CO-WORKERS’ JUSTICE JUDGMENTS, OWN JUSTICE JUDGMENTS AND EMPLOYEE COMMITMENT: A MULTI-FOCI APPROACH

————— Florence Stinglhamber is professor of organisational psychology at the Department of Psychology, Université catholique de Louvain. David De Cremer is professor of Behavioural Business Ethics at Rotterdam School of Management, Erasmus University, the Netherlands. Correspondence concerning this article should be addressed to Florence Stinglhamber, Université catholique de Louvain, Place Cardinal Mercier 10, 1348 Louvain-la-Neuve. E-mail: florence.stinglhamber@uclouvain.be CO-WORKERS’ JUSTICE JUDGMENTS, OWN JUSTICE JUDGMENTS AND EMPLOYEE COMMITMENT: A MULTI-FOCI APPROACH

Using a sample of 212 employees, we conducted a study to examine whether employees use their co-workers' fairness perceptions to generate their own justice judgments and to develop their subsequent affective commitment. The conceptual framework used to investigate these linkages is social exchange theory combined with a multiple foci approach. Results of the structural equation modeling analyses revealed that co-workers' procedural justice judgments strengthened employee's own procedural justice judgments, which in turn influenced their affective commitment to the organisation. Similarly, co-workers' interactional justice judgments increased employee's own interactional justice judgments, which in turn impacted on their affective commitment to both the supervisor and the organisation. As a whole, findings suggest that coworkers' justice judgments strengthened employee's affective attachments toward the justice sources by reinforcing employee's own justice perceptions.
Research on organisational justice has flourished in the last 30 years. Early research in this area focused on distributive justice, i.e., the perceived fairness of outcome distribution such as pay or promotion opportunities (e.g., Adams, 1965;Cropanzano & Byrne, 2000). In the late 1980s, research on fairness in the workplace shifted to the examination of procedural justice, that is the fairness of procedures used to determine the outcomes (e.g., Moorman, 1991). Finally, in the 1990s, researchers started to study the social and relational qualities of justice. In this respect, the construct of interactional justice captures the quality of the interpersonal interaction or treatment people receive as procedures are implemented (e.g., Bies & Moag, 1986).
One important line of inquiry in this research field has been directed toward increasing workplace fairness. For a long time, justice researchers focused solely on one source of information (e.g., supervisor or organisation) affecting people's fairness perceptions. However, more recently, a multiple foci approach of the justice concept has been adopted in the organisational justice literature. This line of research suggests that at least two sources of (in)justice exist: the organisation as a whole and the immediate supervisor (Rupp & Cropanzano, 2002). As a consequence, recently, these two sources have been linked to different dimensions of justice. In accordance with this view, Malatesta and Byrne (1997) suggested, for example that, because perceptions of procedural justice are based on an organisation's formal policies, the organisation can be seen as the most likely source of this form of justice, whereas perceptions of interactional justice derive mostly from interpersonal exchanges, which are most likely to be administered by supervisors.
Thus, a multi-foci perspective (Rupp & Cropanzano, 2002) holds that employees have exchanges with multiple parties and that the fair or unfair behaviour of each of these parties elicits a variety of employee work-related attitudes and behaviours toward that specific party. Indeed, some research has revealed that justice from organisations is associated with outcomes relevant to the organisation and justice from supervisors is related to outcomes relevant to the supervisor (Rupp & Cropanzano, 2002). For example, Malatesta and Byrne (1997) provided initial evidence showing that procedural justice, as lavished by the organisation, is a better predictor of organisational citizenship behaviours beneficial to the organisation (OCBO), whereas interactional justice, as lavished by the supervisor, is a better predictor of organisational citizenship behaviours beneficial to the supervisor (OCBS). In a similar vein, Masterson, Lewis, Goldman, and Taylor (2000) and Cropanzano, Prehar, and Chen (2002) found that job performance was better predicted by supervisory focused interactional justice than by organisationally focused procedural justice.
It can thus be concluded that when examining effects of justice within complex social setting such as organisations it is advised to adopt a multifoci approach. Moreover, such a multi-foci approach is also helpful in addressing a more recent issue that has emerged in the justice literature. That is, a multi-foci approach helps to obtain a more specific and better sense of how fair or unfair the social situation at hand is (because multiple parties involved are evaluated). In a similar vein, more recently, it has been argued that in order to get a better perception of how fair/unfair the situation is one also needs to look at the justice experiences of other employees. Indeed, to date, most justice studies adopt an approach in which it is examined how people personally experience (un)fair treatment. For example, most procedural justice research has examined how I react to the procedures that I experience (Kray & Lind, 2002). In this way, they neglected another important source of social information, that is, the justice of experiences of other organisational members or interaction partners.
Taken together, both a multi-foci approach and an approach in which the justice experiences of the others are taken into account may help us at understanding better how justice in organisations is actually experienced (i.e., which types of justice are experienced and shared by others) and how these perceptions, in turn, affect important organisational outcomes such as affective commitment to the organisation and its supervisors.

