Organizational viability in complex, fast-changing, and turbulent
economic times requires employees willing to exceed the roles and
responsibilities defined by formal job descriptions (Jordan and
Sevastos, 2003). Organizational citizenship behaviors (OCB) can improve
organizational performance and adaptability in environments demanding
complex, ambiguous, and team-oriented work (Organ et al., 2005).
Organizational citizenship behaviors (OCB), defined as volitional
extrarole behaviors not directly related to a specific task or job
description, lead to improved customer and peer relationships, enhanced
teamwork, operational flexibility, and competitiveness (Borman, 2004).
According to Organ et al. (2005), OCB are discretionary employee
behaviors performed for the benefit of the organization or co-workers
that exceed nominal job requirements and not formally recognized by the
organization. Ryan queried, "why would an employee engage in work
that enhances organizational performance, but is not necessarily
recognized or rewarded by his or her employer" (2002: 123)? Ryan
argued that personality characteristics, such as a Protestant work
ethic, conscientiousness, or empathy explained OCB. Kidder and Parks
(2001) contended that employee-defined roles and work-identity
influenced OCB. However, Coyle-Shapiro, Kessler, and Purcell (2004)
found that personality factors such as conscientiousness, positive or
negative affectivity, or agreeableness failed to predict OCB. Job
breadth explained only an additional 11% of the variation in OCB beyond
that explained by perceptions of justice and organizational commitment
(Coyle-Shapiro et al., 2004). Instead, the majority of researchers
(Cardona et al., 2004; Coyle-Shapiro et al., 2004; Kaufman et al., 2001;
Rhoades and Eisenberger, 2002) pose OCB as a form of social exchange for
positive treatment received from the organization.
This study explores the nature of social exchange by assessing if
employees report OCB as an intended method to reciprocate acts of POS
and justice. The study adds to the body of research on OCB antecedents
by investigating the mediating effects of POS on the relationships
between three dimensions of organizational justice (procedural,
distributive, and interactive), and employee self-reported intentions to
enact OCB directed at the organization or peers. The research question
posed: Do employees acknowledge an exchange relationship between
organizational attributes, such as perceived organizational support
(POS) or justice, and their intentions to enact OCB.
This article begins with the purpose and background for the study.
The article then overviews the key constructs of organizational
citizenship behaviors (OCB), perceived organizational support (POS), and
three dimensions of organizational justice. Included is a discussion of
issues associated with self-reported intentions to enact OCB. In the
methods section, the article describes the sample, data collection
procedure, and measurement instrument. Hypothesis testing and a review
of results follow. The article closes with key findings, recommendations
tot future research, and study limitations.
BACKGROUND
Social exchange occurs when a person, motivated by the returns
those acts are anticipated to bring, voluntarily engages in acts
beneficial to another (Blau, 1986). According to the norm of
reciprocity, acts of helping are contingent on the expectation that the
recipient will reciprocate with an act of helping in the future
(Gouldner, 1960). Unlike contractual obligations, which demand
repayment, social exchange creates unspecified reciprocal obligations
enforced through cultural and normative standards of behaviors
(Cropanzano and Mitchell, 2005). According to Organ and Konovsky (1989),
employees perform OCB in anticipation that the organization will
discharge its accrued obligations through increased employee rewards or
other acts favorable to employees. In turn, organizations, with work
environments advantageous to employees, create social and normative
pressures on employees to reciprocate through behaviors valued by the
organization (Eisenberger et al., 2001).
Essential to social exchange is a sense of trust that employees and
the organization will discharge accrued obligations (Organ and Konovsky,
1989). Perceptions of justice influence employee perceptions of trust
when organizational procedures, decision-making processes, and resource
allocations are equitable (Organ and Konovsky, 1989). According to
Cropanzana and Mitchell (2005), social exchange mediates the
relationship between supportive and equitable treatment by the
organization, and employee work effort, attitudes, and commitment.
Moreover, perceived organizational support (POS) reflects the
"quality of the social exchange that takes place between an
employee and the employer" (Cropanzano and Mitchell, 2005: 883).
