CJBS 37-4 9/26/05 1:11 PM Page 250

Organizational Injustice and Psychological Strain

LORI FRANCIS, Saint Mary’s University

JULIAN BARLING, Queen’s University

Abstract sive reviews, see Cohen-Charash & Spector, 2001;

Characterizing perceived injustice as a form of stress, Colquitt, Conlon, Wesson, Porter, & Ng, 2001). For
we examined the main and interactive relationships instance, experiencing injustice contributes to such
among interactional, procedural, and distributive injus- negative outcomes as decreased work performance
tice and psychological strain while controlling for job (e.g., Pfeffer & Langton, 1993) and increased with-
insecurity. Using moderated multiple regression analysis drawal behaviour (e.g., Hulin, 1991). Although the
with a sample of 1,083 government employees, we show violation of organizational justice principles has
that interactional, procedural, and distributive injustice received considerable research attention, the relation-
are all unique predictors of psychological strain that ships most frequently considered by researchers have
account for significant unique variance beyond that involved outcome variables that are attitudinal or
explained by job insecurity. Those individuals who per- behavioural in nature, for example, organizational
ceive more interactional, procedural, or distributive injus- commitment (Alexander, Sinclair, & Tetrick, 1995),
tice at work reported a higher degree of strain. However, theft (Greenberg, 1990), and trust in the organization
there were no significant interactive effects, suggesting (Konovsky & Pugh, 1994). Until recently, the role of
that these three categories of perceived injustice do not injustice in questions pertaining to employee health,
interact to predict symptoms of psychological strain. in particular work-related strain, had not been
broached in organizational research. However, of
Résumé late, organizational researchers have recognized that
Tout en considérant l’injustice perçue comme une forme organizational injustice may play a role in the very
de stress et en tenant compte de la variable de l’insécurité real problem of strain at work. In particular,
d’emploi, nous avons examiné les relations principales et researchers have begun to demonstrate that percep-
interactives propres aux injustices en matière d’interac- tions of organizational injustice are associated with
tion, de procédure et de catégorie, et la fatigue psy- stress-related outcomes, including psychological
chologique. Une analyse de régression multiple soumise strain (e.g., Elovainio, Kivimaki, & Helkama, 2001),
à un échantillon de 1 083 employés du gouvernement depression (Tepper, 2001), and counterproductive
nous a permis de démontrer que les injustices en matière behaviour at work (Fox, Spector, & Miles, 2001). In
d’interaction, de procédure et de catégorie sont toutes des doing so, they acknowledge injustice as an emerging
variables indépendantes de la fatigue psychologique qui issue in the occupational stress literature (e.g.,
peuvent expliquer une variance unique significative dif- Kelloway, Francis, & Montgomery, 2005).
férente de celle de l’insécurité d’emploi. Les personnes In the current research, we investigate relation-
qui perçoivent, dans leur milieu de travail, un plus grand ships among the experience of injustice and one
nombre d’injustices en matière d’interaction, de procé- stress-related outcome, namely, psychological strain.
dure et de catégorie rapportent une fatigue plus intense. This work contributes to the existing literature in
Cependant, aucun effet interactif significatif n’a été three main ways. First of all, we concurrently exam-
observé, ce qui donne à penser que ces trois catégories ine the relationships among three different categories
d’injustice perçue n’interviennent pas dans la prédiction of perceived injustice and strain. Although previous
des symptômes de fatigue psychologique. studies have examined procedural, interactional, and
distributive injustice and reports of strain, few have
considered such relationships in a single study (for
exceptions see Judge & Colquitt, 2004; Tepper, 2000).
A substantial body of organizational research sup- Second, we examine the relationship between per-
ports the claim that perceived injustice in the work ceived injustice and strain in concert with another
place leads to undesirable reactions (for comprehen- variable known to be involved in the stress process,

