Upload
jaewon-lee
View
219
Download
2
Embed Size (px)
Citation preview
Perceived organizational support andaffective commitment: the mediating roleof organization-based self-esteem in thecontext of job insecurity
JAEWON LEE1 AND RICCARDO PECCEI2*1Kyung Hee University, School of Business, 1 Hoegi-dong, Dongdaemun-gu, Seoul, Korea2Department of Management, King’s College London, London, U.K.
Summary Research on the relationship between perceived organizational support (POS) and affectiveorganizational commitment (AC) has primarily adopted a social exchange perspective. In thisstudy we considered complementary socio-emotional explanations of the POS–AC relation-ship. We focused on the mediating role of organization-based self-esteem (OBSE) and testedcompeting models of the POS–OBSE-AC relationship separately on data from two Koreanbanks that experienced different levels of downsizing following the 1997 Korean financialcrisis. We further extended the analysis by examining the extent to which the relationshipbetween POS and OBSE and AC, respectively, was affected by employees’ perceived sense ofjob insecurity in the two banks. The results showed that OBSE was a significant mediator of thePOS–AC relationship in both organizations. POS, however, also retained a strong independentdirect effect on commitment. Moreover, as expected, perceived job insecurity tended toattenuate the POS–OBSE relationship, but augmented the direct relationship between perceivedorganizational support and affective commitment. Copyright# 2006 John Wiley & Sons, Ltd.
Introduction
Perceived organizational support (POS) is commonly defined in the literature as the extent to which
individuals believe that their employing organization values their contributions and cares for their
well-being (Eisenberger, Huntington, Hutchison, & Sowa, 1986; Shore & Shore, 1995). Perceived
organizational support has been shown to be related to a range of positive employee attitudes and
behaviors at work, including, for example, job satisfaction, organizational commitment, various forms of
citizenship and discretionary behavior, attendance and intention to stay (Rhoades & Eisenberger, 2002).
As a result, POS has attracted increasing interest among both organizational scholars and practitioners in
recent years (Rhoades & Eisenberger, 2002). Despite the acknowledged importance of POS to outcomes
that are favorable to both employees and the organization, however, our understanding of the processes
that underlie these various associations is still limited. The specific mechanisms involved are likely to
Journal of Organizational Behavior
J. Organiz. Behav. 28, 661–685 (2007)
Published online 7 December 2006 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/job.431
*Correspondence to: Riccardo Peccei, Department of Management, King’s College London, 150 Stamford Street, London, SE19NN, U.K. E-mail: [email protected]
Copyright # 2006 John Wiley & Sons, Ltd.
Received 12 April 2006Revised 18 September 2006Accepted 2 October 2006
vary depending on the outcome under consideration. Our focus in this study is on the link between
perceived organizational support and employee affective commitment (AC) to the organization.
AC refers to the relative strength of employees’ emotional attachment to, identification with and
involvement in their employing organization (Meyer & Allen, 1991; Mowday, Porter, & Steers, 1982).
The reason for focusing on the POS–AC link is that AC is a key indicator of employee integration at work
which, because of its reported positive effects on important organizational phenomena, such as turnover,
voluntary absenteeism, and various aspects of both in-role and discretionary performance, is of central
interest to organizational scholars and practitioners alike (Mathieu & Zajac, 1990; Meyer & Allen, 1997;
Meyer, Stanley, Herscovitch, & Topolnytsky, 2002; Mowday et al., 1982).
Two main explanations of the link between perceived organizational support and AC have been
advanced in the literature. The first is in terms of reciprocity and social exchange. According to this
view, employees who believe that their organization values them and cares for their well-being are
more likely to feel a sense of obligation toward the organization and, therefore, to reciprocate the
favorable treatment with increased loyalty and commitment. The second explanation is in terms of
socio-emotional needs. This is the idea that perceived organizational support can help to fulfill
important socio-emotional needs of individuals, such as needs for approval, esteem and affiliation, and
that such need fulfillment, in turn, enhances employees’ affective attachment to and identification with
the organization (Eisenberger et al., 1986; Meyer & Allen, 1997).