APPROACH
Organizational Citizenship Behaviors (OCB)
Organ et al. defines OCB as "individual behavior that is
discretionary, not directly or explicitly recognized by the formal
reward system, and in the aggregate promotes the efficient and effective
functioning of the organization" (2005: 8). Empirical and
theoretical research demonstrates that OCB correlates with improvements
in organizational performance (Organ et al., 2005; Podsakoff and
MacKenzie, 1997). In a review of research, Organ et al. (2005) found
that some OCB dimensions (i.e., helping and consciousness) correlated
more strongly with performance than other dimensions (i.e.,
sportsmanship or civic virtue). L. J. Williams and Anderson (1991) found
that employees directed some OCB at peers in the organization, such as
helping or taking a personal interest in others, and directed other OCB
at the organization, such as exemplary attendance, offering suggestions,
participating in the life of the organization, or frugally using
organizational resources.
Perceived Organizational Support (POS)
Perceived organizational support (POS) is an employee belief that
the organization cares for and values his or her contribution to the
success of the organization (Kaufman et al., 2001). Antecedents of POS
include procedural justice, supportive and respectful acts by
supervisors, recognition, fair pay and rewards, promotions, job
security, autonomy, and training (Rhoades and Eisenberger, 2002).
Consequences of heightened POS are increased organizational commitment,
job satisfaction, positive affect, task interest, task performance, and
intentions to remain with the organization (Rhoades and Eisenberger,
2002). Perceived organizational support (POS) encompasses policies
promoting employee well-being and feelings of accomplishment, a sense of
positive contribution to the organization, and personal and
organizational goal attainment (Eisenberger et al., 2001).
Eisenberger et al. (1986), in two studies, correlated perceived
organization support with increased employee efforts to attain
organizational goals. Findings by Cardona et al. (2004) demonstrated
that employees reporting higher POS had stronger social attachment to
the organization, which then led to increased OCB. In contrast, Settoon,
Bennett, and Linden (1996) failed to correlate POS with OCB. However,
Moorman et al. (1998) argued that Settoon et al. failed to assess OCB
directed at the organization. Studies by Kaufman et al. (2001) and Wayne
et al. (2002) confirmed that POS correlated with OCB directed at the
organization but not with OCB directed at peers. Suggested by this
research are the following hypotheses:
[H.sub.1]: There is a positive correlation between employee
self-reported perceptions of organizational support and self-reported
employee intentions to enact OCB directed at the organization.
[H.sub.2]: There is no correlation between employee self-reported
perceptions of organizational support and self-reported employee
intentions to enact OCB directed at peers.
Organizational Justice
Organizational justice encompasses three dimensions: interactive,
procedural, and distributive justice (Williams et al., 2002).
Organizations convey a sense of concern for employee well-being and
their contribution to organizational success when rewards and resources
are fairly distributed (distributive justice), decision-making
procedures are equitable (procedural justice), and supervisors treat
employees with dignity and respect (interactive justice) (Rhoades and
Eisenberger, 2002). Research (Henle, 2005; Stecher and Rosse, 2005)
demonstrated that inequitable treatment influenced employee decisions to
engage in acts intended to harm co-workers (interpersonal deviance) or
the organization (organizational deviance), reduce work effort, or
terminate employment. According to the findings of Kelloway, Loughlin,
Barling, and Nault (2002), employees enacting deviant behaviors were
less inclined to enact OCB. Loi et al. (2006) demonstrated that
procedural and distributive justice, mediated by POS, led to increased
employee commitment to the organization and intentions to remain. Lynch
et al. (1999) also found that supportive and equitable treatment
overcame employee's fear of exploitation and reciprocation
wariness.
Procedural and Interactive Justice. Procedural justice derives from
the perceived equity of organizational polices and procedures
determining resource allocation and other managerial decisions (Williams
et al., 2002). Employees judge the equity of procedures by the amount of
bias, the breadth and accuracy of information gathering, number of
relevant parties given voice in the decision, ethical standards applied,
and the consistency and universality of decision implementation (Stecher
and Rosse, 2005). Interactive justice concerns respectful and equitable
treatment of employees by immediate supervisors (Williams et al., 2002).
Perceptions of interactive justice result from supervisor trust-building
behaviors such as "availably, competence, consistency,
discreetness, fairness, integrity, loyalty, openness, promise
fulfillment, receptivity, and overall trust" (Deluga, 1994: 317).