Canadian Journal of Behavioural Science, 2005, 37:4, 250-261

CJBS 37-4 9/26/05 1:11 PM Page 251

Injustice and Strain 251

namely, job insecurity. A number of other studies stress process (Pratt & Barling, 1988). In its common
addressing the relationship between injustice and usage, stress is a term simultaneously used to
strain have not examined the relationship in the pres- describe something people feel as well as something
ence of other stress-related factors (e.g., Tepper, 2001). to which individuals are exposed in their environ-
We believe that an examination of how injustice pre- ments. However, most researchers now agree on a
dicts strain over and above another organizationally general stress model that distinguishes three closely
relevant factor will result in a more complete under- related terms: stressors, stress, and strain (Pratt &
standing of the relationship between perceived injus- Barling, 1988). Stressors are external, objectively veri-
tice and strain. fiable events such as a heavy workload or poor inter-
As our third contribution, we examine potential personal relations in the workplace that contribute to
interactive patterns among three types of perceived the experience of stress (e.g., Sauter, Murphy, &
injustice in the prediction of strain. Three-way inter- Hurrell, 1990). Stress is an individual’s internal
actions among procedural, distributive, and interac- response to stressors and is characterized by arousal
tional justice perceptions have been demonstrated in and displeasure. Finally, strain describes the conse-
the justice literature (e.g., Skarlicki & Folger, 1997). quences of long-term exposure to stress and includes
However, existing research has not considered the such physiological outcomes as cancer, gastrointesti-
potential for a three-way interaction among interac- nal illness, and cardiovascular disease (Kristensen,
tional, procedural, and distributive justice percep- 1996; Quick et al., 1997) and psychological symp-
tions with psychological strain as the outcome of toms, including anxiety (e.g., Billings & Moos, 1982),
interest. Previous considerations of two-way interac- depression (e.g., Tepper 2001; Zohar, 1995), and cog-
tions between procedural and distributive justice per- nitive failure (Kivimaki & Lusa, 1994).
ceptions when predicting psychological strain yield In this study, we treat perceived injustice as a
mixed results. One study shows that procedural and reflection of stress, that is, an internal state character-
distributive justice perceptions interact to predict ized by arousal and displeasure, and examine its rela-
multiple variables related to psychological strain, tionship to a prevalent stress-related outcome, psy-
including depression, emotional exhaustion, and chological strain. We characterize injustice as stress,
anxiety (Tepper, 2001), but another fails to detect a rather than as a stressor, because it is the individual’s
two-way interaction between procedural and distrib- internal perception of a situation that results in judg-
utive justice when predicting psychological strain ments of injustice. Via our framework, the verifiable
(Francis, 2003). With such mixed results and a lack of external event that contributed to perceived injustice,
research considering the role of interactional justice for instance, being denied a raise, would be the stres-
in justice interactions when predicting strain, the cur- sor. Therefore, we investigate further the issue of
rent study will elucidate an understudied area in the occupational stress by considering organizational
literature. Before detailing the current study, we first injustice as a form of work-related stress and examin-
review relevant research to illustrate the relevance ing its ability to predict psychological strain.
and potential impact of the current study. The study of organizational injustice is concerned
Occupational stress is a common and severe con- with people’s perceptions of fairness violations in the
dition of major concern to both individuals and orga- workplace. However, injustice is not considered as a
nizations. Stress is linked to such deleterious individ- uni-dimensional construct and many studies have
ual outcomes as depression, heart disease, and focused on three dimensions of injustice: distributive,
increased alcohol use (Quick, Quick, Nelson, & procedural, and interactional (Cohen-Charash &
Hurrell, 1997). The organizational consequences of Spector, 2001). Questions about distributive injustice
stress are also severe. For example, work-related focus on people’s perception of the fairness of out-
stress has been linked to such negative outcomes as comes (Greenberg, 1987). Research in the area of pro-
decreased organizational commitment, decreased job cedural injustice is concerned with the extent of per-
performance, increased workers’ compensation ceived fairness for the procedures by which organiza-
claims, and increased sick time (Jex & Crossley, 2005). tional decisions are made (Cropanzano & Greenberg,
Given the damaging outcomes associated with job- 1997). Interactional injustice addresses the degree of
related stress, it is not surprising that organizational perceived fairness in the interpersonal treatment that
researchers have endeavoured to gain an improved a person receives (Bies & Moag, 1986).
understanding of this phenomenon. Research interest in the relationship between
One way that stress researchers have attempted to injustice and strain has recently grown. Procedural,
advance research in this area is by clarifying the ter- interactional, and distributive injustice have all been
minology used to describe various aspects of the linked with increased reports of strain (Elovainio,
CJBS 37-4 9/26/05 1:11 PM Page 252

252 Francis and Barling

Kivimaki, & Vahtera, 2002; Elovanio et al., 2001; Fox examined two aspects of interactional injustice, infor-
et al., 2001; Francis, 2003; Judge & Colquitt, 2004; mational and interpersonal. Judgements of interper-
Kivimaki, Elovainio, Vahtera, & Ferrie, 2003; Tepper, sonal injustice involve assessments of sincerity and
2000, 2001; Wager, Fieldman, & Hussey, 2003; Zohar, respect in interpersonal interactions. Perceptions of
1995). For example, in a study focused on interper- informational injustice consider the extent to which
sonal treatment, Zohar (1995) found that nurses who individuals are given honest and adequate explana-
perceived that their supervisor reacted to their con- tions in the workplace (Bies & Moag, 1986;
cerns in an unfair manner, reported increased psy- Greenberg, 1993). Judge and Colquitt found that,
chological strain. Wager et al. (2003) reported nega- although procedural and interpersonal injustice pre-
tive changes in blood pressure, a manifestiation of dicted a measure of stress, distributive and informa-
physiological strain, among a sample of health-care tional injustice did not emerge as significant predic-
workers when they were working under an unfair tors. It is possible that these mixed results are a
leader relative to when the same individuals com- reflection of measurement or sampling issues pecu-
pleted a shift with a supervisor they perceived as fair. liar to the particular studies or the different treatment
Elovainio and colleagues (Elovanio et al., 2001, 2002; of interactional justice in each case. Whatever the
Kivimaki et al., 2003) have reported a consistent rela- case, the existence of such mixed results regarding
tionship between perceptions of both procedural and the relationship between distributive injustice and
interactional injustice and elevated levels of psycho- other stress-related variables warrants additional
logical strain, increased illness-related work research, particularly those involving multiple cate-
absences, and decreased self-reported health status. gories of injustice perceptions. We do so in the cur-
Fox et al. (2001) found that perceptions of procedural rent study and hypothesize that:
and distributive injustice were associated with
increased reports of negative affect and counterpro- Hypothesis 1: Perceptions of a procedural, interactional,
ductive behaviour in the workplace, such as arriving and distributive injustice will all significantly predict
late, avoiding work, and criticizing the organization. strain. Those who perceive a high degree of procedural,
Similarly, Tepper (2001) found that perceptions of interactional or distributive injustice in their workplace
distributive and procedural injustice were associated will report higher levels of psychological strain.
with increased psychological strain symptoms,
including depression, emotional exhaustion, and Our prediction that distributive injustice will pre-
anxiety. dict strain is contrary to some recent findings (Judge
In the current study, we extend previous investiga- & Colquitt, 2004). However, we note that the majority
tions on this matter in a number of ways. One exten- of studies examining the relationship between dis-
sion is the simultaneous examination of relationships tributive injustice and strain have found a significant
among three common categories of perceived injus- relationship (e.g., Fox et al., 2001; Tepper, 2000, 2001).
tice, namely, interactional, procedural, and distribu- Additionally, the theoretical foundation for the cur-
tive and strain. Most existing studies on injustice and rent conception of distributive injustice, equity theo-
strain have examined a subset of justice-related per- ry (Adams, 1965) purports that exposure to perceived
ceptions. For instance, Tepper (2001) included only inequity contributes to feelings of tension. As such,
procedural and distributive injustice in his work and we propose that distributive injustice will be associ-
found that both types of justice perceptions predicted ated with reports of strain.
psychological strain. Elovainio et al. (2001) on the A second way in which the current study extends
other hand focused on procedural and interactional existing research is by considering perceptions of
injustice. They reported that both of these types of injustice along with job insecurity, another acknowl-
justice judgments had significant relationships with a edged variable in the stress process. An examination
measure of psychological strain. of how injustice operates in concert with other orga-
Two exceptions to the trend of including a subset nizationally relevant factors in the prediction of
of injustice categories that we have identified in strain will result in a more complete understanding
stress-related research are Tepper (2000) and Judge of the relationship between perceptions of injustice
and Colquitt (2004). Tepper (2000) found that proce- and strain, and offer insight into the amount of addi-
dural, interactional, and distributive injustice were tional variance that perceptions of injustice may
all linked with psychological indicators of strain. explain in psychological strain.
Judge and Colquitt (2004) also considered multiple Certainly, the workplace is replete with factors
categories of injustice perceptions. In addition to pro- that contribute to stress among employees. For
cedural and distributive injustice, they separately instance, in the 2003 General Social Survey, Canadian
CJBS 37-4 9/26/05 1:11 PM Page 253