The justice experiences of other employees
The paucity of research examining the impact of other's justice experiences is somewhat surprising. Indeed, there is evidence in research on survivor's reactions to layoffs that people do incorporate the misfortunes of others into their own judgments of situations. This body of research shows that the experiences of those who lost their job can have substantial effects on survivor's own evaluations of procedural justice in layoff situations (Brockner & Greenberg, 1990;Brockner, Konovsky, Cooper-Schneider, Folger, Martin, & Bies, 1994). In a similar vein, research on distributive justice (in particular, on equity theory) has shown that others' outcomes serve as a comparison point in determining the fairness of own outcomes (i.e., Adams, 1965; see also Grienberger, Rutte, & van Knippenberg, 1997). In sum, preliminary support exists for the proposition that the treatment of others should also be attended to when making judgments about the fairness of one's own working environment. Therefore, the last few years, some researchers have started to examine whether procedural fairness judgments are indeed influenced by experiences of others. Lind, Kray, and Thompson (1998) showed that people displayed a strong tendency to use their own experiences to make inferences about the treatment of others, suggesting that personal experience frequently serves as the most important reference point in evaluating procedural justice. Moreover, Kray and Lind (2002) qualified this conclusion by demonstrating that when people do not personally experience any injustice, they do not attend to the treatment of others; however, if they were treated unjustly themselves, they are affected by unjust treatment of others. Further, van den Bos and Lind (2001) showed that a person's reaction was particularly influenced by social reports of another's having received treatment similar to their own, either fair or unfair. As a result, van den Bos and Lind (2001Lind ( , p. 1333 concluded that "Other-oriented justice effects (…) appear to be every bit as strong as were our self-oriented justice effects" (see also De Cremer, Stinglhamber, & Eisenberger, 2005). This conclusion has meanwhile been backed up by some more recent studies. That is, Lamertz (2002) concluded in a field study within a telecommunications company that an employee's perceptions of both procedural and interactional fairness were significantly associated with the interactional fairness perceptions of a peer. Also, Colquitt (2004) showed that in teams, the justice experiences of others influence a person's reactions in the context of how fairly they themselves are treated, but particularly so as a function of the degree of interdependence that these people experienced. Finally, De Cremer and Van Hiel (2006) showed in both experimental and field studies that people use the procedural justice experiences of their team member to regulate their emotions and actions if this team member was helpful to them in the past.
It thus seems likely that in daily conversations and social interactions people may assign weight to other's fairness experiences to assess own judgments and feelings. Indeed, in an organisational context, it appears that the experiences of others can influence employees' perceptions of organisational justice (e.g., De Cremer & Van Hiel, 2006;Lamertz, 2002;van den Bos & Lind, 2001).