Research by Moorman (1991) and S. Williams et al. (2002) found
interactive justice to be the sole significant predictor of OCB. Hubbell
and Chory-Assad (2005) found that procedural justice predicted both
organizational and managerial trust, while interactive justice failed to
correlate with either type of trust. Alternatively, research (Chiaburu
and Marinova, 2006; Cropanzano et al., 2002; Roch and Shanock, 2006)
found that interactive justice correlated with positive
employee-supervisor exchange relationships, while procedural justice
correlated with OCB directed at the organization. Similarly, the
analysis of Colquitt et al. (2001) found moderate correlations between
procedural justice and OCB directed at the organization, no correlation
between interactive justice and OCB directed at the organization, and
weak correlations between procedural and interactive justice and OCB
directed at peers.
According to social exchange theory, enactment of justice
demonstrates organizational support and commitment to employees, which
employees reciprocate through OCB (Cropanzano and Mitchell, 2005). Wayne
et al. (2002) found that POS, but not acts of supervisory support,
correlated with OCB directed at the organization. Stinglhamber, De
Cremer, and Mercken (2006) similarly found that procedural justice
promoted increased POS, which led to increased trust in the
organization, while interactive justice promoted increased perceived
supervisor support, which increased employee trust in supervisors.
Moorman et al. (1998) and Jordan and Sevastos (2003) found that POS
mediated the relationship between procedural justice and OCB, while
Wayne et al. (2002) found that POS mediated the relationships between
both procedural and distributive justice and OCB. These findings suggest
that procedural justice, mediated by POS, correlates with OCB directed
at the organization while interactive justice, with its focus on the
supervisor, does not. Thus, Hypotheses 3 and 4 are as follows.
[H.sub.3]: Perceived organizational support (POS) will mediate the
positive correlation between employee self-reported perceptions of
procedural justice and employee intentions to enact OCB directed at the
organization.
[H.sub.4]: There will be no correlation between employee
self-reports of interactive justice and POS.
Distributive Justice. Distributive justice refers to equitable
distributions of rewards and resources (Williams et al., 2002),
including fair pay and performance evaluations (Flint et al., 2005). S.
Williams et al. (2002) argue that economic and reward equity influences
employee perceptions of organizational fairness, which subsequently
predicts OCB behaviors. Cardona et al. (2004) and others (Roch and
Shanock, 2006; Williams et al., 2002) failed to support a significant
positive correlation between distributive justice and OCB. However,
Wayne et al. (2002) and Rhoades and Eisenberger (2002) found modest
correlations between distributive justice and OCB, when mediated by POS.
Chiaburu and Marinova (2006) also demonstrated that distributive justice
moderated the relationship between organizational trust and OCB directed
at the organization. Colquitt et al. (2001) found distributive justice
moderately correlated with OCB directed at the organization but weakly
correlated to OCB directed at peers. Suggested by this research is the
following hypothesis.
[H.sub.5]: Perceived organizational support (POS) will mediate the
positive correlation between employee self-reported perceptions of
distributive justice and employee intentions to enact OCB directed at
the organization.
Figure I summarizes the hypothesized relationships between study
variables. The study predicts that POS will fully mediate the
relationship between two of three justice dimensions (procedural and
distributive justice) and OCB directed at the organization. Procedural
and distributive justice will correlate with POS, but POS will not
correlate with interactive justice. Similarly, POS will correlate with
OCB directed at the organization, but not with OCB directed at peers.
Self-reported Intentions to Enact OCB
If OCB results out of an explicit quid-pro-quo exchange
relationship, then employee self-ratings of perceived organizational
factors should correlate with self-reported intentions to enact OCB. Yet
Organ et al. (2005) warned against self-reported measurements of OCB.
According to Organ et al. (2005), self-reported intentions to commit OCB
do not correlate with observer-rated OCB measurements and fail to
predict actual employee behaviors.
Self and Other OCB Ratings. Beehr, Ivaniskaya, Hansen, Erofeev, and
Gudanowski (2001), in a study on 360-degree feedback, found that
employee self-rated evaluations of performance failed to correlate with
objective performance data. Allen, Barnard, Rush, and Russell (2000)
also found stronger correlations between supervisor and subordinate
ratings of employee OCB than between employee self-reported ratings and
either supervisor or subordinate ratings. In contrast, Donaldson and
Grant-Vallone (2002), in evaluating the Project WORKWELL database,
failed to support the accuracy of peer reports over self-reports.