Injustice and Strain 253

workers reported that work overload and poor inter- outcomes. For instance, Skarlicki and Folger (1997)
personal relations at work were major sources of reported a significant three-way interaction among
stress (Williams, 2003). Some studies pertaining to interactional, distributive, and procedural justice
organizational injustice and strain have included when predicting organizational retaliatory behaviour.
such additional factors. For instance, Elovainio et al. In that case, the relationship between organizational
(2001) reported that procedural injustice fully medi- retaliatory behaviour and distributive justice only
ates the relationship between job control and strain, emerged when both procedural and interactional jus-
and Francis (2003) found that justice evaluations par- tice were perceived to be low. Existing research has
tially mediate that relationship. Examining the rela- not, however, considered the potential for a three-
tionships among justice perceptions, work-family way interaction among interactional, procedural, and
conflict, and stress, Judge and Colquitt (2004) deter- distributive justice perceptions with psychological
mined that work-family conflict mediates the rela- strain as the outcome of interest. Previous studies
tionship between perceived injustice and measure that tested two-way interactions between procedural
that appears to tap psychological strain. Job insecuri- and distributive justice perceptions when predicting
ty is another often considered stress-related variable psychological strain yield mixed results. Tepper
(Barling & Kelloway, 1996; Sonnentag & Frese, 2003). (2001) reported that procedural and distributive jus-
Those who report a high degree of insecurity in their tice perceptions interact to predict multiple variables
work also report more strain. However, no studies related to psychological strain, including depression,
have examined the incremental predictive ability of emotional exhaustion, and anxiety such that strain
justice perceptions when controlling for job insecuri- was highest when both procedural and distributive
ty in the prediction of strain. Further, job insecurity is justice were perceived to be low. Francis (2003), on
a particularly important variable to consider in the the other hand, did not detect a two-way interaction
context of the current study. These data were collect- between procedural and distributive justice when
ed from employees at a government organization predicting psychological strain across three separate
involved in a substantial organizational transition. As studies, although significant main effects emerged in
such, employees in this organization may have been each case.
particularly concerned about the security of their jobs One potential reason for the mixed findings when
following the change. In this study, we examine the considering the two-way interaction between proce-
extent to which organizational justice perceptions dural and distributive injustice perceptions when
contribute to the prediction of psychological strain predicting strain is the way in which distributive and
beyond the variance accounted for by job insecurity. procedural justice were measured in the separate
The importance of injustice can be demonstrated fur- studies. Francis (2003), who failed to find an interac-
ther if it accounts for a significant proportion of the tion, assessed fairness perceptions at an event- or
variance in individual strain beyond that of job inse- issue-specific level. For example, in one case, student
curity. We offer the following hypothesis: respondents were asked to evaluate the procedural
and distributive fairness of a university exam. Tepper
Hypothesis 2: Perceptions of injustice will account for (2001), who did uncover a significant interaction,
unique variance in psychological strain beyond that asked participants to provide more global assess-
explained by job insecurity. ments about fairness in their organizations. It may be
that the event-specific approach to assessing injustice
The third way that the current study extends exist- is not the most appropriate when strain is the out-
ing work is by considering the potential for interac- come of interest. As we defined earlier, psychological
tive relationships among the different types of per- strain manifests after a longer term exposure to
ceived injustice in the prediction of strain. Certainly, stress. As such, unless the episode in question is a
it seems intuitive that the ability of perceived injus- major one, injustices surrounding a single event may
tice to predict strain would be exacerbated under not, on their own, result in strain. Whereas the accu-
conditions where one perceives that injustice has mulation of many incidences of perceived injustice
been violated along multiple dimensions and miti- may constitute a persistent state of stress, and ulti-
gated when an injustice that occurs along a single mately result in strain (Tepper, 2001). In the current
dimension (e.g., distributive) and is accompanied by study, we, like Tepper (2001), use global assessments
fair treatment on another dimension (e.g., procedur- of fairness, asking participants to provide general jus-
al). Previous research has demonstrated three-way tice evaluations in their workplace.
interactions among procedural, interactional and dis- The contradictory nature of existing findings per-
tributive justice in the prediction of organizational taining to justice interactions illustrate that a clear
CJBS 37-4 9/26/05 1:11 PM Page 254

254 Francis and Barling

understanding of the manner in which various cate- TABLE 1

gories of perceived injustice relate in the prediction of Descriptive Statistics and Intercorrelations for All Study Variables
strain has not been achieved. Additionally, interac- –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
tions involving interactional injustice, including a Variable M SD 1 2 3 4 5
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
three-way interaction of interactional, procedural, 1. Procedural Injustice 2.5 .86 —
and distributive injustice, as predictors of strain, have 2. Distributive Injustice 3.1 1.0 .38 —
not been explored in the literature. The current study 3. Interactional Injustice 2.3 .88 .56 .33 —
4. Insecurity 3.0 .93 .22 .17 .16 —
considers these possible interactions. However,
5. Strain 2.2 .53 .34 .36 .33 .26 —
because of the equivocal findings regarding interac- –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
tions among the categories of perceived injustice, Listwise N = 1,083
specific hypotheses are not offered. All correlations significant at p < .001

