Journal of Applied Psychology © 2009 American Psychological Association

2009, Vol. 94, No. 2, 491–500 0021-9010/09/$12.00 DOI: 10.1037/a0013203

RESEARCH REPORTS

The Role of Overall Justice Judgments in Organizational Justice Research:

A Test of Mediation
Maureen L. Ambrose and Marshall Schminke
University of Central Florida

Organizational justice research traditionally focuses on the unique predictability of different types of
justice (distributive, procedural, and interactional) and the relative importance of these types of justice on
outcome variables. Recently, researchers have suggested shifting from this focus on specific types of
justice to a consideration of overall justice. The authors hypothesize that overall justice judgments
mediate the relationship between specific justice facets and outcomes. They present 2 studies to test this
hypothesis. Study 1 demonstrates that overall justice judgments mediate the relationship between specific
justice judgments and employee attitudes. Study 2 demonstrates the mediating relationship holds for
supervisor ratings of employee behavior. Implications for research on organizational justice are dis-
cussed.

Keywords: overall justice, mediation, justice and attitudes, fairness

Organizational justice research has flourished in the last 25 Byrne, Bobocel, & Rupp, 2001; Konovsky, 2000). Further, recent
years. In general, this work has been characterized by examining research has suggested there may be four distinct justice types:
the unique effects of different types of justice on various outcomes. distributive, procedural, interpersonal, and informational (Colquitt,
That is, the focus of this research has been on how distributive 2001; Colquitt, Conlon, Wesson, Porter, & Ng, 2001). Indeed,
justice, procedural justice, and interactional justice independently empirical support exists for each of these conceptualizations, and
affect individuals’ attitudes and behaviors. Recently, researchers it has demonstrated the relationship between each type of justice
have suggested that this singular focus on the effects of specific and a broad range of individuals’ attitudes and behavior (see
types of justice may not capture the depth and richness of indi- Cohen-Charash & Spector, 2001; Colquitt et al., 2001, for meta-
viduals’ justice experiences. These researchers suggested that a analytic reviews). However, researchers have recently questioned
shift in focus to a consideration of overall fairness judgments may the benefits of focusing exclusively on specific types of justice,
provide a more complete understanding of justice in organizational suggesting a shift toward examining overall justice judgments
settings (Ambrose & Arnaud, 2005; Hauenstein, McGonigle, & (Ambrose & Arnaud, 2005; Hauenstein et al, 2001; Lind, 2001a,
Flinder, 2001; Lind, 2001a, 2001b; Lind & van den Bos, 2002; 2001b; Tornblom & Vermunt, 1999).
Tornblom & Vermunt, 1999; van den Bos & Lind, 2002). Several reasons exist for this interest in overall justice. First,
In this article, we examine overall justice and its relationship there is an increasing acknowledgment in the justice literature that
with distributive justice, procedural justice, interactional justice, the focus on different types of justice may not accurately capture
and employee outcomes. We suggest overall justice mediates the individuals’ justice experiences. Second, a focus on overall justice
relationship between specific justice facets (distributive, proce- may broaden the questions justice researchers consider and over-
dural, and interactional) and outcomes. We present two studies. In come some limitations in current examinations of justice. We
the first we examine whether overall justice mediates the relation- discuss each of these below.
ship between specific justice types and employee attitudes. In the
second, we examine this mediating relationship for supervisors’
Overall Justice and Individuals’ Justice Experience
ratings of employee behavior.
Recently, a number of researchers have suggested weaknesses
Overall Justice related to the exclusive focus on specific types of justice. For
Most justice research accepts that three distinct justice types example, Greenberg (2001) suggested when individuals form im-
exist: distributive, procedural, and interactional (Cropanzano, pressions of justice, they are making a holistic judgment. Simi-
larly, Lind (2001b) noted that although individuals can distinguish
between the sources of their justice experience when asked, what
Maureen L. Ambrose and Marshall Schminke, Department of Manage-
ment, University of Central Florida.
drives behavior is an overall sense of fairness. Likewise, Shapiro
Correspondence concerning this article should be addressed to Maureen (2001) suggested victims of injustice react to their general expe-
L. Ambrose, Department of Management, University of Central Florida, rience of injustice.
P. O. Box 161400, Orlando, FL 32813-1400. E-mail: mambrose@bus Other researchers have made similar assertions. Hauenstein et
.ucf.edu al. (2001) suggested researchers should consider models of justice

