the way respondents were treated by particular staff members, those who received an unfavorable outcome, an outcome they did not expect, or who felt they had no control over the process were still less likely to think the ombuds service as a whole procedurally fair.
It is important to note that we are not claiming that outcome favorability, or any of the other variables in these models, “causes” perceptions of procedural justice; it is clear these associations are all likely to be bidirectional. Most obviously, the communication between parties that generates a sense of procedural justice may shape complainants’ expectations in relation to the outcome of their case. Treating people fairly may also encourage in them a sense that they are in control of the process (Reference Tyler and BladerTyler and Blader 2000). What our results do suggest, however, is that concerns about outcome and process are heavily intertwined in the ombuds context, and that it is highly likely people infer the quality of one from the quality of the other.
Discussion
We presented six hypotheses to test if the model of procedural justice applies to the ombuds context. The focus was on understanding motivations for decision acceptance in users of the ombuds system. First, our results suggest that the “standard” procedural justice model works well in this context: procedural justice is very strongly associated with decision acceptance (Hypothesis 1). Procedural justice is also very strongly associated with legitimacy—so much so, in fact, that is unclear whether, in this context, they constitute separate constructs (separate aspects of the way respondents experienced and thought about the ombuds service they encountered). We were not therefore able to test Hypothesis 2, although it almost certainly true in a general sense that here, as elsewhere, legitimacy predicts decision acceptance (since procedural justice does). However, unlike other procedural justice studies conducted in other contexts, we find that outcome favorability is also a predictor of decision acceptance (Hypothesis 3). Support was also found for Hypothesis 4a, which posited that procedural justice would be more important to those who wanted fairness from the outset of the process; however, Hypothesis 4b, which suggested that procedural justice would be less important for those who sought reward or refund, was not supported by our analysis. We also found little support for Hypothesis 5, which posited that procedural justice should be more important to users of state-funded ombuds services.
Finally, we find that, as suggested by Hypothesis 6, there is a strong association between outcome favorability and perceptions of fairness. Controlling for the quality of respondents’ interactions with particular ombuds staff, those who received a positive outcome were significantly more likely to believe the procedure, as a whole, was fair.Footnote 17 Nevertheless an important finding from this analysis is that the quality of interaction with ombuds staff was a strong influence on individuals’ procedural and outcome perceptions.
All of this said, we can return to the broader question posed at the outset of this paper, of where the ombuds sits within people's value system, in relation to both our findings and existing procedural justice literature. Outcome favorability predicted perceptions of procedural justice to a larger extent than is often found elsewhere (e.g., in the courts and police). One interpretation of this is that unfamiliarity with ombuds services may be an important factor shaping people's experiences of them.Footnote 18 This would be in line with van den Bos, Vermunt, and Wilke's (1997) finding that if people are unsure about what to expect from a process, the perceived fairness of the outcome has larger weight in their assessments of the fairness of the process. Moreover, and again unlike much procedural justice research in other contexts, we find that other factors predict decision acceptance independent of perceptions of procedural justice—namely, receiving an unexpected outcome. Finally, and perhaps most importantly, we find very significant overlap between perceptions of procedural justice, outcome favorability and outcome expectation. These seemingly diverse aspects of the experience appear, in this context, to be mutually constitutive of one another.
Viewing these findings through the lens of the legal socialization literature may provide some valuable insights. For example, it could be that the location of ombuds services within people's concept of the justice system is uncertain. People's values of law are learned and developed over time; they are not only formed through direct interaction with an authority. As a much less visible service or institution it may be that the experiences of the ombuds are not subject to the same enduring social conditioning effects as the courts or police. Many or even most people have some form of understanding and acceptance of these representatives of legal authority, and thus some expectations of how they should behave and of the outcomes they are meant to secure—broadly, justice, and security. People may be less clear about how ombuds are meant to behave and the outcomes they are intended to secure, making them more sensitive to the multiple aspects of the process through which they interact with these services. These questions would be a fruitful area for future enquiry.
Another possibility is of course that in the ombuds context people are simply more instrumental than seems to often be the case in criminal justice contexts. That is, it may be that users of ombuds services did understand the process and knew what they wanted—which was their money back or getting the outcome they desired. Our analysis cannot distinguish whether this was in fact the case (although it seems to us prima facie believable to claim that most people are not familiar with ADR processes), and this again would be an interesting area for future exploration.
Despite this procedural justice remained arguably the most important factor predicting, in the current case, decision acceptance. In addition, the legitimacy of the ombuds appeared to collapse into perceptions of procedural justice, leaving us with little choice but to concentrate in our analysis on procedural justice alone. On this basis, it might be argued that, to the extent that people are uncertain about what ombuds services are intended to do, or find it difficult to “fit” them into their value systems, procedural justice might be the best way to foster such understanding—not least because it seems procedural justice is just as strongly associated with legitimacy here as elsewhere.
Study Limitations
We should note the limitations of our study. While the response rate was high, for a mail out/return sampling strategy, a rate of 17 percent implies our findings are unlikely to be truly generalizable to the population of users of the ombuds services involved. The fact that we used a cross-sectional survey design precludes any attempt at causal inference and means we are unable to model processes of change.
More specifically, the closeness of procedural justice and legitimacy in our data may be as much a product of the survey questions used as one of conceptual overlap. This was, as far as we know, the first study of the users of ombudsman services to field such measures, and we were, therefore, reliant on existing measures, primarily from the criminological literature. Future work might profitably attempt to refine the concept of legitimacy in this context, not least to probe whether it is as close to procedural justice as we have found here.
Our primary outcome measure was also somewhat limited in its scope. Respondents were simply asked whether they were willing to accept the outcome. To be sure, this is an important indicator, which taps into people's state of mind on hearing the outcome of their case. However, it might have been preferable to also include a behavioral indicator of acceptance, such as whether complainants sought a review of the decision or outcome.
Conclusions
This paper set out to explore whether the theory of procedural justice could explain, as it does in other contexts, why people accept decisions handed down by authorities. The authority studied, an ombuds, is an interesting hybrid in the justice system. It is a body that provides ADR outside of the national court system, an additional path to accessing justice.
We asked what motivates people to accept decisions made by an ombuds. We found that outcome favorability and procedural justice are key factors shaping decision acceptance; however, outcome favorability has a more important weighting in this context than is often the case in other studies, for example of policing. Further, we found that procedural justice and legitimacy may not be empirically separable in this context.
We conclude that there are many factors that motivate people to accept ombuds decisions. It may be that the ombuds institution has yet to establish itself in peoples’ minds as an institution similar in nature to the police and courts. While this process is not likely to ever be complete—since ombuds are by definition different to more overtly “legal” institutions—greater familiarization might lead to an increased focus on procedural justice among users of ombuds services. Equally, however, it may simply be that in this context people are more outcome-focused than seems to be the case in more overtly criminal justice settings. More research on these questions is needed. Our data did show, however, that staff procedural justice had a significant statistical effect on people's overall perceptions of fairness of the ombuds, and therefore on decision acceptance, underlining that here, as elsewhere, the fairness of a process is likely to be an important factor shaping readiness to accept its conclusion.