Method TABLE 2
Participants Fit Indices for the Models
Employee opinion surveys were distributed to –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
4,591 employees of a Canadian government organi- Model χ2 df GFI AGFI CFI NFI RMSEA
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
zation. Completed surveys were returned by 1,378 3 Factor 995.13* 116 .91 .88 .96 .95 .08
individuals for a 30% response rate. Respondents 2 Factor 4,320.20* 118 .63 .52 .79 .78 .20
worked for various ministries (e.g., transportation, 1 Factor 8,745.05* 119 .45 .30 .56 .56 .29
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
health, and education). Age was assessed using a Note. N = 1,219; *p < .01.
series of age intervals, and most participants (68%)
were between the ages 30 and 50. item is, “If someone laid a complaint, my organiza-
tion would collect all the information necessary for
Procedure and Measures decision making.” Higher scores indicate a high
All employees of the government organization degree of perceived injustice. This scale showed
were invited to complete an anonymous employee acceptable internal consistency (α = .95).
opinion survey. Surveys were delivered to employees
via payroll distribution and completed surveys were Interactional injustice was measured using five
returned separately to the labour relations division of items developed by Moorman (1991). Responses
the government. were given based on a 5 point scale ranging from 1
(strongly disagree) to 5 (strongly agree). An example
Psychological strain was measured using the 12- item is, “If I were to speak to my supervisor about a
item version of the General Health Questionnaire complaint, my supervisor would treat me with kind-
(GHQ; Banks et al., 1980). The GHQ includes items ness and consideration.” Higher scores indicate a
pertaining to such factors as depression and self-con- high degree of perceived injustice. This scale had
fidence. The respondents were asked to consider acceptable internal consistency (α = .93).
their psychological strain symptoms over the past six
months. The items were rated on a 4-point response Distributive injustice was measured using five
scale ranging from 1 (better than usual) to 4 (much items described by Price and Mueller (1986).
worse than usual). Due to a printing error, one item Responses were given based on a 5-point scale rang-
from the GHQ 12 was missing in this survey (been ing from 1 (strongly disagree) to 5 (strongly agree).
able to face up to your problems). Thus, the psycho- An example item is, “In my workplace I am fairly
logical strain score was computed using 11 items. rewarded considering my responsibilities.” Higher
This variable was computed such that higher scores scores indicate a high degree of perceived injustice.
are indicative of more strain. Prior research has This scale was internally consistent (α = .94).
shown high internal consistency for this measure (α’s
ranging from .82 to .90; Banks et al., 1980). Internal Job insecurity was assessed using five items devel-
consistency for the GHQ was high in this study (α = oped for the current study. Responses were given
0.92). based on a 5 point scale ranging from 1 (strongly dis-
agree) to 5 (strongly agree). The items are presented
Procedural injustice was measured using seven in the Appendix. Higher scores indicate a high
items developed by Moorman (1991). Responses degree of insecurity. This measure was internally
were given based on a 5-point scale ranging from 1 consistent (α = .81).
(strongly disagree) to 5 (strongly agree). An example
CJBS 37-4 9/26/05 1:11 PM Page 255

Injustice and Strain 255

TABLE 3
Standardized Parameter Estimates and Squared Multiple Correlations for the Three-Factor Model

–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Distributive Interactional Procedural
Item R2
Justice Justice Justice
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
I am fairly rewarded…
…considering my responsibilities. .83 .70
…in view of the amount of experience I have. .82 .67
…for the amount of effort I put forth. .93 .87
…for the work I have done well. .91 .83
…for the stresses and strains of my job. .82 .67

If I were to speak to my supervisor about a complaint, my supervisor would…

…consider my viewpoint. .86 .75
…consider the situation objectively. .87 .76
…provide me with timely feedback about any decision. .78 .60
…treat me with kindness and consideration. .88 .77
… show concerns for my rights as an employee. .90 .80

If someone in my workplace laid a complaint my organization would…

…collect all the information necessary for decision making. .86 .73
…provide opportunities to appeal or challenge the decision. .83 .70
…follow standards or policies so that decision could be made consistently. .88 .77
… hear the concerns of all those affected by the decision. .90 .81
…provide useful information regarding the decision and its implementation. .91 .82
…allow requests for clarification about the decision. .90 .82
…ensure the privacy of all those involved. .73 .53
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Results TABLE 4
Variables were computed using listwise deletion Beta Weights and Multiple Correlation for Psychological Strain
Following Step 4 of the Hierarchical Moderated Multiple
of missing data at the item level. Prior to testing the
Regression
stated hypotheses, we examined the data for viola-
tions of the assumptions of normality, linearity, –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
homoskedasticity, and multicollinearity using SPSS Variable β R R2
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
for Windows Version 10.1. The distributions for Job Insecurity .17* .475* .225*
strain, interactional injustice, and procedural injustice Interactional Injustice .18*
were somewhat positively skewed. However, as the Procedural Injustice .13*
Distributive Injustice .24*
proposed analysis is robust to violations of the
Interactional x Procedural .003
assumption of normality, transformations were not Interactional x Distributive -.06
performed. No univariate (greater than four standard Procedural x Distributive .06
deviations from the mean) or multivariate outliers Interactional x Procedural x Distributive - .05
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
were detected. No other violations of assumptions Note. * p < .01; Listwise N = 1,083.
were detected. Descriptive statistics and intercorrela-
tions for all study variables are presented in Table 1. loaded on a single factor. Unfortunately, due to the
As there is some debate in the literature regarding nature of the interactional justice items, we were
the number of dimensions of organizational justice, unable to test a four-factor model (Colquitt, 2001) in
before testing the hypotheses pertaining to a three- which interactional justice was separated into infor-
dimensional model of organizational justice, we veri- mational and interpersonal components.
fied the factor structure using Confirmatory Factor All model tests were conducted using maximum
Analysis. We compared the fit of three separate mod- likelihood estimate in LISREL 8.51 (Joreskog &
els. The first model contained three factors in which Sorbom, 2001) and were based on the covariance
distributive, interactional, and procedural justice matrix. Fit indices for the three models are presented
loaded on separate factors. The second was a two- in Table 2. An inspection of the fit indices suggests
factor model, in which distributive justice loaded on that the three-factor model provides the best fit to the
one factor, and procedural and interactional justice data. In particular, the three factor model provides a
loaded on a second factor. The third model was one better fit to the data than does the model hypothesiz-
in which all three categories of justice perceptions ing two factors, χ2difference (2) = 3,325.97, p < .01, and
CJBS 37-4 9/26/05 1:11 PM Page 256