491
492 RESEARCH REPORTS

in which perceptions of general fairness provide the central causal would be most appropriately examined in conjunction with a
mechanism. Similarly, Tornblom and Vermunt (1999) argued that global measure of justice. Finally, Lind (2001a) has suggested that
individuals consider fairness as a Gestalt, that the components of the relationship between specific types of justice and overall
fairness “are meaningful only in relation to the overall fairness of justice may vary as a function of the salience and relevance of the
the situation” (p. 51). The common thread running through all of specific justice type. However, current conceptualizations of jus-
this research is that a focus on distinct forms of justice may not tice, which are limited to exploring specific justice facets only,
provide either a complete or an accurate picture of how individuals cannot shed light on such issues.
make and use justice judgments. The idea that overall justice judgments mediate the relationship
In many ways, this current focus on overall justice brings the between specific types of justice and outcomes is present in many
field full circle. For example, Leventhal’s (Leventhal, 1980; Lev- theories of justice, yet no empirical research has examined this
enthal, Karuza, & Fry, 1980) seminal work on procedural justice conceptualization. In this article, we examine whether overall
framed procedural and distributive rules as the foundation of justice mediates the relationship between specific justice types and
overall justice judgments. Lind and Tyler’s (1988) work also employee attitudes and behavior. We predict the following:
acknowledged distributive and procedural justice as components
of overall justice in noting that procedural fairness “plays at least Hypothesis 1: Overall justice mediates the relationship be-
as large a role as distributive fairness in determining overall justice tween specific types of justice (distributive, procedural, and
judgments” (p. 135, emphasis added). interactional) and individuals’ attitudes and behavior.
If, as some researchers suggest, individuals react to their assess-
ment of overall justice, then excluding this assessment from our We note there is no discussion in the literature regarding full
research may omit an important construct and limit our theoretical versus partial mediation in this relationship. Although Lind’s
understanding of individuals’ justice experiences and their reac- (2001a) schematic of the relationship between specific fairness
tions to them. Clearly, there are research questions for which a judgments, overall fairness judgments, and outcomes suggests full
focus on specific facets of justice is desirable. However, research mediation, the issue is not explicitly addressed. Baron and Kenny
examining the overall justice construct and its relationship to (1986) suggested partial mediation is most likely the norm in
specific types of justice and outcomes is also warranted. psychology research. However, James, Mulaik, and Brett (2006)
indicated full mediation should serve as the baseline model in
evaluating mediation. Therefore, we employ a full mediation
The Mediating Role of Overall Justice model as the foundation for our investigation but explore partial
Implicit in the early conceptualizations of justice is a mediating mediation as well.
role for overall justice. That is, specific types of justice affect We address the hypothesis in two studies. In Study 1 we
overall justice judgments, which, in turn, affect outcomes (Lev- examine whether overall justice mediates the relationship between
enthal, 1980). Recent discussions also have suggested overall specific types of justice and employee attitudes. In Study 2 we
justice is the proximal driver of outcomes, with specific justice examine this mediating relationship for supervisors’ ratings of
types serving a more distal role. For example, both Greenberg’s employee behavior.
(2001) suggestion that individuals respond to a holistic assessment
of justice and Shapiro’s (2001) assertion that individuals respond Study 1
to their general justice experience imply overall justice is the more
proximal influence on outcomes. Method
Some researchers have been more overt about the mediating role
overall justice plays in the relationship between specific types of Sample and Procedures
justice and outcomes. For example, Lind’s (2001a) Fairness Heu- Surveys were distributed to 425 employees from 54 organiza-
ristic Theory explicitly suggests overall justice judgments mediate tions in the southeast United States including technology, govern-
the relationship between specific justice judgments and outcomes. ment, insurance, financial, food service, retail, manufacturing, and
Colquitt and his colleagues (Colquitt & Shaw, 2005; Scott, medical organizations. Survey packets were hand-delivered to five
Colquitt, & Zepata-Phelan, 2007) have been explicit in suggesting to seven employees in each department. The cover letter indicated
that overall justice mediates the relationship between specific
justice experiences and outcomes as well.1
Conceptualizing overall justice as a mediator of the relationship 1
Research on specific justice facets has sometimes placed justice facets
between specific justice facets and outcomes changes the way we judgments in the role of mediators. For example, research on applicant
think about justice and broadens the avenues of research that reactions to selection processes demonstrates that procedural justice judg-
scholars might pursue. For example, there are circumstances in ments mediate the relationship between selection process characteristics
which a focus only on overall justice may be warranted, such as such as job relatedness, consistency, and bias, and applicant reactions (e.g.,
when researchers are interested in the influence of fairness relative Chapman & Webster, 2006; Hausknecht, Day, & Thomas, 2004; Smither,
Millsap, Stoffey, Reilly, & Pearlman, 1996). Our conceptualization differs
to other individual characteristics (e.g., affect, personality) or
from this previous work by suggesting that overall justice judgments
contextual characteristics (e.g., trust, support). Additionally, mediate between these facet judgments and outcomes. Thus, our concep-
Colquitt and Shaw (2005) suggested researchers should match the tualization would add an additional link to these conceptual models.
level of specificity of the justice construct being investigated to the Attributes of the experience would affect specific facet judgments (proce-
outcomes of interest. Therefore, questions involving global atti- dural, distributive, and interactional). Overall justice judgments would then
tudes such as organizational commitment or job performance mediate the relationship between these facet judgments and outcomes.
 RESEARCH REPORTS 493

the study was being conducted for academic research purposes in POJ also includes three items to assess the fairness of the organi-
an effort to better understand some of the issues that affect people zation generally: “Usually, the way things work in this organiza-
at work. Participants were assured of the confidentiality of their tion are not fair” (POJ2, reverse scored); “For the most part, this
responses. A postage-paid envelope was included in the packet to organization treats its employees fairly” (POJ5); “Most of the
return the survey to the researchers. Employees were not compen- people who work here would say they are often treated unfairly”
sated for their participation in the study. A total of 330 surveys (POJ6, reverse scored). Individuals reported their agreement with
were returned for a 78% response rate. Forty-two percent of the each POJ statement (as well as those for satisfaction, commitment,
respondents were male, and 58% were female. The average age of and turnover intentions below) on a 7-point scale ranging from 1
respondents was 34.1 years, with 3.1 years of tenure in the depart- (strongly disagree) to 7 (strongly agree). Responses to the items
ment and 4.8 years of tenure in the organization. were recoded to parallel the specific justice items, such that higher
The survey contained demographic questions and questions ratings reflect greater perceptions of fairness. The ␣ for overall
assessing perceptions of specific types of justice, overall justice, justice was .93.2
and job attitudes (job satisfaction, organizational commitment, and Job attitudes. Job satisfaction was assessed by using five items
turnover intentions). Ten different forms of the survey were used. (␣ ⫽ .82) from the Job Satisfaction Index (Brayfield & Rothe,
The order for the measures was determined randomly for each 1951). Organizational commitment was measured with five items3
form. (␣ ⫽ .76) from the Affective Commitment Scale (Meyer & Allen,
1997; Meyer, Allen, & Smith, 1993). Two items from Cropanzano,
James, and Konovsky (1993) assessed turnover intentions (␣ ⫽ .67).
Measures
Distributive, procedural, and interactional justice. Specific
Results
justice perceptions were assessed with scales developed by
Colquitt (2001). Distributive justice was assessed by four items Correlations, means, and standard deviations for all variables
(Cronbach’s ␣ ⫽ .95). Procedural justice was assessed by seven are presented in Table 1.
items (␣ ⫽ .89). Interactional justice was assessed by nine items
(␣ ⫽ .95). For all items, individuals responded on a 5-point
Likert-type scale ranging from 1 (to a small extent) to 5 (to a great Test of the Measurement Model
extent). We conducted a series of confirmatory factor analyses (CFAs)
Overall justice. Two approaches have been suggested in the on the justice items and attitude items. Because more complex
literature for how to measure overall justice (Colquitt & Shaw, models might obscure the fit of the justice measures, we first
2005; Lind, 2001a). Both approaches reflect what have been examined the measurement model for the justice items only. Fol-
termed entity judgments, which ask individuals to assess some lowing these CFAs, we examined the measurement model includ-
entity (e.g., organization, group, or supervisor) as a whole. Entity ing the justice items and the attitude items jointly.
judgments reflect a general assessment of the fairness of the entity For the justice items, we conducted a CFA analysis to assess the
(Cropanzano et al., 2001). However, the two approaches differ in fit of our theorized four-factor model (distributive, procedural,
their specific suggestions for assessing overall fairness. Lind interactional, and overall justice). Results in Table 2 indicate the
(2001a) suggested overall justice represents an individual’s global four-factor model provides an acceptable fit to the data. We
evaluation of the fairness of his or her experiences. Consequently, compared this model with two alternative models: a two-factor
it should be assessed by items focusing on the individual’s per- model (all of the specific justice items in one factor and the overall
sonal experiences (e.g., Overall, how fairly treated am I?; Lind, justice items in a second factor; this model is analogous to a
2001a, p. 85). Colquitt and Shaw (2005) provided an alternative composite overall justice factor versus a global overall justice
approach. Drawing on Cropanzano et al. (2001), they suggested factor) and a one-factor model. The four-factor model was a
overall justice items be composed of general statements about the significantly better fit than was the two-factor model: ␹2 differ-
organization (e.g., How fair is this organization?). These state- ence(df ⫽ 5) ⫽ 3,815.20, p ⬍ .01; or the one-factor model: ␹2
ments do not refer directly to individuals’ own personal experi- difference(df ⫽ 6) ⫽ 6,010.49, p ⬍ .01. Thus, the four-factor
ences. Indeed, research has demonstrated that individuals also use
information about the fairness experiences of others to form their
impressions of fairness (Kray & Lind, 2002; Lind, Kray, & 2
Although we believe it is useful to include both types of items in an
Thompson, 1998). Thus, in forming a global assessment of an assessment of overall justice, we conducted all analyses for both studies by
entity, individuals may consider the fairness of the organization using only the three personal experience items (Cronbach’s ␣ ⫽ .90 for
generally (e.g., In general, how fairly does the organization treat its both Study 1 and Study 2) or the three general experience items (Cron-
employees?). bach’s ␣ ⫽ .85 for Study 1; Cronbach’s ␣ ⫽ .84 for Study 2). In all cases,
the pattern of results remains the same.
Using a deductive approach (Hinkin, 1998), we developed a 3
six-item measure consistent with both Lind’s (2001a) and Colquitt Our survey included the seven-item version of the Affective Commit-
ment Scale (Meyer & Allen, 1984). Because this scale has been revised, we
and Shaw’s (2005) suggestions for measuring overall justice,
deemed it best to utilize the five items that were consistent between the
called the Perceived Overall Justice (POJ) scale. The POJ scale 1984 scale and the revised 1997 scale; one item from the 1997 scale was
consists of three items to assess individuals’ personal justice experi- not assessed. The pattern of results with the 1984 (seven-item) measure is
ences: “Overall, I’m treated fairly by my organization” (POJ1); “In the same as with the five-item version of the 1997 measure. The correlation
general, I can count on this organization to be fair” (POJ3); “In between the two commitment measures is .97, and the reliabilities for the
general, the treatment I receive around here is fair” (POJ4). The two scales are also the same (.79).
494 RESEARCH REPORTS