A multi-foci approach to the justice experiences of others
Most of the studies examining the impact of other's justice experiences have mainly focused on the effect of others' procedural fairness judgments on own procedural justice perceptions. In the present paper, we wish to argue that a multi-foci approach can also be useful by taking other forms of justice into account. Therefore, using a multiple foci view of the justice construct, the first objective of this study is to demonstrate that both procedural and interactional justice perceptions can be socially constructed.
The assumption that an employee's perceptions of procedural justice should be influenced by the corresponding perceptions of co-workers is based on several arguments. First, employees are likely to seek cues from their co-workers to reduce the ambiguity about the meaning of formal procedures and the subsequent uncertainty about their relationship to the organisation (cf. for example, Ashford & Cummings, 1983). In addition, because each organisation has its own standards, co-workers can provide, especially to newcomers who do not know the norms of this specific company, a reference point that will allow the contextualization of the justice associated with a particular procedure (Lamertz, 2002). In sum, indications that employees may receive from their close colleagues should contribute to their understanding of how fair organisational procedures are, consequently shaping their own fairness judgments.
In a similar way, it seems plausible to assume that co-workers' perceptions of interactional justice should impact employees' corresponding judgments. Indeed, employees may seek cues from their colleagues to help them to decode their supervisor's interpersonal behaviours. In the face of an example of poor or good interpersonal treatment from his/her supervisor, an employee should then be able to understand its real meaning. Is this treat-ment due to an impression management from the supervisor, to external facts or constraints that are not under his/her control, or to a deliberate action from his/her part? Getting indications from co-workers should help an employee to respond to such a question and, as a consequence, make his/her own judgment about the justice of the treatment he/she received (Lamertz, 2002).
In general terms, due to situational features like uncertainty, concerns about status within the organisation and so forth, it seems likely that others' (procedural and interactional) justice judgments may serve as source for an employee's judgments to the extent that he/she feels uncertain about his/her own experiences or just does not have enough experience yet. Accordingly, we hypothesised in this research that people's own reactions will be determined by information concerning the interactional and procedural treatment of others. Specifically, our assumption is that an employee's perceptions of procedural justice should be affected by his/her co-workers' judgments of procedural justice, whereas the co-workers' judgments of interactional justice should be a potent predictor of his/her interactional justice perceptions. We thus suggest that informal communication channels among colleagues (i.e., among people who interact on an everyday basis in the organisational context) are primary routes through which social comparisons and social cues may affect the formation of a meaningful interpretation of organisational events and institutions, especially formal rules and norms of interpersonal treatment (Hartman & Johnson, 1989;Johanson, 2000;Lamertz, 2002).
The second goal of this paper is to link this social influence effect to the nomological network of organisational justice, and in particular to the consequences of this construct. The underlying assumption here is that socially constructed justice judgments (emerging from others) can serve as a heuristic to evaluate the relationship with the organisation and the supervisor. That is, if employees perceive relationships with the organisation and the supervisor as fair (i.e., based on the experiences of peers), they may infer that the employees (including themselves) have a good relationship with these two justice sources.