Either explicitly or implicitly, the social exchange approach has been the dominant framework
underpinning much of the research dealing not only with the specific link between perceived
organizational support and AC, but also with the antecedents of organizational commitment more
generally (Rhoades & Eisenberger, 2002; Meyer & Allen, 1997; Mottaz, 1988). Socio-emotional
explanations have received far less attention both in the general commitment literature and in studies
dealing specifically with the POS–AC relationship. In this study we seek to address this gap by focusing on
organization-based self-esteem (OBSE) and exploring the role that this potentially key socio-emotional
mechanism plays in mediating the relationship between perceived organizational support and AC.We test
for the mediating role of OBSE using employee survey data from two Korean banks that experienced
varying levels of job insecurity following Korea’s major financial crisis in 1997.
As we explain more fully below, in line with both social exchange and socio-emotional arguments, job
insecurity can be expected to have amoderating effect on the pattern of relationships between POS, OBSE,
and AC. A systematic consideration of the potential moderating effects of job insecurity can, therefore,
contribute to a fuller understanding of the conditions under which key relationships between POS, OBSE,
and AC are more likely to hold. Hence, in a second step of the analysis, we explicitly focus on employees’
perceived level of job insecurity as a potential moderator of the observed links between POS, OBSE, and
AC. Specifically, in linewith socio-emotional and social exchange arguments, we first present a number of
hypotheses about the moderating effects of job insecurity on the links between POS and OBSE, and POS
and AC, respectively, and then test these hypotheses using data from the two case study banks.
Theoretical Framework and Hypotheses
Social exchange and felt obligation
The most common explanation of the relationship between perceived organizational support and AC is
in terms of reciprocity and social exchange. Central to this explanation is the idea that, for a variety of
reasons, employees tend to ascribe human-like characteristics to the organization (Levinson, 1965) and
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
662 J. LEE AND R. PECCEI
that, in line with the norm of reciprocity (Gouldner, 1960), favorable treatment by the organization
creates a sense of indebtedness among employees to the personified organization (Eisenberger et al.,
1986). This sense of indebtedness, in turn, becomes the basis for a stronger affective attachment to the
organization. Specifically, organizational support theory (Eisenberger, Cummings, Armeli, & Lynch,
1997; Eisenberger et al., 1986; Shore & Shore, 1995) suggests that perceived organizational support
increases AC by creating a felt obligation to care about the organization and help it meet its objectives.
Employees who perceive the organization as caring for their well-being are, therefore, assumed to be
more likely to reciprocate not only by engaging in various forms of prosocial behavior directed toward
the organization, but also by developing a stronger AC to the organization.
As noted by Eisenberger, Armeli, Rexwinkel, Lynch, and Rhoades (2001), the felt obligation to
reciprocate caring with caring, in both behavioral and attitudinal terms, can be expected to be more
pronounced among employees who have a stronger exchange ideology. It can also be expected to be
more marked where the organization is thought to provide valued support on a voluntary basis, rather
than out of necessity or because it is forced to do so by outside agencies (Eisenberger et al., 1986;
Eisenberg et al., 1986, 1997). In the former situation perceived organizational support can be expected
to have a stronger effect on AC because its impact on felt obligation is likely to be correspondingly
more pronounced. Felt obligation, however, remains the core factor through which POS is assumed to
affect AC. In other words, dominant social exchange explanations suggest that felt obligation, via the
norm of reciprocity, is a key mechanism mediating the relationship between perceived organizational
support and AC.
The fulfillment of socio-emotional needs
An alternative, but complementary, explanation of the POS–AC link focuses on socio-emotional needs.
The standard argument in this respect is that advanced by Eisenberger and his colleagues in their
original 1986 article on perceived organizational support. This posits that to the extent that perceived
organizational support meets important socio-emotional needs for esteem, approval and affiliation, ‘the
employee would incorporate organizational membership into self-identity and thereby develop a
positive emotional bond (affective attachment) to the organization’ (Eisenberger et al., 1986: 501).