Table 1
Study 1: Descriptive Statistics and Correlations

Variable M SD 1 2 3 4 5 6 7

1. Distributive justice 3.14 1.17 (.95)

2. Procedural justice 3.01 0.88 .57 (.89)
3. Interactional justice 3.80 1.01 .45 .53 (.95)
4. Overall justice 5.19 1.23 .46 .46 .37 (.93)
5. Satisfaction 5.30 1.18 .27 .31 .36 .54 (.82)
6. Commitment 4.29 1.21 .26 .33 .30 .55 .66 (.76)
7. Turnover intentions 4.03 1.61 ⫺.29 ⫺.35 ⫺.27 ⫺.51 ⫺.55 ⫺.70 (.67)

Note. All correlations significant at the 0.01 level (two-tailed), N ⬎ 285 for all variables. Reliabilities (coefficient alpha) are in parentheses.

model (distributive, procedural, interactional, and overall justice) the mediator (overall justice) and three from the mediator to the
represents the best fit to the data.4 outcome variables. As such, it assumes that all effects of the
We next assessed the fit of a seven-factor model reflecting antecedent variables on the outcome variables are exerted indi-
distributive, procedural, interactional, and overall justice, and the rectly, through the mediator. The partial mediation model adds to
three job attitude measures (see Table 2). Because our previous this model nine direct paths, connecting each of the three anteced-
CFAs demonstrate the usefulness of four distinct justice factors, ent variables to each of the three outcome variables. Thus, it
we compared the fit of this seven-factor model with that of a assumes that the antecedent variables may exert either direct or
five-factor model that retained the four justice factors, and we indirect effects on the outcome variables.
combined all attitude items into a single factor. The seven-factor Fit indices indicate that the full mediation model provides an
model (reflecting three attitude factors) was a significantly better acceptable fit to the data: ␹2(df ⫽ 656) ⫽ 3,624.19 (root-mean-
fit than was the five-factor model: ␹2 difference(df ⫽ 11) ⫽ 71.8, square error of approximation [RMSEA] ⫽ .10, incremental fit
p ⬍ .01. index [IFI] ⫽ .94, and comparative fit index [CFI] ⫽ .94), as does
the partial mediation model: ␹2(df ⫽ 647) ⫽ 3,605.53 (RMSEA ⫽
Structural Model .10, IFI ⫽ .94 and CFI ⫽ .94). However, the partial mediation
model does not provide a better fit to the data: ␹2 difference(df ⫽
Our hypothesis predicts overall justice will mediate the relation-
ship between antecedent variables (distributive, procedural, and
interactional justice) and outcome variables (job satisfaction, com- 4
We also considered a five-factor model that distinguishes between
mitment, and turnover intentions). We analyzed the data via struc- interpersonal justice and informational justice. The CFA demonstrated the
tural equation modeling, utilizing LISREL 8.80. Following Mayer five-factor model was a good fit to the data, ␹2(df ⫽ 289) ⫽ 683.10
and Gavin (2005) and Schneider, Ehrhart, Mayer, Saltz, and Niles- (RMSEA ⫽ .06, IFI ⫽ .95, CFI ⫽ .95), and an improvement over the
Jolly (2005), we sought to identify the most appropriate structural four-factor model. However, the correlation between interpersonal justice
model before testing our hypothesis. Existing theory and prior and informational justice is .79. This correlation falls well above the
research do not provide a compelling rationale for whether overall guideline (.70) suggested by Colquitt and Shaw (2005) for the aggregation
justice will partially or fully mediate the relationship between of justice constructs. This suggests that although the five-factor model may
provide a better fit, a four-factor model is more appropriate. This decision
specific justice types and outcomes. Thus, we follow the recom-
is also consistent with Colquitt and Shaw’s caveat that, “Confirmatory
mendation of James et al. (2006) that full mediation represents the factor analyses can provide some guidance, but it must be noted that the fit
best choice of a baseline model. statistics in those analyses almost always reward the more complex factor
In this case, we compare the fit of this full mediation model with structure, even when correlations among the factors are particularly high”
that of an alternative partial mediation model. The full mediation (p. 138).
model includes six paths, three from the antecedent variables to A comment on the magnitude of the correlation between interpersonal
justice and informational justice is warranted. Although the correlation in
this study is higher than that found in Colquitt et al.’s (2001) meta-analysis
Table 2 (rc ⫽ .66), it is consistent with more recent studies that distinguish between
Study 1: Confirmatory Factor Analyses interpersonal justice and informational justice. We reviewed justice re-
search published in the Journal of Applied Psychology, Personnel Psychol-
Item ␹2 df RMSEA IFI CFI ogy, Academy of Management Journal, Organizational Behavior and Hu-
man Decision Processes, and Journal of Management since the publication
Justice items only of Colquitt’s scale in 2001. Of the 87 articles we found, only 7 studies
Four-factor model 1,480.02 293 .11 .96 .96 (Bell, Wiechmann, & Ryan, 2006; Humphreys, Ellis, Conlon, & Tinsley,
Two-factor model 5,295.22 298 .23 .86 .86 2004; Judge & Colquitt, 2004; Kernan & Hanges, 2002; Liao & Rupp,
One-factor model 7,490.51 299 .27 .80 .80 2005; Roch & Shanock, 2006; Scott et al., 2007) distinguished between
Justice and attitude items
interpersonal justice and informational justice and included both in their
Seven-factor model 2,491.93 644 .09 .95 .95
Five-factor model 2,563.73 655 .09 .94 .94 study. In these 7 articles, the correlation between interpersonal justice and
informational justice often exceeds the guideline provided by Colquitt and
Note. RMSEA ⫽ root-mean-square error of approximation; IFI ⫽ incre- Shaw. Six of the 13 correlations reported are above .70, and 2 of these are
mental fit index; CFI ⫽ comparative fit index. above .80.
 RESEARCH REPORTS 495