256 Francis and Barling

the model proposing a one-factor solution, χ2difference .001; and for distributive injustice, β =.24, t(1,074) =
(3) = 7,749.92, p < .01. Standardized parameter esti- 7.65, p < .001. As predicted in Hypothesis 1, those
mates and squared multiple correlations for the who feel that their workplaces are interactionally,
three-factor model are presented in Table 3. All para- procedurally, or distributively unfair report more
meter estimates were significant and explained sub- psychological strain.
stantial amount of item variance.
To test our hypotheses, we conducted a moderated Discussion
multiple regression. According to the procedure rec- The current investigation provides additional sup-
ommended by Aiken and West (1991), those predic- port for the proposition that exposure to injustice is a
tors involved in the interaction were standardized stressful experience. Individuals who perceived high-
before the interaction terms were computed. With er degrees of procedural, interactional, or distributive
strain as the criterion measure, job insecurity was injustice tended to report higher incidence of psycho-
entered on the first step. Interactional, procedural, logical strain symptoms. Certainly, these results are
and distributive injustice were all entered on Step 2. in keeping with those of a number of other recent
The 3 two-way interactions between the injustice studies that show perceived injustice as an important
terms were entered on Step 3, and the three-way predictor of stress-related outcomes (Elovanio et al.,
injustice interaction term was included on the fourth 2001; Elovainio et al., 2002; Fox et al., 2001; Francis,
and final step of the regression. 2003; Judge & Colquitt, 2004; Kivimaki et al., 2003;
Listwise deletion of missing data resulted in a Tepper, 2000, 2001; Wager et al., 2003; Zohar, 1995).
sample size of 1,083 for this analysis. R2 was signifi- These results are similar to those found in Tepper’s
cantly different from zero for all steps. With all vari- (2000) investigation treating procedural, interaction-
ables entered in the equation, R2 = .23, F(8, 1074) = al, and distributive justice as predictors of psycholog-
39.06, p < .001. Following Step 1, with job insecurity ical strain. However, the results of the current study
entered as the lone predictor, R2 = .07, F(1, 1081) = differ somewhat from another recent study. Judge
79.34, p < .001. and Colquitt (2004) reported that procedural injustice
As predicted, the addition of the three injustice and one component of interactional injustice, inter-
variables on Step 2 accounted for significant incre- personal injustice, were significant predictors of a
mental variance in psychological strain, ΔR2 = .152, measure of perceived stress. However, distributive
Fchange (3, 1,078) = 70.14, p < .001. Thus perceptions of injustice and an additional component of interaction-
injustice accounted for an additional 15.2% of the al injustice, informational injustice, did not emerge as
variance in psychological strain beyond that significant predictors.
explained by job insecurity. Therefore, Hypothesis 2 It is possible that a number of differences between
is supported. The addition of the two-way interaction the current study and Judge and Colquitt (2004) may
terms on the third step did not account for additional account for the discrepant results. First of all, Judge
variance in the psychological strain reported by the and Colquitt examined work-family conflict as a
government employees (ΔR2 = .003, p > .15). Similarly, mediator of justice perceptions and a measure of per-
the three-way interaction term introduced on Step 4 ceived stress. Therefore, this study considered indi-
did not significantly contribute to the variance rect, rather than direct, relationships between per-
accounted for in the criterion measure (ΔR2 = .001, p > ceived injustice and stress. Additionally, Judge and
.15). Thus, in the present study, procedural, interac- Colquitt considered four, rather than a categories of
tional, and distributive injustice do not appear to organizational justice. Their treatment separately
interact in any manner to predict incremental vari- considered interpersonal and informational justice
ance in psychological strain. perceptions, which they viewed as components of
The beta weights for each predictor following Step interactional injustice. Future research should
4 of the hierarchical moderated multiple regression address these discrepancies in an effort to more clear-
are presented in Table 4. Examination of these beta ly ascertain the roles of distributive and interactional
weights indicates that job insecurity was a significant injustice. Clearly, additional research should consider
predictor of reported psychological strain, β = .17, t the components of interpersonal and informational
(1,074) = 6.09, p < .001. Not surprisingly, those who injustice to see if informational injustice arises as a
felt higher levels of job insecurity also reported more predictor of strain under other circumstances.
strain. All three types of perceived injustice emerged Similarly, additional work should be done to identify
as significant predictors of psychological strain. For the role of other potential mediators in the relation-
interactional injustice, β =.18, t(1,074) = 5.07, p < .001; ship between injustice and strain.
for procedural injustice, β =.13, t(1,074) = 3.56, p < Another interesting finding of the current investi-
CJBS 37-4 9/26/05 1:11 PM Page 257