With minor variations, this formulation has been repeated in most subsequent references to the
socio-emotional approach in the POS literature (e.g., Armeli, Eisenberger, Fasolo, & Lynch, 1998;
Eisenberger, Fasolo, & Davis-LaMastro, 1990; Rhoades & Eisenberger, 2002; Rhoades, Eisenberger,
& Armeli, 2001).
Although not always systematically articulated, central to the socio-emotional argument is the idea
that perceived organizational support helps to meet important individual needs for esteem and approval
at work. Because it is explicitly associated with the personified organization in the mind of employees,
this need fulfillment, in turn, leads to the development of a stronger affective attachment to the
organization. Specifically, as argued by Meyer and Allen (1991, 1997), the satisfaction of important
higher-order needs by the organization through the provision of valued rewards and support is
experienced as psychologically fulfilling by employees. And it is precisely these psychologically
rewarding experiences that are at the basis of the development of AC. This is in line with the associative
theory of attitude formation and with the tenets of classical conditioning (Byrne & Clore, 1970;
Zimbardo & Leippe, 1991). As noted both by Meyer and Allen (1997) and by Eisenberger and his
colleagues (2001), repeated association of psychologically fulfilling experiences with the organization
can be expected to result in a positive emotional link of the individual to the organization whereby the
individual not only develops a strong affective attachment to the organization, but also identifies with it
and internalizes its goals and values. The development of this emotional attachment can be quite
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
PERCEIVED ORGANIZATIONAL SUPPORT AND AFFECTIVE COMMITMENT 663
unconscious (Meyer & Allen, 1997). However, to the extent that perceived organizational support
fulfills important socio-emotional needs, it can be expected to have a positive impact on AC. In this
case, though, the impact of perceived organizational support is not assumed to be mediated by felt
obligation, but rather, by the satisfaction of socio-emotional needs reflected, for example, in a
heightened sense of self-worth and self-esteem by employees.
To our knowledge, the socio-emotional argument outlined above has not been systematically tested by
researchers. It has only been examined indirectly in a recent study by Eisenberger and his colleagues (2001)
that compared the relative importance of felt obligation and of positive mood as potential mediators of the
POS–AC relationship among a sample ofUS postal workers. The results of this study indicated that, although
the path through felt obligation was stronger than the one via positive mood, there were additional
mechanisms through which perceived organizational support was likely to have an impact on AC. Included
among thesemechanisms, as the researchers themselves note, are ones directly related to the socio-emotional
perspective, such as POS’ enhancement of organizational identification through its fulfillment of key
individual needs for esteem and affiliation at the workplace (Eisenberger et al., 2001: 49–50).
In summary, the existing evidence suggests that felt obligation only partially mediates the impact of
perceived organizational support on AC and that socio-emotional factors may well be important in
explaining the relationship between POS and AC. However, the mediating role of socio-emotional
variables in the POS–AC relationship has not been explicitly examined to date and, therefore, is the
central focus of the present study.
The role of organization-based self-esteem
As part of the socio-emotional perspective, there are a number of psychological needs and/or feelings
that, as we have seen, have been identified as potentially important mediators of the relationship between
perceived organizational support and AC. These include, for example, esteem, self-worth, approval,
respect, and affiliation. To a degree, these various feelings and needs are likely to be interrelated
(Brockner, 1988). In practice, therefore, they may be difficult to distinguish clearly from one another.
While acknowledging this potential overlap, in the present study we focus on self-esteem as one of the
core contributors to individuals’ sense of psychological well-being. Specifically, self-esteem refers to the
‘self-evaluation that individuals make with regard to themselves’ (Pierce, Gardner, Cummings, &
Dunham, 1989: 625). It expresses their attitude of approval or disapproval toward themselves, their
overall assessment of their own value as a person. In other words, self-esteem reflects individuals’
personal judgment and belief about their own worthiness and, as such, serves a crucial anxiety-buffering
function (Greenberg et al., 1999). Hence the importance of self-esteem which is viewed as a fundamental
requirement of a healthy consciousness (Locke, McClear, & Knight, 1996), or, as Baumeister (1995) has
argued, as one of the main factors that make life meaningful, regardless of cultural differences.