9) ⫽ 18.66, ns. The rule of parsimony, therefore, suggests the fully Neither employees nor supervisors were compensated for their
mediated model is the preferred model (James & Brett, 1984; participation in the study. A total of 125 matched employee–
James et al., 2006). In the fully mediated model, distributive supervisor dyads (250 participants) returned surveys, representing
justice exerted a significant direct effect on overall justice (struc- a 91% response rate. Forty-five percent of the employee respon-
tural coefficient ⫽ .25, p ⬍ .01), as did procedural justice (.43, p ⬍ dents were male, and 55% were female. The average age of
.01) and interactional justice (.14, p ⬍ .01), with an R2 for variance employee respondents was 31.2 years, with 3.1 years of tenure in
explained in overall justice of .33. Overall justice in turn exerted a the department and 4.0 years of tenure in the organization.
significant effect on job satisfaction (.50, p ⬍ .01), commitment The employee survey contained demographic questions and
(.56, p ⬍ .01), and turnover intentions (–.53, p ⬍ .01), with an R2 questions assessing perceptions of specific types of justice and
for variance explained of .33, .46, and .42, respectively. Figure 1 overall justice. Eight different forms of the employee survey were
illustrates the full mediation model. used. The order for the measures was determined randomly for
each form. The supervisor survey contained questions to assess
Study 2 ratings of employee job behaviors, including task performance,
OCBs, and organizational deviance. Two forms of the supervisory
Study 1 provides strong support for our hypothesis. Overall survey were used, with employee performance ratings appearing
justice mediates the relationship between specific justice types and first in one version and last in the other.
employee attitudes. However, one limitation of Study 1 is the
reliance on same-source data; both the justice perceptions and
attitude assessments came from the employee. In Study 2 we Measures
address this limitation by examining the mediating role of overall
justice on the relationship between the specific justice facets and Procedural, distributive, interactional, and overall justice. As
outcomes by utilizing supervisors’ reports of employee behavior. in Study 1, distributive, procedural, and interactional justice were
Specifically, we examine three types of behavior previously dem- assessed with Colquitt’s (2001) items. Overall justice was assessed
onstrated to be related to fairness: organizational citizenship be- with the same six items as in Study 1. Reliabilities (Cronbach’s ␣)
havior (OCB), task performance, and organizational deviance. for procedural, distributive, interactional, and overall justice, were
Thus, we examine whether our Study 1 mediation results can be .90, .95, .95, and .92, respectively.
replicated in a second sample that uses non-self-report outcomes. Task performance. Employee task performance was assessed
by supervisors with a seven-item (␣ ⫽ .84) task performance scale
(Williams & Anderson, 1991). This and all subsequent measures
Method
utilized a 7-point scale ranging from 1 (strongly disagree) to 7
Sample and Procedures (strongly agree).
Organizational citizenship behavior. Employee OCB was as-
Surveys were distributed to 137 employee–supervisor dyads sessed by supervisors with an eight-item (␣ ⫽ .93) OCB scale
(274 individuals) from 58 organizations in the southeast United
(Williams & Anderson, 1991).
States. The organizations represented a broad range of functions,
Organizational deviance. Organizational deviance on the part
including retail, service, manufacturing, medical, food service,
of the employee was assessed by supervisors with a 12-item (␣ ⫽
entertainment, technology, and educational organizations. Survey
.94) organizational deviance scale (Bennett & Robinson, 2000).
packets were hand-delivered to participants. As with Study 1, the
cover letter indicated the study was being conducted for academic
research purposes in an effort to better understand some of the Results
issues that affect people at work. Participants were assured of the
confidentiality of their responses. A postage-paid envelope was Correlations, means, and standard deviations for all variables
included in each packet to return the survey directly to the authors. are presented in Table 3.

R2

Distributive Job .33

Justice .25** Satisfaction
.50**

Procedural .43** .56**

Overall Justice Commitment .46
Justice

–.53**
.14** R 2 = .33
Interactional Turnover
Justice Intentions .42

Figure 1. Study 1 structural equation modeling results: Full mediation model. ⴱ p ⬍ .05, one-tailed; ⴱⴱ p ⬍ .01,
one-tailed.
496 RESEARCH REPORTS

Table 3
Study 2: Descriptive Statistics and Correlations

Variable M SD 1 2 3 4 5 6 7

1. Distributive justice 4.64 1.61 (.95)

2. Procedural justice 4.22 1.32 .53ⴱⴱ (.90)
3. Interactional justice 5.54 1.39 .33ⴱⴱ .54ⴱⴱ (.95)
4. Overall justice 5.06 1.34 .47ⴱⴱ .68ⴱⴱ .52ⴱⴱ (.92)
5. Task performance 4.90 0.61 .06 .19ⴱ .29ⴱⴱ .19ⴱ (.84)
6. OCB 4.02 0.81 .08 .23ⴱ .31ⴱⴱ .12 .69ⴱⴱ (.93)
7. Organizational deviance 1.39 0.59 ⫺.19ⴱ ⫺.30ⴱⴱ ⫺.31ⴱⴱ ⫺.22ⴱ ⫺.57ⴱⴱ ⫺.58ⴱⴱ (.94)

Note. N ⬎ 115 for all variables. Reliabilities (coefficient alpha) are in parentheses. OCB ⫽ organizational citizenship behavior.
ⴱ
p ⬍ .05. ⴱⴱ p ⬍ .01.