Ferris, Judge, Rowland, and Fitzgibbons (1994) also found that
supervisor-focused employee impression tactics increased supervisor
liking for those employees, which resulted in higher performance ratings
and resource allocations. As Piercy et al. noted, "the manager may
be more aware of the good actor, who seeks to make a favorable
impression, rather than the good soldier" (2006: 257).
[FIGURE I OMITTED]
Self and Other OCB Definitions. Studies by Van Dyne and LePine
(1998), Lam, Law, and Hui (1999), and Vey and Campbell (2004) found that
employees, employee peers, and supervisors held different definitions of
what constituted in-role or extra-role behaviors. These findings
demonstrated that the target benefiting from increased OCB tended to
frame OCB as in-role while the person enacting OCB tended to frame OCB
as extra-role (Vey and Campbell, 2004). Lain et al. concluded, "if
we are interested in subordinates' work behaviors, we should use
subordinate definitions and ratings; if we are interested in how
supervisors appraise subordinates, we should use supervisor definitions
and ratings" (1999: 599).
Intentions to Enact OCB and Actual OCB. Organ et al. (2005) also
argued that self-reported intentions to commit OCB failed to predict
actual employee OCB. Becker, Randall, and Reigel (1995), in a study of
112 employees at 16 fast-service restaurants, discovered that self-rated
multi-dimensional measurements of intentions-to-act-altruistically,
founded on commitment, attitudes and normative behaviors, predicted
subsequent altruistic behaviors, as measured by supervisors. Though
Becket et al. (1995) failed to find that self-reported intentions
predicted subsequent behaviors when intention measurements were limited
to cognitive assessments, the richer and more encompassing
multi-dimensional scale found a link between self-reported intentions
and actual behaviors. Becker et al. (1995) did not suggest that intent
and behavior failed to correlate, as argued by Organ et al. (2005), but
that measurements of self-reported intentions, when limited to cognitive
representations, might under-report actual correlations between intent
and actual behaviors. Research by Morwitz, Johnson, and Schmittlein
(1993) also demonstrated a relationship between participant responses on
an intention-type questionnaire and subsequent actual behaviors.
METHOD
Sample and Procedure
Participants in this correlation study were employees of a small
North American manufacturing organization. Each member of the
organization received a single survey containing questions concerning
OCB intentions directed at peers, OCB intentions directed at the
organization, perceptions of distributive, procedural, and interactive
justice, and perceptions of perceived organizational support (POS). The
total number of surveys distributed was 102, with 58 issued to union
factory employees and 44 to sales, technical, and administrative staff.
Surveyed factory employees averaged 43.5 years of age (SD = 13.7), with
an average tenure with the present company of 14.7 years (SD = 12.2).
The surveyed factory population was male. Surveyed office employees
averaged 41.9 years of age (SD = 10.0), with an average tenure of 11.4
years (SD = 9.5) and a 17% female population. Usable surveys totaled 93,
with 49 from factory employees and 44 from sales, technical, and
administrative employees. The survey covered 92% of employee population
in attendance.
Self-Reports. This study employed self-reports, in concert with
Spector's conclusion that "it is difficult to get accurate
information about internal states, such as attitudes or emotions, with
anything other than self-reports" (2006: 229). Both Van Dyne and
LePine (1998) and Piercy et al. (2006) argued for self-reports when
assessing self-conceptualizations of OCB behaviors. Yet according to
Podsakoff et al. (2003), instruments using self-reported responses were
subject to common method variance. Self-reports introduce common method
variance when raters attempt to maintain consistency across
questionnaire items, report illusionary relationships, fit responses to
socially desirable norms, or alter responses in concert with a negative
or positive framing of events (Podsakoff et al., 2003). Yet research
(Chan, 2001; Ganster et al., 1983; Moorman and Podsakoff, 1992; Werner,
1994) failed to support significant biasing effects from social
desirability, impression management, or negative and positive
affectivity on self-reported assessments of organizational behavioral
constructs, such as job satisfaction, organizational commitment, and
perceived organizational support. According to Spector (2006), if common
method variance persisted in all self-reported measures then variables
would consistently correlate, a finding not supported by Spector's
review.