Many studies have been performed in order to identify the outcomes of organisational (in)justice, both at the attitudinal and the behavioural levels (see Colquitt, Conlon, Wesson, Poter, & Ng, 2001 for meta-analytic results). An important observation in this respect is that recent research and theory is increasingly making clear that emotional reactions play a central role in understanding the psychology of justice (see De Cremer & van den Bos, 2007;Folger & Cropanzano, 1998;Mikula, Scherer, & Athenstaedt, 1998;Weiss, Suckow, & Cropanzano, 1999). Therefore, we decided to focus in the present study on the links between justice and one of its well-known emotionally-oriented consequences, affective commitment (or AC).
AC to the organisation has been defined as an individual's identification with, involvement in, and emotional attachment to, the organisation (Allen & Meyer, 1990;Meyer & Allen, 1991. Several empirical studies and recent meta-analytic results Meyer, Stanley, Herscovitch, & Topolnytsky, 2002) reported organisational justice (and, in particular, procedural justice) and organisational AC to be strongly related. Fair treatment implies the formation of social exchange relationships and, as such, creates a sense of obligation in the employee to pay back the source of this fair treatment via work-related attitudes or behaviours that are beneficial to it (e.g., AC) (see Moorman, Blakely, & Niehoff, 1998;Pillai, Schriescheim, & Williams, 1999).
To the best of our knowledge, however, these studies on the links between justice and AC did not distinguish among the multiple foci pertaining to the workplace (for one exception, see Malatesta & Byrne, 1997). Yet, just as it is now broadly accepted that employees make the distinction between the different possible sources of justice (i.e., the organisation versus the supervisor), recent research has emphasised the value of distinguishing among multiple foci of employee commitment in the workplace to predict employee behaviour. Several studies have shown that just as employees develop affective attachments to the global organisation, they may commit themselves to foci that are internal to it, such as the supervisor (Becker, 1992;Becker & Billings, 1993;Becker, Billings, Eveleth, & Gilbert, 1996;Clugston, Howell, & Dorfman, 2000;Gregersen, 1993;Stinglhamber, Bentein, & Vandenberghe, 2002). As a consequence, AC to the supervisor can be defined as an attachment characterised by an identification and emotional attachment to the supervisor (Clugston et al., 2000).
To the extent that recent research has shown that justice from organisations is associated with outcomes relevant to the organisation and justice from supervisors is related to outcomes relevant to the supervisor (see above; Rupp & Cropanzano, 2002), we thus hypothesised that the multiple foci approach of the justice concept should also have considerable impact on the examination of AC as one of its potential consequences. In agreement with this view, procedural justice as lavished by the organisation was assumed to be a better predictor of AC to the organisation, whereas interactional justice as lavished by the supervisor was supposed to be a primary source of AC to the supervisor. Evidence has been reported that employees do engage in enduring exchanges with both the organisation as a whole and their immediate supervisor (Settoon, Bennett, & Liden, 1996;Wayne, Shore, & Liden, 1997). This suggests that, even if employees often perceive their supervisors as representatives of the organisation (Eisenberger, Huntington, Hutchison, & Sowa, 1986;Levinson, 1965), they may develop exchange relationships with them that are distinct from those that they experience with their organisation.
Before moving to the present study, we briefly summarise our hypotheses. First, we hypothesised that co-workers' procedural justice judgments would strengthen employee's affective commitment to the organisation by reinforcing employee's own procedural justice judgments (Hypothesis 1). Second, we predicted that co-workers' interactional justice will increase employee's affective commitment to the supervisor by reinforcing employee's own interactional justice judgments (Hypothesis 2). Figure 1 depicts the hypothesised linkages among variables. Whereas, to date, most research on the effect of other's justice experiences has been experimental in nature (for the only exceptions see De Cremer et al., 2005; see also Colquitt, 2004), the present research examines these two hypotheses in the setting of a field study (i.e., with high external validity).