Test of the Measurement Model interactional justice) and outcome variables (task performance,
OCB, and organizational deviance). We analyzed the data via
We conducted a series of CFAs on the justice items and job
structural equation modeling, utilizing LISREL 8.80.
behavior items. As in Study 1, we first examined the measurement
Consistent with Study 1, we consider both a fully mediated
model for the justice items only. Following this series of CFAs, we
model and a partially mediated model. Results indicate that the full
examined the measurement model including the justice items and
mediation model provides a good fit to the data: ␹2(df ⫽ 1316) ⫽
the job behavior items jointly.
2,439.45 (RMSEA ⫽ .08, IFI ⫽ .92 and CFI ⫽ .92). As in Study
For the justice items, we conducted a CFA to assess the fit of our
1, fit indices for an alternative partial mediation model (which
theorized four-factor model (distributive justice, procedural jus-
allows for direct effects of distributive, procedural, and interac-
tice, interactional justice, and overall justice). Results in Table 4
tional justice on outcomes, as well as their hypothesized indirect
indicate the four-factor model provides an acceptable fit to the
effects via overall justice) does not significantly improve the fit to
data. We compared this model with two alternative models. First,
the data: ␹2(df ⫽ 1307) ⫽ 2,415.65 (RMSEA ⫽ .08, IFI ⫽ .92,
we examined a two-factor model with all of the specific justice
CFI ⫽ .92); ␹2 difference(df ⫽ 9) ⫽ 23.80, ns. Thus, the rule of
items in one factor and the overall justice items in a second factor.
parsimony again indicates the fully mediated model is the pre-
Second, we examined a one-factor model. The four-factor model
ferred model (James et al., 2006).
was a significantly better fit than was the two-factor model: ␹2
Procedural justice (structural coefficient ⫽ .58, p ⬍ .01) and
difference(df ⫽ 13) ⫽ 1,596.48, p ⬍ .01; or the one-factor model:
interactional justice (.24, p ⬍ .01) each exerted direct effects on
␹2 difference(df ⫽ 14) ⫽ 2,007.51, p ⬍ .01. Thus, the four-factor
overall justice, while distributive justice did not (.08, ns). Overall
model (distributive, procedural, interactional, and overall justice)
R2 for variance explained in overall justice was .61. Overall justice
represents the best fit to the data.
in turn exerted a significant effect on task performance (.25, p ⬍
We next assessed the fit of a seven-factor model reflecting
.05), OCB (.18, p ⬍ .05), and organizational deviance (–.28, p ⬍
distributive, procedural, interactional, and overall justice, as well
.01), with an R2 for variance explained of .06, .03, and .08,
as the three job behavior measures (see Table 4). We compared the
respectively. These results also provide support for our hypothesis.
fit of this seven-factor model with that of a five-factor model that
Figure 2 illustrates this model and results.
retained the four justice factors and combined all the job behavior
items into a single factor. The seven-factor model was a signifi-
cantly better fit than was the five-factor model: ␹2 difference(df ⫽ Discussion
11) ⫽ 1,798.45, p ⬍ .01.
In this article, we presented two studies to examine the relation-
Structural Model ship between overall justice judgments, specific justice judgments,
and outcomes. We hypothesized overall justice would mediate the
Our hypothesis predicts overall justice will mediate the relation-
relationship between specific justice judgments and attitudes and
ship between antecedent variables (distributive, procedural, and
behavior. The results provide clear support for this hypothesis.
Table 4 Study 1 demonstrated overall justice mediated the effect of the
Study 2: Confirmatory Factor Analyses specific justice judgments on employees’ self-reports of job satis-
faction, commitment, and turnover intentions. Study 2 demon-
Item ␹2 df RMSEA IFI CFI strated this mediating effect for supervisors’ evaluations of em-
ployees’ OCB, task performance, and organizational deviance. The
Justice items only results of both studies suggest these relationships are best de-
Four-factor model 595.12 285 .09 .96 .96
Two-factor model 2,191.60 298 .23 .82 .82 scribed by full mediation.
One-factor model 2,602.63 299 .26 .79 .79 Two aspects of the results warrant additional attention. First, it
Justice and job behavior items is useful to examine further the relationship between the specific
Seven-factor model 2,299.12 1304 .08 .93 .93 justice judgments and overall justice. In Study 1, all three justice
Five-factor model 4,097.57 1315 .14 .88 .89
facets were significant predictors of overall justice judgments.
Note. RMSEA ⫽ root-mean-square error of approximation; IFI ⫽ incre- However, in Study 2, distributive justice did not significantly
mental fit index; CFI ⫽ comparative fit index. predict overall justice. Perhaps the more robust bivariate rela-
 RESEARCH REPORTS 497

R2

Distributive Task .06

Justice .08 Performance
.25*

Procedural .58** .18*

Overall Justice OCB .03
Justice

.24** R 2 = .61 –.28**

Interactional Organizational
Justice Deviance .08

Figure 2. Study 2 structural equation modeling results: Full mediation model. OCB ⫽ organizational citizen-
ship behavior. ⴱ p ⬍ .05, one-tailed; ⴱⴱ p ⬍ .01, one-tailed.