Injustice and Strain 257

gation is that perceived injustice accounted for con- justice. Given that significant interactions did not
siderable incremental variance in psychological emerge in the current study, the level at which the
strain. Perceptions of procedural, interactional, and justice variables are measured does not explain the
distributive injustice explained an additional 15.2% inconsistent results. However, other differences
of the variance in stress beyond that accounted for by between Tepper (2001) and the current study may
job insecurity alone. The results lend support to the account for the seemingly incompatible results.
notion that injustice is, in and of itself, an important One difference between Tepper (2001) and the cur-
aspect of organizational stress. We note that per- rent study is the degree of relationship between dis-
ceived injustice explaining variance above and tributive and procedural justice. In two separate
beyond job insecurity may take on particular impor- studies, Tepper (2001) reported correlations between
tance given the current sample. The data for the pre- distributive and procedural justice that were above
sent study were collected at a time when the govern- .6, whereas in the current study the same correlation
ment organization in question was undergoing sub- was .38. This difference is somewhat surprising given
stantial change, and individuals may have been con- that both studies appear to have used similar opera-
cerned about their job security in the postchange tionalizations of procedural and distributive injus-
environment. tice. It is possible, especially noting that Tepper’s sig-
One of our goals in the current study was to nificant interactions accounted for between 1% and
examine potential interactive relationships among 4% of the variance in strain, that the differences in the
procedural, interactional, and distributive injustice in extent to which the two justice variables are related
the prediction of strain. There were no significant affects the emergence of an interaction when they are
interactions, even though with such a large sample used to predict strain.
size there was sufficient power to detect such effects Another pertinent factor that distinguishes Tepper
if they were present. This is the only study of which (2001) and the current study is that the former used
we are aware that examines the potential interactions longitudinal investigations to assess the relationship
involving interactional, procedural, and distributive between injustice and strain. As strain manifests as a
injustice when predicting a measure of psychological result of longer-term exposure to stress, it is possible
strain. Thus, the lack of a three-way interaction, and that a longitudinal approach, which would permit
the failure to detect any two-way interactions involv- longer time frames for the individuals to experience
ing interactional injustice, is a unique contribution to the injustice they initially reported at Time 1, may
the literature. increase the likelihood of detecting a justice interac-
Other studies have examined the interactive tion when strain is the outcome of interest.
effects of perceived procedural and distributive injus- The final difference between Tepper (2001) and the
tice in the prediction of strain. The lack of a signifi- current study that we will consider as a potential rea-
cant interaction between procedural and distributive son for the contradictory results between the two
injustice in the current study is in keeping with some investigations is the nature of the samples. Tepper
previous research. For instance, Francis (2003) failed (2001) sampled respondents from a large public orga-
to detect such an interaction in three separate studies nization in one study and residents of a large city in
examining the relationship between perceived injus- another. In the current study, we also sampled from a
tice and strain. However, the current results are con- public organization; however, the organization con-
trary to those reported by Tepper (2001), who found sidered in the current study was undergoing a major
that procedural and distributive injustice consistently transition. It is possible that the respondents in our
interact to predict outcomes related to psychological study were experiencing strain as a result of stress
strain, such as exhaustion and depression. Such con- stemming from multiple stressors associated with the
flicting results suggest that procedural and distribu- change. Other than job insecurity, we were unable to
tive injustice can interact to predict psychological control for these types of stress. In this case, such
strain, but that the presence of such an interaction multiply determined symptoms of strain may have
may be dependent upon the context and conditions masked justice interactions that actually exist.
of the study. Earlier, we suggested that the reason for Whatever the reason for the contradictory nature
the conflicting findings of Tepper (2001) and Francis of the current findings and those reported by Tepper
(2003) was that the former relied on global assess- (2001), the mixed results appear to reflect the state of
ments of fairness and the latter on more event- the existing literature on justice interactions.
focused justice evaluations. We addressed this possi- Previously published results on the relationship
bility in the current study by using general assess- between procedural and distributive injustice in the
ments of procedural, interactional, and distributive prediction of varied outcomes, beyond those pertain-
CJBS 37-4 9/26/05 1:11 PM Page 258

258 Francis and Barling

ing to strain, are also mixed (see Cropanzano & relate to important outcomes (Fox et al., 2001).
Greenberg, 1997). Folger and Cropanzano (1998) note We do, however, agree that future research efforts
that when interactions between justice categories do should involve objective measures of strain. Stress
arise, they can take on different forms. One possible researchers and consumers of the organizational
form is that the negative outcomes associated with a stress literature have expressed a great deal of con-
perceived distributive unfairness only arise when cern regarding the widespread use of self-report
procedural and interactional fairness evaluations are measures (e.g., Kahn & Byosiere, 1992) and have
also low. Alternatively, it is also possible that the encouraged the increased use of physiological indices
interaction is such that high distributive justice can of strain (Sonnentag & Frese, 2003). Certainly, physio-
mitigate the effects of unfair processes, be they proce- logical indicators of strain, such as blood pressure,
dural or interactional. are now increasingly incorporated in occupational
It is obvious that, as of yet, a clear understanding stress research (e.g., Wager et al., 2003). In the future,
of the manner in which the various types of per- investigators should include such measures in
ceived injustice relate in the prediction of outcomes research that explores further the complexities of the
has not been achieved. Researchers should continue relationships among justice perceptions and strain. In
to entertain the potential for and nature of interactive doing so, these researchers will move away from the
relationships among procedural, interactional, and focus on psychological outcomes and into a broader
distributive injustice in the prediction of organiza- conceptualization of strain.
tionally relevant outcomes. For instance, future Nonresponse bias is often a concern in large, sur-
research should look for interactive patterns in stud- vey-based investigations such as this one. The cur-
ies that explicitly compare event-specific versus glob- rent response rate of 30% is not particularly low for
al assessments of fairness. A similar comparison of an organizational survey of this type. However, it is
longitudinal versus cross-sectional designs would possible that those who are most highly vulnerable to
also be valuable. In doing such research, investiga- stress and strain were more likely to complete the
tors would provide a necessary exploration of the survey than were other individuals. This may be of
boundaries of the conditions under which an interac- particular concern when response rates are low.
tive relationship emerges and what form such inter- However, fears regarding nonresponse bias are
actions might take, as well as identify the circum- allayed by research conducted by Schalm and
stances under which the categories of injustice oper- Kelloway (2002), who show that low response rates
ate independently. Understanding the impact of con- do not necessarily bias study results. Their premise is
text and research design on the interactive patterns that nonresponse is only a concern if it distorts the
among the different types justice variables would be effects under consideration. That is, response bias is
valuable to organizational practice, as there are work- problematic if low response rates are associated with
place situations in which management has more con- increased effect sizes. Their investigation of previous-
trol over the fairness of procedures or interpersonal ly published work resulted in a nonsignificant corre-
treatment than they do the outcome, for example, lation between response rate and effect size.
during a layoff decision. Additionally, given that the data for the present
Like all research, the present study is not without study were drawn from a much larger organizational
limitations. Our reliance on self-reported measures survey, there is no reason to believe that individuals
for justice, job insecurity, and stress is not ideal. who were particularly sensitive to issues of injustice
Although we acknowledge the potential limitations or strain per se would be more likely to respond to
associated with self-report measures, we nonetheless the survey.
believe that the current research makes a valuable Although demonstrations that the experience of
contribution to the literature. The literature pertain- injustice contributes to strain are becoming more
ing to injustice and strain is in its infancy. At this commonplace, very little is known about the actual
point, when little is known about the nuances of the mechanisms that govern the relationship among per-
relationship between perceived injustice and strain, ceived injustice, stress, and strain (Judge & Colquitt,
valuable information can be obtained from self- 2004), and there is no agreement regarding the partic-
report data. Additionally, with respect to the mea- ular role that perceived injustice plays in the stress
sures of organizational injustice in work environ- process. For instance, we treat perceived injustice as a
ments, it is often not desirable to move away from reflection of stress that occurs in response to an exter-
self-reported information as it is the individuals’ per- nal stressor. That is, perceived injustice is an internal
ceptions regarding the fairness of their workplaces, state that develops in response to an external event,
rather then objective assessments of fairness, that for instance, a promotion decision-making process.
CJBS 37-4 9/26/05 1:11 PM Page 259