Sample and procedure
Questionnaires were mailed to a selected sample of 470 employees from a telecommunication company. We sent prospective participants a packet including a cover letter, the measures of procedural and interactional justices for self and for co-workers, and affective commitment to the organisation and the supervisor, as part of a larger survey, and a postage-paid return envelope Figure 1 Theoretical model specifying the hypothesised structural linkages among variables that respondents could use for mailing back their completed questionnaire to the university. The cover letter explained to the employees the purpose of the study, stressed the importance of responding to the questionnaire, and provided assurances of anonymity and confidentiality. Two weeks later, we mailed follow-up letters to non respondents stressing the value of the survey and the importance of their participation.
Two hundred twelve employees returned usable questionnaires (a 45% response rate). Of this final sample, 66.5% were females. Average age was 30.18 years (SD = 9.65), average organisational tenure was 2.50 years (SD = 1.62), and average tenure with the supervisor was .87 years (SD = .79).

Measures
As the study was conducted in a Dutch-speaking context, all measures were translated from English into Dutch by a first translator and then independently back-translated into English by a second translator, following the procedure recommended by Brislin (1980). Minor discrepancies between the original English version and the back-translated version resulted in adjustment in the Dutch version based on direct discussion between the translators.

Control variables
In order to rule out the possibility that a significant relationship in our model may be explained via the concomitant effect of some demographic variables on the two variables included in this significant relationship, we controlled for tenure in the organisation and tenure with the supervisor in our analyses. That is, we introduced these two control variables as additional exogenous variables predicting own justice judgments and affective commitments to the organisation and the supervisor. These variables were assessed with single item measures that asked participants how long they had been working in the organisation and under the supervision of their current supervisor.

Own justice judgments
To measure employees' own perceptions of procedural justice, we relied on Colquitt's (2001) 7-item scale. A sample item was 'To what extent have you been able to express your views and feelings during the procedures?'. Own perceptions of interactional justice was assessed via the 6-item scale of Bies and Moag (1986) used by Colquitt (2001). One of the items was 'To what extent has your supervisor treated you in a polite manner?'.
Respondents were asked to answer to these items using a 5-point Likert-type scale with anchors of 1 = to a small extent and 5 = to a large extent.