tionships between procedural, interactional, and overall justice specific. The idea here is similar to the differential effects ap-
in Study 2 overshadowed the relationship between distributive proach in that the predictor and the criterion are matched, but it
justice and overall justice (which is relatively consistent across differs in that it suggests it is not just the focus of the attitude or
the studies). behavior that matters, but the specificity of the assessment as well.
Second, a comment on full versus partial mediation is war- This idea is consistent with that of Colquitt and Shaw (2005), who
ranted. The fully and partially mediated models exhibited similar suggested a fit should exist between the level of specificity of the
fit to the data in both studies. Given the rule of parsimony, the fully justice construct being investigated and the outcomes of interest.
mediated model is preferred. Nonetheless, the adequate fit for the Specifically, they suggested questions concerning how improving
partially mediated model raises the question about circumstances the fairness of performance appraisal or compensation systems
in which partial mediation may be a more appropriate model than would necessarily require an equally specific justice construct.
is full mediation. We consider a few such situations below. However, for questions involving global attitudes such as organi-
Much research on organizational justice focuses on the effect of zational commitment or job performance, a more global measure
different types of justice on different types of outcomes. Specifi- of justice would be more appropriate. Thus, it may be that full
cally, researchers have suggested there should be a stronger rela- mediation is less likely when judgments of a specific process,
tionship between distributive justice and specific outcome- outcome, or interaction are related to specific attitudes about that
referenced attitudes and behavior, procedural justice and process, outcome, or interaction (e.g., fairness of performance
organization-referenced outcomes and behavior, and interactional appraisal associated with satisfaction with performance appraisal)
justice and supervisor-referenced outcomes and behavior. Am- than when specific facet judgments are related to global outcomes
brose and her colleagues (Ambrose & Arnaud, 2005; Ambrose, (e.g., commitment, performance).
Hess, & Gansar, 2007) called this the differential effects approach. Other factors might also affect full versus partial mediation. For
Although one might intuitively expect evidence of partial media- example, Lind’s (2001a) framework suggests that partial media-
tion in situations where the foci of the specific justice judgment tion is perhaps most likely during the judgment phase (as a general
matches the foci of the attitude or behavior, we do not think there justice judgment is developing) whereas full mediation is most
is a strong basis for predicting partial mediation in these cases. likely during the use phase (in which the general justice judgment
Despite the appeal of the differential effects conceptualization, is driving attitudes and behavior).5
meta-analytic results suggest the support for these differential There are a number of theoretical and conceptual benefits to
effects is mixed (Colquitt et al., 2001; Cohen-Charash & Spector, considering a global approach to justice. The results suggest the
2001). Indeed, recently Ambrose et al. (2007) reviewed these importance of overall fairness in understanding individuals’ justice
findings, tested competing models, and did not find support for a experiences and reactions. In particular, overall justice judgments
differential effects conceptualization of the relationships between mediate the relationship between specific justice judgments and
specific justice judgments and attitudes. outcomes. These findings are consistent with recent discussions of
However, we believe there may be some situations in which individuals’ justice experiences.
outcomes are strongly associated with specific justice facets. For Recognizing the role of overall justice as the link between
example, bandwidth-fidelity theory suggests it is important to specific justice types and outcomes raises some interesting ques-
match the breadth or generality (bandwidth) of a predictor variable tions. Consider Lind’s (2001a) suggestion that the relationship
to that of the criterion variable to be predicted (Cronbach, 1970; between the specific justice facets and overall justice may vary as
Cronbach & Gleser, 1965). Stronger relationships emerge, and a function of the salience and relevance of the specific type of
optimal explanatory power is achieved, when the complexity of the justice. Justice experiences that occur early in the relationship with
predictor matches that of the criteria being predicted. the organization (often procedural in nature) might be particularly
Based on bandwidth fidelity, one might expect full mediation by
overall justice to be less likely when the target of the justice
5
assessment and the target of the attitude or behavior are similarly We thank an anonymous reviewer for this suggestion.
498 RESEARCH REPORTS

influential in the development of overall justice judgments. Addi- others, suggesting the specific justice judgments and overall justice
tionally, contextual variables such as organizational structure may judgments demonstrate similar stability over time. Although not a
influence the salience and relevance of the specific justice facets definitive test, these data suggest overall justice judgments are not
(Ambrose & Schminke, 2003; Schminke, Ambrose, & Cropan- more stable than specific justice judgments over time. Rather,
zano, 2000; Schminke, Cropanzano, & Rupp, 2002). Future ex- specific justice experiences may continue to influence the overall
amination of contextual variables like structure or climate on the justice judgments. We suspect there is an ongoing interplay be-
relationship between specific justice types and overall fairness tween specific justice experiences and overall justice judgment.
would be useful. However, the precise nature of that relationship is a topic for future
Finally, an overall justice construct provides a more parsimoni- research.
ous approach to examining justice. It is quite common in the As with all studies, there are limitations. First, in Study 1, we
justice literature for researchers to make identical predictions used a single method and a single respondent. However, Study 2
about the effect of some moderating or mediating variable on all demonstrates similar effects by using supervisor ratings of em-
three subtypes of justice (e.g., Kernan & Hanges, 2002; Liao & ployee behavior. Second, we used a cross-sectional design and
Rupp, 2005; Wayne, Shore, Bommer, & Tetrick, 2002). Under data collection method, which might inflate the relationship be-
such circumstances, the underlying construct of interest appears to tween our variables. Third, our overall justice scale is a direct
be a general sense of justice rather than the specific justice sub- measure of fairness. That is, it is a measure that asks respondents
types. An overall assessment of justice provides a better indicator directly how fair something is. In contrast, the specific justice
of that underlying construct and a more parsimonious approach to scales are indirect measures (i.e., a measure that describes at-
studying justice. Therefore, unless a clear theoretical basis exists tributes of fairness such as voice, consistency, courteous treat-
for making differential predictions across different subtypes of ment). Indirect measures such as scales developed by Folger and
justice, researchers should assess overall justice instead. Konovsky (1989), Moorman (1991), and Colquitt (2001) are the
In addition to these implications for scholars, our results also most common measures of the specific justice facets. Colquitt and
have implications for practitioners. For example, in employee Shaw (2005) suggested direct measures be used when justice is an
surveys, length is almost always an issue. A general measure of endogenous variable (such as a mediator), while indirect measures
fairness may allow organizations to assess fairness more parsimo- are appropriate when justice is an exogenous variable (i.e., the
niously and, consequently, to understand better its relation to other predictor variable). We followed this advice. Nonetheless,
important outcomes. Similarly, previous research has suggested the relationship between direct measures of the justice facets and
that although individuals are capable of expressing perceptions of the direct measure of overall justice may differ from the relation-
specific justice facets when asked to do so, they tend to think in ship with indirect measures.
terms of general justice impressions. Thus, the overall justice Colquitt and Shaw (2005) also suggested direct measures
construct may provide practitioners, managers, and employees may be more susceptible to biases because words such as “fair”
alike with a language of justice that is both more accessible to may be morally charged. Similarly, an overall measure may be
them and more accurately descriptive of their concept of organi- more likely to capture general affect. In collecting data for
zational justice. Study 2, we measured negative affectivity as well as the vari-
Of course, there are questions our study cannot answer about ables utilized in the main analyses. In a subsequent structural
overall justice. For example, we suggest specific justice experi- equation modeling analysis, we controlled for negative affec-
ences are the foundation for overall justice judgments. However, it tivity in the model. The pattern of results is the same as those
is likely that once formed, overall justice may influence percep- we report above. Moreover, the model fit, ␹2(df ⫽ 1875) ⫽
tions of specific justice experiences. Indeed, Lind (2001a) sug- 3,303.65 (RMSEA ⫽ .08, IFI ⫽ .89, CFI ⫽ .89), was worse
gested overall justice judgments are quite stable over time. As this than that of our hypothesized model. Thus, for these data at
is an important issue, we collected additional data to provide a least, general affect does not seem to account for the effect of
preliminary examination of the stability of the specific justice overall justice.
judgments and overall justice judgments over time. We assessed
both specific justice judgments and overall justice judgments of
fifty-seven respondents at Time 1 and then again 4 months later Conclusion
(Time 2). All respondents were in the same job with the same
supervisor at Time 1 and Time 2. Dependent sample t tests In this article, we strive to contribute to the organizational
revealed respondents’ judgments of procedural justice (PJ) and justice literature by examining the relationship between overall
distributive justice (DJ) were stable across the 4 month period: PJ justice, specific types of justice, and employee attitudes and be-
Time 1 ⫽ 4.42, Time 2 ⫽ 4.60, t(df ⫽ 52) ⫽ –1.25, ns; DJ Time havior. Our hope is that demonstrating the mediating role of
1 ⫽ 4.85, Time 2 ⫽ 4.75, t(df ⫽ 54) ⫽ .50, ns. In contrast, overall justice in the relationship between specific types of justice
judgments of interactional justice (IJ) and overall justice (OJ) and outcomes will encourage scholars to think differently about
changed significantly from Time 1 to Time 2: IJ Time 1 ⫽ 5.53, justice and to pursue new avenues of inquiry. We do not suggest
Time 2 ⫽ 5.14, t(df ⫽ 50) ⫽ 2.84, p ⬍ .01; OJ Time 1 ⫽ 5.26, that work on specific types of justice and their effect on outcomes
Time 2 ⫽ 4.99, t(df ⫽ 50) ⫽ 2.18, p ⬍ .05. We also compared the should cease. Rather, we suggest justice research could benefit
correlations between the Time 1 and Time 2 justice measures. The from an alternative, contemporaneous, and complementary line of
correlations were .62, .64, .69, and .74 for distributive justice, inquiry— one including overall justice judgments. We believe
procedural justice, interactional justice, and overall justice, respec- overall justice is particularly useful in considering questions such
tively. None of the correlations was significantly different from the as the relative impact of justice versus other organizational mo-
 RESEARCH REPORTS 499