Injustice and Strain 259

According to our treatment, perceived injustice con- This research was supported by funding from the
tributes to the experience of strain. Elovainio and col- Nova Scotia Health Research Foundation to the first
leagues (Elovanio et al., 2001; Elovainio et al., 2002; author and the Social Sciences and Humanities Research
Kivimaki et al., 2003) take a similar approach and Council of Canada to the second author. The authors
discuss how justice evaluations predict indicators of thank Noel Keeley for his assistance with data collection.
various indicators of strain. Other researchers, how- Portions of this study were presented at the Work, Stress
ever, appear to treat perceived injustice as a stressor and Health Conference, Toronto, ON, March 2003.
and examine its relationship to stress (e.g., Judge & Correspondence should be addressed to Lori Francis,
Colquitt, 2004). To that extent, considerable work in Department of Psychology, Saint Mary’s University,
articulating, describing, and dissecting this process is Halifax, Nova Scotia, Canada B3H 3C3 (E-mail:
needed. Lori.Francis@ SMU.ca).
The results of the present study carry a number of
important practical implications for Human References
Resources practice with respect to the management Adams, J. S. (1965). Inequity in social exchange. In L.
of the costly problem of occupational stress (Karasek Berkowitz (Ed.), Advances in experimental social psychol-
& Theorell, 1990). The clear demonstration that injus- ogy (Vol. 2, pp. 267-299). New York: Academic Press.
tice contributes to psychological manifestations of Aiken, L.S., & West, S.G. (1991). Multiple regression: Testing
strain should direct attention toward injustice as a and interpreting interactions. Newbury Park, CA: Sage
potential component of stress management initia- Publications.
tives. One approach to managing stress in the work- Alexander, S., Sinclair, R. R., & Tetrick, L. E. (1995). The
place is Preventive Stress Management (Quick et al., role of organizational justice in defining and maintain-
1997). Preventive Stress Management emphasizes ing the employment relationship. In L. E. Tetrick & J.
that the well-being of an organization and its Barling (Eds.), Changing employment relations:
employees are interdependent. As such, this Behavioral and social perspectives (pp. 61-89).
approach notes that stress-related interventions ideal- Washington, DC: American Psychological Association.
ly include both organizational and individual efforts. Banks, M. H., Clegg, C. W., Jackson, P. R., Kemp, N. J.,
It also highlights both the reduction of stressors in the Stafford, E. M., & Wall, T. D. (1980). The use of the
workplace and the management of existing stress and General Health Questionnaire as an indicator of men-
strain. The recognition that perceived injustice is tal health in occupational studies. Journal of
stressful might guide program development under Occupational Psychology, 53, 187-194.
the principles of Preventive Stress Management. Barling J., & Kelloway, E. K. (1996). Job insecurity and
Given the goal of reducing the number of stresses health: The moderating role of workplace control.
in the workplace, acknowledging that a state of per- Stress Medicine, 12, 253-260.
ceived injustice is stressful is a starting point for the Bies, R. J., & Moag, J. S. (1986). Interactional justice:
design of preventative interventions. For instance, if Communication criteria of fairness. In R. J. Lewicki, B.
perceived injustice is recognized as a type of stress, H. Sheppard, & M. Bazerman (Eds.), Research on nego-
Human Resources experts might include organiza- tiation in organizations, (Vol. 1; pp. 43-55). Greenwich,
tional justice principles in curricula for management CT: JAI Press.
training programs. Those managers who are aware of Billings, A. G., & Moos, R. H. (1982). Stressful life events
the damaging effects of injustice may be more mind- and symptoms: A longitudinal model. Health
ful to enact fair procedures, treatment, and outcomes Psychology, 1, 99-117.
when dealing with employees, thus reducing the Cohen-Charash, Y., & Spector, P. E. (2001). The role of jus-
incidence of unfair scenarios in the workplace. tice in organizations: A meta-analysis. Organizational
With respect to the goal of improved management Behavior and Human Decision Processes, 86, 278-321.
of existing employee stress and strain, the characteri- Colquitt, J. A. (2001). On the dimensionality of organiza-
zation of injustice as a type of stress may benefit tional justice: A construct validation of a measure.
counselling and employee assistance initiatives. For Journal of Applied Psychology, 86, 386-400.
instance, these programs may help employees recog- Colquitt, J. A., Conlon, D. E., Wesson, M .J., Porter, C. O.,
nize situations that lead to perceptions of unfairness & Ng, K. Y. (2001). Justice at the millennium: A meta-
as a contributing factor in their experience of stress analytic review of 25 years of organizational justice
and strain. As such, employees may be able to learn research. Journal of Applied Psychology, 86, 425-445.
to cope with their feelings about the perceived Cropanzano, R., & Greenberg, J. (1997). Progress in orga-
unfairness or manage the source of the perceived nizational justice: Tunneling through the maze. In C.
injustice. L. Cooper & I. T. Robertson (Eds.), International review
CJBS 37-4 9/26/05 1:11 PM Page 260