Variables and Measures
In this study, data collection relied on a single questionnaire,
encompassing scales measuring OCB intentions directed at the
organization, OCB intentions directed at peers, POS, distributive
justice, procedural justice, and interactive justice. Questionnaire
design and data collection methods insured respondent anonymity, linkage
of data by respondent, and minimized the impact of this research on the
organization. Copyright holders of each questionnaire granted permission
for use in this study. Appendices A and B list the survey questions and
their sources.
The four-item consideration directed-at-peers scale from S.
Williams and Shiaw (1999) measured employee intentions to enact OCB
directed at peers. The three-item civic virtue scale from S. Williams
and Shiaw (1999), supplemented with questions from Farh, Earley, and Lin
(2002) assessing acts of conscientiousness beneficial to the
organization, measured employee intentions to enact OCB directed at the
organization. The nine-item scale by Eisenberger, Huntington,
Hutchinson, and Sowa (2002) measured perceived organizational support.
Three questions from the six-item distributive justice index by
Price and Mueller (2002) and one question from Spector (2002) measured
distributivejustice. Excluded were questions from Price and Mueller
(2002) which reflected topics similar to those assessed by the three
included questions. The addition of one question from Spector (2002)
added an assessment of benefit satisfaction, a topic missing in the
Price and Mueller (2002) scale. The Niehoff and Moorman (2002) six-item
formal procedures assessment measured procedural justice. Also taken
from the Niehoff and Moorman (2002) scale were five of nine questions
assessing interactive justice. The four excluded questions from the
Niehoff and Moorman (2002) scale assessed topics implied in the five
included questions.
Questionnaire Design. According to Podsakoff et al. (2003), the
arrangement and format of questions on a questionnaire might induce
common method effects when questionnaires include ambiguous questions
requiring rater interpretations or use similar scaling anchors. To
reduce ambiguity of the question phrasing, three factory employees, and
one technical and one administrative employee reviewed and recommended
changes to the questionnaire prior to issuance. The questionnaire also
used a sevenpoint Likert-type scale with two differently phrased
anchors. Never (1) and always (7) anchored the OCB intentions scales.
Strongly disagree (1) and strongly agree (7) anchored the other scales.
Each scale used a numeric value of four as the neutral rating. To
interrupt rater tendencies to fall into patterned responses, the
questionnaire included two reversed scored items (see Appendix A).
Consistent with Podsakoff et al. (2003), the questionnaire in this
study used grouped constructs instead of counterbalancing construct
questions. According to Podsakoff et al. (2003), grouping constructs
avoided disrupting the logical flow of the questionnaire. The
questionnaire assessed, in order, OCB, distributive justice, POS,
procedural justice, and interactive justice. Research by Spector and
Michaels (1983) found that the order of constructs within a single
self-rated survey did not alter the resuits.
RESULTS
Data Screening
Data screening insures that data conform to the assumptions
required for statistical tests (Tabachnick and Fidell, 2001). In this
study, data screening proceeded by first assessing missing data, then
examining normality, mitigation of outliers, and, finally, testing for
homogeneity of variance and multicollinearity. Cronbach's alpha and
factor analysis assessed construct validity and reliability (Kline,
2005). Appendices A and B lists the results for principal components
factor analysis with varimax rotation. Aggregated into single variable
scores were questions associated with each variable dimension with
factor scores in excess of .55 (considered good by Tabachnick and Fidell
(2001)). Appendices A and B highlight the factor scores for questions
included in each aggregated variable score.
Cronbach's alpha indicated sufficient reliability, defined as
scores in excess of [alpha] = .70 (George and Mallery, 2004), and
supported aggregation into a single variable score for POS ([alpha] =
.92), distributive justice ([alpha] = .89), procedural justice ([alpha]
= .82), and interactive justice ([alpha] = .95). Though Cronbach's
alpha indicated questionable reliability for OCB intentions directed at
peers ([alpha] = .70) and OCB intentions directed at the organization
([alpha] = .67), each were aggregated separately into two measurement
scores, one representing OCB intentions directed at peers and the other
representing OCB intentions directed at the organization. Table 1
displays descriptive statistics, Cronbach's alpha, and bivariate
correlations for POS, justice dimensions, and OCB intentions directed at
peers and at the organization.