tives. Moreover, we believe this is an area that has largely been tions about “what, why, and how.” Journal of Vocational Behavior, 58,
overlooked in the current literature. 210 –219.
We have learned much from examining the unique effects and Hauenstein, N. M. T., McGonigle, T., & Flinder, S. W. (2001). A meta-
relative impact of distributive, procedural, and interactional jus- analysis of the relationship between procedural justice and distributive
tice. We will continue to do so. However, as the field of organi- justice: Implications for justice research. Employee Responsibilities and
zational justice continues to mature, considering the holistic influ- Rights Journal, 13, 39 –56.
Hausknecht, J. P., Day, D. V., & Thomas, S. C. (2004). Applicant reactions
ence of overall justice on employee outcomes is a path worth
to selection procedures: An updated model and meta-analysis. Personnel
traveling as well.
Psychology, 57, 639 – 683.
Hinkin, T. R. (1998). A brief tutorial on the development of measures for
References use in survey questionnaires. Organizational Research Methods, 1,
104 –121.
Ambrose, M. L., & Arnaud, A. (2005). Distributive and procedural justice:
Humphreys, S. E., Ellis, A. P. J., Conlon, D. E., & Tinsley, C. H. (2004).
Construct distinctiveness, construct interdependence, and overall justice.
Understanding customer reactions to brokered ultimatums: Applying
In J. Greenberg & J. Colquitt (Eds.), The handbook of organizational
negotiation and justice theory. Journal of Applied Psychology, 89, 466 –
justice (pp. 59 – 84). Mahwah, NJ: Erlbaum.
Ambrose, M. L., Hess, R. L., & Gansar, S. (2007). The relationship 482.
between justice and attitudes: An examination of justice effects on James, L. R., & Brett, J. M. (1984). Mediators, moderators, and tests for
specific and global attitudes. Organizational Behavior and Human De- mediation. Journal of Applied Psychology, 69, 307–321.
cision Processes, 103, 21–36. James, L. R., Mulaik, S. A., & Brett, J. M. (2006). A tale of two methods.
Ambrose, M. L., & Schminke, M. (2003). Organization structure as a Organizational Research Methods, 9, 233–244.
moderator of the relationship between procedural justice, interactional Judge, T. A., & Colquitt, J. A. (2004). Organizational justice and stress:
justice, perceived organizational support, and supervisory trust. Journal The mediating role of work–family conflict. Journal of Applied Psy-
of Applied Psychology, 88, 295–305. chology, 89, 395– 404.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable Kernan, M. C., & Hanges, P. J. (2002). Survivor reactions to reorganiza-
distinction in social psychological research: Conceptual, strategic, and tion: Antecedents and consequences of procedural, interpersonal, and
statistical considerations. Journal of Personality and Social Psychology, informational justice. Journal of Applied Psychology, 87, 916 –928.
51, 1173–1182. Konovsky, M. (2000). Understanding procedural justice and its impact on
Bell, B. S., Wiechmann, D., & Ryan, A. M. (2006). Consequences of business organizations. Journal of Management, 26, 489 –511.
organizational justice expectations in a selection system. Journal of Kray, L. J., & Lind, E. A. (2002). The injustices of others: Social reports
Applied Psychology, 91, 455– 466. and the integration of others’ experiences in organizational justice judg-
Bennett, R. J., & Robinson, S. L. (2000). The development of a measure of ments. Organizational Behavior and Human Decision Processes, 89,
workplace deviance. Journal of Applied Psychology, 85, 349 –360. 906 –924.
Brayfield, A. H., & Rothe, H. F. (1951). An index of job satisfaction. Leventhal, G. S. (1980). What should be done with equity theory? In K. J.
Journal of Applied Psychology, 35, 307–311. Gergen, M. S. Greenberg, & R. H. Willis (Eds.), Social exchange:
Chapman, D., & Webster, J. (2006). Toward an integrated model of Advances in theory and research (pp. 27–55). New York: Plenum.
applicant reactions and job choice. International Journal of Human Leventhal, G. S., Karuza, J., & Fry, W. R. (1980). Beyond fairness: A
Resources Management, 17, 1032–1057. theory of allocation preferences. In G. Mikula (Ed.), Justice and social
Cohen-Charash, Y., & Spector, P. E. (2001). The role of justice in orga- interaction (pp. 167–213). New York: Springer-Verlag.
nizations: A meta-analysis. Organizational Behavior and Human Deci- Liao, H., & Rupp, D. E. (2005). The impact of justice climate and justice
sion Processes, 86, 278 –321. orientation on work outcomes: A cross-level multifoci framework. Jour-
Colquitt, J. A. (2001). On the dimensionality of organizational justice: A nal of Applied Psychology, 90, 242–256.
construct validation of a measure. Journal of Applied Psychology, 86, Lind, E. A. (2001a). Fairness Heuristic Theory: Justice judgments as
386 – 400.
pivotal cognitions in organizational relations. In J. Greenberg, & R.
Colquitt, J. A., Conlon, D. E., Wesson, M. J., Porter, C. O., & Ng., K. Y.
Cropanzano (Eds.), Advances in organizational justice (pp. 56 – 88).
(2001). Justice at the millennium: A meta-analytic review of 25 years of
Stanford, CA: Stanford University Press.
organizational justice research. Journal of Applied Psychology, 86, 425–
Lind, E. A. (2001b). Thinking critically about justice judgments. Journal of
445.
Vocational Behavior, 58, 220 –226.
Colquitt, J. A., & Shaw, J. C. (2005). How should organizational justice be
Lind, E. A., Kray, L. J., & Thompson, L. (1998). The social construction
measured? In J. Greenberg & J. A. Colquitt (Eds.), The handbook of
of injustice: Fairness judgments in response to own and others’ unfair
organizational justice (pp. 113–152). Mahwah, NJ: Erlbaum.
Cronbach, L. J. (1970). Essentials of psychological testing. New York: treatment by authorities. Organizational Behavior and Human Decision
Harper & Row. Processes, 75, 1–22.
Cronbach, L. J., & Gleser, G. C. (1965). Psychological tests and personnel Lind, E. A., & Tyler, T. R. (1988). The social psychology of procedural
decisions (2nd ed.). Urbana: University of Illinois Press. justice. New York: Plenum.
Cropanzano, R., Byrne, Z. S., Bobocel, D. R., & Rupp, D. R. (2001). Moral Lind, E. A., & van den Bos, K. (2002). When fairness works: Toward a
virtues, fairness heuristics, social entities, and other denizens of organi- general theory of uncertainty management. In B. M. Staw & R. M.
zational justice. Journal of Vocational Behavior, 58, 164 –209. Kramer (Eds.), Research in organizational behavior (Vol. 24, pp. 181–
Cropanzano, R., James, K., & Konovsky, M. (1993). Dispositional affec- 223). Boston: Elsevier.
tivity as a predictor of work attitudes and job performance. Journal of Mayer, R. C., & Gavin, M. B. (2005). Trust in management and perfor-
Organizational Behavior, 14, 595– 606. mance: Who minds the shop while the employees watch the boss?
Folger, R., & Konovsky, M. A. (1989). Effects of procedural and distrib- Academy of Management Journal, 48, 874 – 888.
utive justice on reactions to pay raise decisions. Academy of Manage- Meyer, J. P., & Allen, N. (1984). Testing the “side bet theory” of organi-
ment Journal, 32, 115–130. zational commitment: Some methodological considerations. Journal of
Greenberg, J. (2001). Setting the justice agenda: Seven unanswered ques- Applied Psychology, 69, 372–378.
500 RESEARCH REPORTS