260 Francis and Barling

of industrial and organizational psychology, (Vol. 12; Management of occupational health and safety, (3rd ed).
pp. 317- 372). London: John Wiley & Sons. Scarborough, ON: Nelson.
Elovainio, M., Kivimaki, M., & Helkama, K. (2001). Kivimaki, M., Elovainio, M., Vahtera, J., & Ferrie, J. E.
Organizational justice evaluations, job control, and (2003). Organizational justice and the health of
occupational strain. Journal of Applied Psychology, 86, employees: Prospective cohort study. Occupational and
418-424. Environmental Medicine, 60, 27-34.
Elovainio, M., Kivimaki, M., & Vahtera, J. (2002). Kivimaki, M., & Lusa, S. (1994). Stress and cognitive per-
Organizational justice: Evidence of a new psychosocial formance of fire fighters during smoke diving. Stress
predictor of health. American Journal of Public Health, Medicine, 10, 63-68.
92, 105-108. Konovsky, M. A., & Pugh, S. D. (1994). Citizenship behav-
Folger, R., & Cropanzano, R. (1998). Organizational justice iour and social exchange. Academy of Management
and human resource management. Thousand Oaks, CA: Journal, 37, 656-669.
Sage Publications. Kristensen, T. S. (1996). Job stress and cardiovascular dis-
Fox, S., Spector, P. E., & Miles, D. (2001). ease: A theoretical critical review. Journal of
Counterproductive work behaviour (CWB) in Occupational Health Psychology, 3, 246-260.
response to job stressors and organizational justice: Moorman, R. H. (1991). Relationship between organiza-
Some mediator and moderator tests for autonomy and tional justice and organizational citizenship behavior.
emotions. Journal of Vocational Behavior, 59, 291-309. Do fairness perceptions influence employee citizen-
Francis, L. (2003). Organizational justice, sensitivity to ship? Journal of Applied Psychology, 76, 845-855.
injustice and the experience of stress. (Doctoral disser- Pfeffer, J., & Langton, N. (1993). The effects of wage dis-
tation, University of Guelph, 2003). Dissertation persion on satisfaction, productivity, and working col-
Abstracts International, 64(1-B), 451. laboratively: Evidence from college and university fac-
Greenberg, J. (1987). A taxonomy of organizational justice ulty. Administrative Science Quarterly, 38, 382-407.
theories. Academy of Management Review, 12, 9-22. Pratt, L. I., & Barling, J. (1988). Differentiating between
Greenberg, J. (1990). Employee theft as a reaction to daily events, acute and chronic stressors: A framework
underpayment inequity. Journal of Applied Psychology, and its implications. In J. J. Hurrell, Jr., L. R. Murphy,
75, 561-568. S. L. Sauter, & Cooper (Eds.), Occupational stress: Issues
Greenberg, J. (1993). The social side of fairness: and developments in research (pp. 41-53). New York:
Interpersonal and informational causes of organiza- Taylor & Francis.
tional justice. In R. Cropanzano (Ed.), Justice in the Price, J. L., & Mueller, C. W. (1986). Handbook of organiza-
workplace: Approaching fairness in human resource man- tional measurement. Marshfield, MA: Pittman.
agement (pp. 79-103). Hillsdale, NJ: Erlbaum. Quick, J. C., Quick, J. D., Nelson, D. L., & Hurrell, Jr., J. J.
Hulin, C. L. (1991). Adaptation, persistence and commit- (1997). Preventive stress management in organizations.
ment in organizations. In M. D. Dunnette & L. M. Washington, DC: APA Books.
Hough (Eds.), Handbook of industrial and organizational Sauter, S. L., Murphy, L. R., & Hurrell, Jr., J. J. (1990).
psychology (2nd ed.; Vol. 2; pp. 445-506). Palo Alto, CA: Prevention of work-related psychological disorders: A
Consulting Psychologists Press. national strategy proposed by the National Institute
Jex, S. M., & Crossley, C. D. (2005). Organizational conse- for Occupational Safety and Health (NIOSH).
quences. In J. Baring, E. K. Kelloway, & M. Frone American Psychologist, 45, 1146-1158.
(Eds.), Handbook of work stress (pp. 575-599). Thousand Schalm, R. L., & Kelloway, E. K. (2002). The relationship
Oaks, CA: Sage Publications. between response rate and effect size in occupational
Joreskog, K. G., & Sorbom, D. (2001). LISREL 8.51 user’s health psychology research. Journal of Occupational
guide. Chicago, IL: Scientific Software International. Health Psychology, 6, 160-163.
Judge, T. A., & Colquitt, J. A. (2004). Organizational jus- Skarlicki, D. P., & Folger, R. (1997). Retaliation in the
tice and stress: The mediating role of work-family con- workplace: The roles of distributive, procedural and
flict. Journal of Applied Psychology, 89, 395-404. interactional justice. Journal of Applied Psychology, 82,
Kahn, R. L., & Byosiere, P. (1992). Stress in organizations. 434-443.
In M. D. Dunnette & L. M. Hough (Eds.), Handbook of Sonnentag, S., & Frese, M. (2003). Stress in organizations.
industrial and organizational psychology (Vol. 3, pp. 571- In W. C. Borman, D. R. Ilgen, & R. J. Klimoski (Eds.),
650). Palo Alto, CA: Consulting Psychologists Press. Handbook of psychology, Vol. 12: Industrial and
Karasek, R., & Theorell, T. (1990). Healthy work: Stress, pro- Organizational Psychology (pp. 453-491). Hoboken, NJ:
ductivity, and the reconstruction of working life. New Wiley.
York: Basic Books. Tepper, B. J. (2000). Consequences of abusive supervision.
Kelloway, E. K., Francis, L., & Montgomery, J. (2005). Academy of Management Journal, 43, 176-190.
CJBS 37-4 9/26/05 1:11 PM Page 261

Injustice and Strain 261

Tepper, B. J. (2001). Health consequences of organization- Perspectives on Labour and Income, 4(6), 5-12.
al injustice: Tests of main and interactive effects. Zohar, D. (1995). The justice perspective of job stress.
Organizational Behavior and Human Decision Processes, Journal of Organizational Behavior, 16, 487-495.
86, 197-215.
Wager, N., Fieldman, G., & Hussey, T. (2003). The effect
on ambulatory blood pressure of working under Received May 15, 2004
favourably and unfavourably perceived supervisors. Revised January 6, 2005
Occupational and Environmental Medicine, 60, 468-474. Revised June, 2005
Williams, C. (2003). Sources of workplace stress. Accepted June 26, 2005

Appendix

Items Used to Measure Job Insecurity

Please indicate the extent to which you agree or disagree with each of the following statements where:

Strongly Disagree Disagree Neither Agree Nor Disagree Agree Strongly Agree
1 2 3 4 5

1. I can keep my current job for as long as I want it. (R)

2. This job has retirement security. (R)
3. I can be sure of my present job as long as I do good work. (R)
4. I am not really sure how long my present job will last.
5. I am afraid of losing my present job.