The bivariate correlation between interactive and procedural
justice measured r = .609. According to George and Mallery (2004),
bivariate correlations in excess of r = .50 indicate high linear
dependency. Kline (2005) argued that linearly dependent variables create
instability in regression analysis as linearly dependent variables pose
as proxies for one another. Instead of interactive justice or procedural
justice replacing one another, assessment of variable relationships
suggested a separate analysis of the effects of interactive justice and
procedural justice on the study variables.
Two employee groupings, classified as office and unionized factory,
provided data. Means, aggregated by office and factory groupings,
exhibited statistically significant differences for POS ([t.sub.(91)] =
3.37, p = .001, two-tailed), interactive justice [t.sub.(91)] = 2.98, p
= .004, two-tailed), and OCB intentions directed at the organization
[t.sub.(91)] = 4.33, p = .000, two-tailed). Discriminant analysis,
however, failed to support aggregation of data into distinct office and
factory clusters, since clusters based on office and factory
classifications correctly predicted group membership in only 70% of the
original grouped cases (Garson, 2006). Instead, data analysis used a
single aggregated grouping. Table 2 displays descriptive statistics for
POS, justice dimensions, and OCB intentions directed at peers and at the
organization, aggregated by office and union factory groupings.
Variable Relationships and Hypothesis Testing
In support of Hypothesis 1, this study found a significant, though
modest correlation, between the level of POS and employee self-reported
intentions to enact OCB directed at the organization (r = .387, p <
.01). Supporting Hypothesis 2, there was no significant correlation
between POS and OCB intentions directed at peers (r = .021, p > .05).
The findings also indicated a significant correlation between OCB
intentions directed at the organization and OCB intentions directed at
peers (r = .341, p> .01).
Bivariate correlation analysis (Table 1) indicated no significant
correlation between distributive justice and OCB intentions directed at
the organization (r = .120, p > .05), and modest correlations between
OCB intentions directed at the organization and interactive justice (r =
.248, p < .05), and between OCB intentions directed at the
organization and procedural justice (r = .244, p < .05). No
significant correlations existed between OCB intentions directed at
peers and POS, distributive, procedural, or interactive justice. The
findings also revealed a statistically significant correlation between
POS and distributive justice (r = .341, p < .01), and strong
correlations between POS and interactive justice (r = .732, p < .01)
and between POS and procedural justice (r = .672, p < .01).
Mediating Effects of Perceived Organizational Support (POS). The
findings indicated that POS fully mediated the relationships between
procedural justice and OCB intentions directed at the organization,
supporting Hypothesis 3. Applying the methodology outlined by Baron and
Kenny (1986), correlations between OCB intentions directed at the
organization, POS, and procedural justice satisfied the first three of
Baron and Kenny's requirements. Multiple regression analysis
confirmed the fourth Baron and Kenny requirement for mediation, as the
standardized beta weight of procedural justice ([beta] = -.029, p =
.824) reduced to non-significance when POS was entered into the
regression equation with OCB intentions directed at the organization as
the dependent variable.
Unexpectedly, the findings indicated that POS fully mediated the
relationships between interactive justice and OCB intentions directed at
the organization, rejecting Hypothesis 4. Correlations between OCB
intentions directed at the organization, POS, and interactive justice
satisfied the first three of Baron and Kenny's (1986) requirements.
Fulfilling the fourth Baron and Kenny requirement for mediation,
multiple regression reduced the standardized beta weight of interactive
justice ([beta] = -.076, p = .597) to non-significance when POS was
entered into the regression equation with interactive justice as the
independent variable and OCB intentions directed at the organization as
the dependent variable.
Distributive justice also exhibited a bivariate correlation with
POS (Table 1). However, hierarchal regression analysis (step 1)
indicated that interactive justice explained 54% of the variation in POS
([beta] = .732, p = .000). Marginally non-significant was the
incremental additive effect of distributive justice (step 2) to the
regression equation including interactive justice ([beta] = .144, p =
.053), thus rejecting Hypothesis 5 in part. The incremental predictive
value of distributive justice explained less than 2% of the variance in
POS scores. Hierarchal regression analysis (step 1) also found that
procedural justice explained 45% of the variation in POS ([beta] = .672,
p = .000). Unlike interactive justice, the results indicated that
distributive justice (step 2) had a small but statistically significant
incremental additive effect (13 = .185, p = .021), explaining an
additional 3.2% of the variance in POS, in partial support of Hypothesis
5.