Co-workers' justice judgments
To measure procedural justice perceived by co-workers, we adapted the seven items used to assess own perceptions of procedural justice (Colquitt, 2001), replacing the references to the self with the terms my co-workers. One of these items was thus 'To what extent do you think that your co-workers in general perceive that they have had influence over the decisions arrived at by the procedures?'. The same adaptation was done to measure co-workers' perceptions of interactional justice. As a consequence, one of the six items adapted was 'To what extent do you think that your co-workers perceive that your supervisor has treated them with dignity?'. Again, respondents rated their agreement with each statement using a 5-point Likert-type scale (1 = to a small extent; 5 = to a large extent).

Affective commitment
To measure employees' affective attachment to the organisation, we relied on a revised version of Meyer, Allen, and Smith's (1993) scale which has been previously validated Vandenberghe, Stinglhamber, Bentein, & Delhaise, 2001). For use in this study, we retained the four items displaying the highest loadings on the intended factor, in the confirmatory factor analysis results reported by Stinglhamber et al. (2002) ('I really feel that I belong in this company'; 'This organisation has a great deal of personal meaning for me'; 'I am proud to belong to this company'; 'I do not feel emotionally attached to this organisation'). To assess employees' affective attachment to their supervisor, we selected the four highest-loading items from the scale used in Stinglhamber et al.'s (2002) study. These items were: 'I feel respect for my supervisor', 'I appreciate my supervisor', 'I have little admiration for my supervisor', and 'I feel proud to work with my supervisor'. For these measures, respondents answered to the proposed statements using a 5-point Likert-type scale (1 = strongly disagree; 5 = strongly agree).

Results
Data were analysed using the structural equation modeling (SEM) approach by LISREL 8 (Jöreskog & Sörbom, 1993). Following Anderson and Gerbing's (1988) recommendations, a two-stage process was followed to analyse the data. First, we assessed the measurement model to evaluate the independence of constructs examined in this study. Second, we proceeded with the assessment of the hypothesised structural relationships among latent variables.
Note that, given the sample size and in order to limit the number of parameters to be estimated, we first reduced the number of items per factor by combining them so as to create a limited number of indicators per construct.
To do so, we used the single-factor method (Landis, Beal, & Tesluk, 2000), that is we paired the item with the highest loading with the item having the lowest loading in order to form the first composite, and so on until all items have been assigned to composites. The purpose of this specific method is to distil the original set of items to a reduced number of indicators that are empirically balanced measures of the underlying concept. As a result, we created three indicators for each justice variable and for each commitment construct. Note. *p < .05; **p < .01; ***p < .001; cw-PJ = Co-workers' procedural justice; own-PJ = Own procedural justice; cw-IJ = Co-workers' interactional justice; own-IJ = Own interactional justice; AC-org = Affective commitment to the organisation; AC-sup = Affective commitment to the supervisor; df = degrees of freedom; CFI = Comparative fit index; RMSEA = Root-meansquare error of approximation; ∆χ 2 = Chi-square difference tests between the six-factor model and alternative models.

Confirmatory factor analyses
The distinctiveness of all the variables included in our investigation was tested via a sequence of nested models (Bentler & Bonett, 1980). We compared the fit of seven nested models, ranging from the hypothesised six-factor model to a one-factor model. Details on how the variables that were distinguished in the six-factor model were combined to obtain the more constrained measurement models are provided in Table 1. The table also displays fit indices for all the measurement models that were tested.
As can be seen, the hypothesised six-factor model had the most acceptable values for the root-mean-square error of approximation (RMSEA) and the Comparative Fit Index (CFI) (Medsker, Williams, & Holahan, 1994). Moreover, the chi-square difference tests (Bentler & Bonett, 1980;James, Mulaik, & Brett, 1982) indicated that all more constrained models displayed significant decrements in fit as compared with the six-factor model. Finally, all the indicators loaded reliably on their predicted factors, with standardised loadings ranging from .87 to .88 for co-workers' procedural justice, .86 to .95 for co-workers' interactional justice, .77 to .84 for own procedural justice, .81 to .96 for own interactional justice, .73 to .82 for affective commitment to the organisation, and .78 to .90 for affective commitment to the supervisor. Consequently, we treated the six constructs separately in subsequent statistical analyses in order to examine our hypotheses.