Meyer, J. P., & Allen, N. (1997). Commitment in the workplace: Theory, Shapiro, D. (2001). The death of justice theory is likely if theorists neglect
research, and application. Thousand Oaks, CA: Sage. the “wheels” already invented and the voices of the injustice victims.
Meyer, J. P., Allen, N., & Smith, C. (1993). Commitment to organizations Journal of Vocational Behavior, 58, 235–242.
and occupations: Extension and test of a three component conceptual- Smither, J. W., Millsap, R. E., Stoffey, R. W., Reilly, R. R., & Pearlman,
ization. Journal of Applied Psychology, 78, 538 –551. K. (1996). An experimental test of the influence of selection procedures
Moorman, R. H. (1991). Relationship between organizational justice and on fairness perceptions, attitudes about the organization, and job pursuit
organizational citizenship behaviors: Do fairness perceptions influence intentions. Journal of Business and Psychology, 10, 297–318.
employee citizenship? Journal of Applied Psychology, 76, 845– 855. Tornblom, K. Y., & Vermunt, R. (1999). An integrative perspective on
Roch, S. G., & Shanock, L. R. (2006). Organizational justice in an social justice: Distributive and procedural fairness evaluations of posi-
exchange framework: Clarifying organizational justice distinctions. tive and negative outcome allocations. Social Justice Research, 12,
Journal of Management, 32, 299 –322. 39 – 64.
Schminke, M., Ambrose, M. L., & Cropanzano, R. (2000). The effect of van den Bos, K., & Lind, E. A. (2002). Uncertainty management by means
organizational structure on perceptions of procedural fairness. Journal of of fairness judgments. In M. P. Zanna (Ed.), Advances in experimental
Applied Psychology, 85, 294 –304. social psychology (Vol. 34, pp. 1– 60). San Diego, CA: Academic Press.
Schminke, M., Cropanzano, R., & Rupp, D. E. (2002). Organization Wayne, S. J., Shore, L. M., Bommer, W. H., & Tetrick, L. E. (2002).
structure and fairness perceptions: The moderating effects of organiza- The role of fair treatment and reward in perceptions of organizational
tional level. Organizational Behavior and Human Decision Processes, support and leader–member exchange. Journal of Applied Psychol-
89, 881–905. ogy, 87, 590 –598.
Schneider, B., Ehrhart, M. G., Mayer, D. M., Saltz, J. L., & Niles-Jolly, K. Williams, L. J., & Anderson, S. E. (1991). Job satisfaction and organiza-
(2005). Understanding organization– customer links in service settings. tional commitment as predictors of organizational citizenship and in-role
Academy of Management Journal, 48, 1017–1032. behaviors. Journal of Management, 17, 601– 617.
Scott, B. A., Colquitt, J. A., & Zapata-Phelan, C. P. (2007). Justice as a
dependent variable: Subordinate charisma as a predictor of interpersonal Received September 11, 2006
and informational justice perceptions. Journal of Applied Psychology, Revision received May 5, 2008
92, 1597–1609. Accepted June 2, 2008 䡲