When controlling for POS, the statistically significant bivariate
correlation between distributive and interactive justice (r = .285, p
< .01) was reduced to non-significance (r = .056, p = .599).
Similarly, the statistically significant bivariate correlation between
distributive and procedural justice (r = .250, p < .05), when
controlling for POS, also reduced to non-significance (r = .030, p =
.778). In contrast, the strong correlation between interactive and
procedural justice (r = .609, p < .01) remained significant even
after controlling for POS (r = .232, p = .026). While the three-way
correlation between distributive, procedural, and interactive justice
was consistent with Colquitt et al. (2001), the findings indicated that
POS fully mediated the relationship between procedural and distributive
justice, between interactive and distributive justice, and partially
mediated the relationship between procedural and interactive justice.
Summary of Variable Interrelationships
The findings of this study indicated a modest reciprocal
relationship between intentions to enact OCB directed at the
organization and intentions to enact OCB directed at peers, suggesting
that, to a limited degree, intentions reported by individuals on one
form of OCB also reported similar intentions of the other form of OCB.
Though OCB intentions directed at peers varied with OCB intentions
directed at the organization, the organizational elements of support or
justice failed to explain a statistically significant variation in
scores of OCB intentions directed at peers. Perceived organization
support (POS) explained a modest 15% of the variation in the scores of
OCB intentions directed at the organization. Perceived organization
support (POS) fully mediated the relationship between procedural and
interactive justice, and OCB intentions directed at the organization.
Distributive justice failed to explain any significant variation in
self:reported OCB intentions directed at the organization.
Procedural, interactive, and distributive justice predicted POS.
The linearly dependent procedural and interactive scales offered the
greatest predictive power, explaining 55% and 49%, respectively, of the
variation in POS. Distributive justice marginally failed as a
significant predictor of POS when controlling for interactive justice,
explaining less than 2% of the variation. Distributive justice did
explain a small but statistically significant 3.2% of the variation in
POS scores when controlling for procedural justice. Moreover, POS, in
part or in whole, mediated the relationships between distributive,
procedural, and interactive justice dimensions.
DISCUSSION
Three significant findings resulted from this study. The first was
the significant, though modest, correlation between POS and OCB
intentions directed at the organization. The second finding was that POS
fully mediated the relationships between perceptions of justice and OCB
intentions directed at the organization. Finally, the study uncovered
the mediating effect of POS on the relationships between procedural,
distributive, and interactive justice.
Reciprocating Perceived Organizational Support (POS)
The findings supported the presence of a mutual exchange
relationship where employees expressed intentions to reciprocate POS
through OCB directed at the organization. The failure of OCB intentions
directed at peers to correlate with POS or justice dimensions adds
support for reciprocity, as POS and justice were organizational-level
constructs not malleable by peers (Kaufman et al., 2001). These findings
suggest that when organizations provide POS, employees signal an
intention to reciprocate through OCB directed at the organization.
Leaders seeking to influence employee intentions to enact OCB directed
at the organization might consider interventions intended to improve
employee perceptions of POS. This study offers hope that POS can
overcome fears of exploitation, lead to trust in the organization, and
prompt employees to risk reciprocation through acts beneficial to the
organization.
However, intentions to enact OCB might not lead to actual OCB
(Organ et al., 2005). Various workplace features--including leadership,
tightly coupled processes, work rules, job and work group
characteristics, organizational context, morale, motivational
conditions, along with fairness in social exchange--evoke or restrain
actual enactment of intended OCB (Organ et al., 2005). Moreover, this
study's failure to find a significant correlation between OCB
intentions directed at peers coupled with the modest correlation between
OCB intentions directed at the organization and POS suggest that a
variety of factors beyond those included in this study might influence
variations in OCB scores. While the findings of this study provide
support for a statistically significant relationship between POS and
OCB, the modest correlations cannot dismiss conjectures that OCB might
occur without expectations for organizational reciprocation.