Relationships among variables
Descriptive statistics and correlations among latent variables Means, standard deviations, internal reliabilities and both raw correlations and correlations among latent variables are given in Table 2. Of great interest, correlations within a particular focus (i.e., organisation and supervisor) are stronger than correlations between foci. For instance, own procedural justice is more strongly associated with affective commitment to the organisation than with affective commitment to the supervisor (rs = .32 and .19, respectively; t(209) = 1.74, p < .10; Cohen & Cohen, 1983), while own interactional justice is more strongly related to affective commitment to the supervisor than to affective commitment to the organisation (rs = .67 and .23, respectively; t(209) = 6.64, p < .001; Cohen & Cohen, 1983). These preliminary results are in accordance with our hypotheses. Table 2 Means, standard deviations, reliabilities, and correlations among variables Note. *p < .05; **p < .01; ***p < .001; cw-PJ = Co-workers' procedural justice; cw-IJ = Co-workers' interactional justice; own-PJ = Own procedural justice; own-IJ = Own interactional justice; AC-org = Affective commitment to the organisation; AC-sup = Affective commitment to the supervisor. Correlations among latent variables are provided below the diagonal while the raw correlations are presented above the diagonal. Cronbach's alphas are provided in parentheses on the diagonal. Table 3 Fit indices for nested structural models .06

1.27
Note. *p < .05; ***p < .001; cw-PJ = Co-workers' procedural justice; cw-IJ = Co-workers' interactional justice; own-PJ = Own procedural justice; own-IJ = Own interactional justice; AC-org = Affective commitment to the organisation; AC-sup = Affective commitment to the supervisor.; df = degrees of freedom; CFI = Comparative fit index; RMSEA = Root-mean-square error of approximation; ∆χ 2 = Chi-square difference tests between the best fitting model (Model 2) and alternative models. Table 3 presents fit indices for the hypothesised structural model (Model 1), along with those for a series of alternative models (Model 2 to Model 9). In all models, the disturbance terms associated with endogenous variables relating to the same core construct were allowed to correlate, to account for the fact that own procedural and interactional justices on one hand and affective commitments to the organisation and to the supervisor on the other hand share common causes that were unmeasured in the present study (cf Bollen, 1989;Hunt & Morgan, 1994).

SEM analyses
Model 1 (depicted in Figure 1) accurately explained the data as indicated by a RMSEA of .06 (p < .05) and a CFI of .97. However, as indicated by the chi-square difference test, adding a direct path from own interactional justice to affective commitment to the organisation (Model 2) resulted in a significant improvement in model fit (∆χ 2 (1) = 8.75, p < .001), with a RMSEA of .06 (ns), and a CFI of .97. This suggests that own interactional justice exerts a significant influence on both affective commitment to the organisation and affective commitment to the supervisor.
To assess whether Model 2 was the best depiction of the data, we compared its fit to that of a number of alternative models containing additional paths that were theoretically plausible. Details on the paths that were added in these alternative models are provided in Table 3 (see Models 3 to 9). As indicated by the chi-square difference tests reported in this table, none of these alternative models improved significantly over Model 2. We thus retained Model 2 as the best fitting model. Standardised parameter estimates for Model 2 are shown in Figure 2. For ease of presentation, we show the structural model rather than the full measurement model, and we describe the effects of the two control variables in the text. Organisational tenure was significantly related to own interactional justice (γ = -.17, p < .05) but not to own procedural justice (γ = .00, ns), affective commitment to the organisation (γ = .03, ns), nor affective commitment to the supervisor (γ = -.04, ns). Tenure with the supervisor was significantly related to own procedural justice (γ = -.16, p < .05) and affective commitment to the supervisor (γ = .15, p < .01), but not to own interactional justice (γ = -.03, ns), nor affective commitment to the organisation (γ = .13, ns). Controlling for organisational tenure and tenure with the supervisor, coworkers' procedural justice predicted own procedural justice which, in turn, had an influence on affective commitment to the organisation. Also, coworkers' interactional justice influenced own interactional justice which, in turn, had an impact on affective commitment to the supervisor. Unexpectedly, a cross-foci effect was significant. Indeed, own interactional justice also exerted an influence on affective commitment to the organisation.
In order to test for mediation, we relied on the Sobel test (Baron & Kenny, 1986;MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002) to calculate the significance of the indirect effects of independent variables on dependent variables through mediating or intervening variables. First, results indicated that the indirect effect of co-workers' procedural justice on affective commitment to the organisation through own procedural justice was significant (indirect effect = .26, z = 3.29, p < .001), thereby demonstrating that own procedural justice mediated the influence of co-workers' procedural justice on affective commitment to the organisation, which yields support to Hypothesis 1. Second, results showed that the indirect effect of co-workers' interactional justice on affective commitment to the supervisor through own interactional justice was significant (indirect effect = .39, z = 5.04, p < .001), thereby demonstrating that own interactional justice mediated the influence of co-workers' interactional justice on affective commitment to the supervisor and providing support to Hypothesis 2. Finally, results demonstrated that the indirect effect of co-workers' interactional justice on affective commitment to the organisation through own interactional justice was significant too (indirect effect = .12, z = 2.52, p < .05), thereby demonstrating that own interactional justice mediated the influence of co-workers' interactional justice on affective commitment to the organisation.

Discussion
Using a multi-foci approach, the present investigation sought (1) to explore the impact of the co-workers' perceptions of both procedural and interactional justice on own justice experiences, and (2) to link these socially constructed justice judgments with a well-known consequence of justice, affective commitment. Contrary to the limited number of studies in this field, these research questions were examined in a non-experimental field setting.
Of importance, the confirmatory factor analyses of our data revealed that employees are able to develop precise representations of their co-workers' justice judgments, which are distinct from their own perceptions of justice. This distinction occurs for both procedural and interactional justice, that is, whether the source of justice is the organisation or the supervisor. Furthermore, findings suggest that organisations and supervisors are distinct sources of justice and separate targets of employee affective commitment. More generally, these findings contribute to current streams in the literature showing that employees engage in distinct exchange processes with their organisation and their supervisor. The growing interest for the examination of employee-supervisor linkages aside from employee-employer relationships (e.g., Eisenberger, Stinglhamber, Vandenberghe, Sucharski, & Rhoades, 2002;Masterson et al., 2000;Tan & Tan, 2000) could be explained by the current trend toward globalisation of markets, which has generated tremendous transformations in companies around the world such as mergers and acquisitions (Kozlowski, Chao, Smith, & Hedlund, 1993). A consequence of these dramatic changes is that it becomes harder for employees to identify who is their employer, whose visibility to employees may be more limited that in the past. In such a context, it is plausible that immediate supervisors, especially if they are "company (wo)men", become a more critical focus of exchange with employees than ever before, so that employees may expect more of them in order to guide their behaviour (Butz, Dietz, & Konovsky, 2001). Supervisors may not only 'represent' but also partly 'replace' the organisation in keeping their subordinates motivated and willing to stay. In addition, as organisations are getting more decentralised, supervisors may have some room for developing their own relationships with employees.
The results of our SEM analyses revealed that co-workers' perceptions of procedural and interactional justice exert a significant influence on an employee's judgments of procedural and interactional justice, respectively. In addition, findings indicate that employee's own judgments of procedural and interactional justice significantly predict his/her attachment to the sources of these justices, that is the organisation and the supervisor, respectively. In other words, the results show that co-workers' perceptions of pro-cedural justice indirectly impact an employee's affective commitment to his/her organisation, through his/her own procedural justice judgments. In the same way, co-workers' perceptions of interactional justice influence an employee's affective commitment to his/her supervisor via his/her own interactional justice judgments.
Overall, these findings provide evidence that perceptions of fairness in the workplace are subject to the social influence of co-workers, but, even more importantly, show that this influence takes place on different dimensions of justice (prior research did not adopt a multi-foci approach and thus only focused on how people personally experience justice). Social influence processes are proposed to affect fairness judgments because employees experience uncertainty about their relationship to the organisation or the supervisor. As a consequence, they use fairness judgments about the way the organisation treats its members and the supervisor treats his/her subordinates as heuristics for evaluating the quality of those relationships (van den Bos, Bruins, Wilke, & Dronkert, 1999). Reactions to procedures and interpersonal treatment are thus affected by other-oriented considerations.
Another interesting finding was that when procedures and their enactment are separated, procedural justice is a primary source of affective commitment to the organisation whereas interactional justice is related to affective commitment to the supervisor. Procedural justice applied to the exchange between an employee and his/her organisation, whereas interactional justice applied to a similar process of exchange between an employee and his/her supervisor. However, these exchange processes appeared not to be totally independent of each other because supervisor-related variables also affected organisation-related ones. Indeed, results of our SEM analyses unexpectedly showed that employee's perceptions of interactional justice exert a significant influence of his/her affective commitment to the organisation. This cross-foci effect suggests that, even though the organisation and the supervisor are considered by employees as distinct entities pertaining to the workplace, the fact remains that supervisors are also viewed as representatives of the whole organisation. As "agents" of the organisation, their behaviours toward employees are perceived as indirect acts of the organisation itself (Stinglhamber, Eisenberger, Aselage, Becker, Sucharski, & Eder, 2008). Consequently, it is not so surprising that justice from supervisors (i.e., interactional justice) also contributes to the general attachment of employees to their whole organisation. Future research is needed to examine potential moderators that determine when exactly interactional justice has the most influence on variables assessing organisational consequences.
Several limitations of the study demand caution in interpreting the results. First, the cross-sectional nature of the data precludes any inference of causality. Moreover, future research should examine whether the linkages between variables examined in this study could be replicated using longitudinal designs. Furthermore, future research using longitudinal analyses with inclusion of the same variables on repeated occasions is needed to allow stronger inferences concerning causality among the variables. In particular, we assumed in the present study that what employees think co-workers' views of justice lead to their own justice judgments. However, one could argue that the causal relationship between these two variables takes place in the opposite direction.
Second, this study might be criticised for relying on self-reported variables (i.e., making common method variance a potential problem) rather than on, for example, co-workers ratings. However, we believe that self-report measures are most appropriate for assessing the constructs included in this study. Indeed, as argued in the introduction, we were here more interested in perceptions than in reality, that is in employees' perceptions of the co-workers' justice judgments and not in the actual co-workers' justice judgments. Additionally, we were able to partially address the concern over method bias by assuring participants of the anonymity of their responses. Furthermore, as recommended by Podsakoff, MacKenzie, Lee, and Podsakoff (2003), we performed analyses showing that all the structural parameters that were significant when common method variance was not controlled remained significant even when we controlled for the effects of a single unmeasured latent method factor. The results of this test indicate that the relationships found were robust to common method effects and therefore suggest that common method variance is not a pervasive problem in this study.
A strength of the present research, however is that we are the first (at least to our knowledge) to combine both the multi-foci approach with the approach of looking at the influence of co-workers' justice experiences to gain a better understanding of how justice experiences in organisations are constructed and consequently influence affective commitment. In addition, our present findings also have important implications for managers too. Above the well-known direct influence of employee's own justice perceptions on his/her affective commitments at work, results indicate that employee's attachments at work are also indirectly impacted by co-workers' justice perceptions. That means that, contrary to the current beliefs, the justice-commitment link is not a two-by-two issue. In other words, this implies that managers have to consider the social network of each person at work. Indeed, fairly treating an employee will not necessarily have positive effects on his/her justice judgments and, subsequently, on his/her attachment to the source of this fair treatment if other persons around this target employee perceive a lack of fairness in the procedures or the interpersonal treatment they are confronted to.
Taken together, justice experiences are important for organisations to pro-mote a good and healthy working environment. Understanding how fairness is experienced in organisations and larger collectives, however, requires that we as justice researchers focus on different sources of justice information and how our colleagues experience the justice climate when predicting issues of affective commitment. It is our hope that future justice research will adopt both perspectives when pursuing a better understanding of organisational justice.