Mediating Effects of POS on Justice and OCB
The findings of this study demonstrated that improvements in
procedural or distributive justice which enhance POS correlate with
increased OCB intentions directed at the organization. While
improvements in procedural or distributive justice -such as
strengthening disciplinary procedures, performance requirements, or more
closely linking pay with performance--might improve decision
implementation, consistency, and perceived equity, only policies
conducive to POS create an employee perceived environment of respect,
dignity, and a commitment to well-being (Coyle-Shapiro et al., 2004).
Justice interventions that embitter or disenfranchise employees, even if
changes in policies promote equity and consistency, may not be conducive
to increased POS and higher reported levels of intended OCB directed at
the organization.
Mediating Effects of POS on Justice Dimensions
The mediating effect of POS between dimensions of justice suggests
that a change in one dimension of justice affecting POS might also
influence changes in other dimensions of organizational justice. These
findings indicate a potential leveraging effect between changes in
procedural and distributive justice on POS and interactive justice. In
other words, managers attempting to treat employees with respect or
explain decisions might still receive low interactive justice scores if
guided by flawed decision making (procedural justice) or poor resource
allocation procedures (distributed justice). Similarly, employees might
sense a lack of organization support when policies and procedures
inconsistently reward or punish employees, or misallocate resources.
Without policies and procedures insuring equitable treatment, employees
might receive mixed signals concerning organizational concern for
employee well-being, commitment, and support. In sum, the findings
support an assertion that improvements in POS, perceptions of
interactive justice, and increased OCB intentions directed at the
organization might follow from increased procedural and distributive
justice.
Recommendations for Future Research
In this study, office and union-factory employees exhibited
statistically significant differences in reported level of OCB
intentions directed at the organization, POS, and interactive justice.
Important differences might result when union negotiated wage and
work-rules influence reward and resource allocations (distributive
justice), and workplace policies and grievance resolution (procedural
justice). There is a need for research in environments where the target
for reciprocation of supportive treatment is ambiguous, as might occur
when both the union and the organization influence employee pay and
benefits, work policies and procedures, and employee-supervisor
relationships.
The overwhelming balance of research on OCB reflects
managerially-oriented concerns for performance, customer satisfaction,
levels of effort, and efficient use of resources (Borman, 2004; Organ et
al., 2005; Podsakoff and MacKenzie, 1997). Consistent with this
managerial focus was this study's search for antecedents capable of
leveraging from employees increased OCB intentions beneficial to the
organization. However, attempts to capture, classify, divide, and
harness OCB might reduce employee autonomy or increase work intensity.
There is a need for research to test the implicit assumption that
amplification of OCB is beneficial to both employees and the
organization.
Limitations
Measurement of data at a single point in time and the use of
correlation analysis limit any suggested causality between variables.
Limiting generalization from the findings in this study was the small
sample size, use of a single organizational site, and a survey
instrument limited to self-reported intentions. The use of a self rated
instrument, completed in a single session, posed a threat of common
method bias (Podsakoff et al., 2003). However, the failure to find
correlations between all variables argues against significant common
method bias in this study. Other potential biases threatening validity
included impression management, consistency bias and illusionary
relationships, social desirability, and negative and positive
affectivity. There is a need for additional research employing controls
for various potential biasing effects. There is also a need for
additional research that explores relationships between self-reported
intentions to OCB and actual enacted behaviors, while controlling for
observer biases, tendencies to define in-role and extra-role behaviors
differently, and organizational inhibitors preventing realization of
intended OCB.
Further, this study also limited the search for antecedents to a
set of predefined predictors. Rioux and Penner (2001) argued that OCB
were voluntary acts that satisfied certain needs or motivations unique
to the individual, and were independent of organizational influences. If
different individuals performed the same set of behaviors for different
reasons, then a priori determination of predictors might mask underlying
employee motives for enacting volitional behaviors.
Conclusion
This study was an assessment of organizationally controllable
factors of perceived organizational support (POS) and perceptions of
justice on employee self-reported intentions to enact organizational
citizenship behaviors (OCB). The findings offered evidence that
employees reported that intended OCB was one method to reciprocate POS.
Perceived organizational support (POS) also mediated the effect of
organizational justice on OCB intentions directed at the organization,
and interrelationships between distributive, interactive, and procedural
justice. The strong positive correlations between procedural justice,
interacti