Different forms of self-esteem have been identified in the literature including, for example, global,
organization, role and task or situation-specific self-esteem (Pierce et al., 1989). Our focus is on OBSE
defined as ‘the degree to which organizational members believe that they can satisfy their needs by
participating in roles within the context of the organization’ (Pierce et al., 1989: 625). In other words, OBSE
reflects employees’ evaluation of their personal adequacy and worthiness as organizational members so that,
as noted by Pierce and his colleagues (1989), individuals with high OBSE perceive themselves as important,
meaningful, and worthwhile within their employing organization. The reason for focusing on OBSE, rather
than on other forms of self-esteem, is that the principle of compatibility suggests that attitudes are likely to be
more closely related to other attitudinal or behavioral variables of interest when they are framed at a level that
is similar to that of these other variables (Ajzen & Fishbein, 1977). Our interest here is in examining the role
of self-esteem as a potential mediator of the relationship between perceived organizational support and AC.
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
664 J. LEE AND R. PECCEI
The point of reference of both POS and AC is the organization. The appropriate referent for self-esteem in
this case, therefore, is also the organization. Hence the specific focus on OBSE.
Underpinning the analysis of OBSE as a mediator of the POS–AC relationship are two key
arguments. The first, based on symbolic interactionism, concerns the link between perceived
organizational support and OBSE. Central to symbolic interactionism is the idea that individuals,
through a variety of processes, view themselves through how they are viewed by others (Blumer, 1969;
Brockner, 1988; Hewitt, 1997). Thus, as individuals interact with other people and entities that
encourage certain systems of beliefs about the self, it becomes more likely that they will internalize
those beliefs and that others’ views will become a part of the self (Korman, 1971). Particularly
important in this respect are the views and reactions of significant others, of people and entities that
individuals respect and regard as important (Hewitt, 1997). If significant others hold the self in high
regard, one’s own sense of self-esteem will be high (Harter, 1993). Within an organizational context,
the organization itself may be an important significant other for employees (Van Dyne, Vandewalle,
Kostova, Latham, & Cummings, 2000). Thus, the appraisal of the organization, reflected in employees’
perceived sense of organizational support, may play a key role in enhancing individuals’ sense of
OBSE. Consistent with symbolic interactionist arguments, therefore, to the extent that employees
perceive the organization as valuing their contributions and caring for their well-being, they can be
expected to develop a stronger sense of organization-based self esteem.
The second argument concerns the link between OBSE and AC. Consistent with the arguments of the
socio-emotional perspective outlined above, a heightened sense of OBSE is likely to be
psychologically satisfying to employees. To the extent that, in the minds of individuals, this
psychological satisfaction is associated with the organization through perceived organizational
support, OBSE can be expected to enhance their AC to the organization. In line with Meyer and Allen’s
(1997) arguments and with the tenets of classical conditioning and the associative theory of attitude
formation (Byrne &Clore, 1970; Zimbardo&Leippe, 1991) highlighted above, therefore, a heightened
sense of POS–related OBSE is expected to facilitate the development of AC.
In this context it is worth noting that there may also be other mechanisms underpinning the link
between OBSE and AC. These include some of the self-regulatory mechanisms reviewed by Pierce and
Gardner (2004), such as the self-consistency motivations identified by Korman (1970). Self-
consistency arguments suggest that individuals who have a positive image of themselves will tend to
adopt attitudes that reinforce that positive image (Korman, 1970). On this basis, therefore, employees
who enjoy a stronger sense of self-esteem related to their perception of themselves as capable,
significant and worthy organizational members, can be expected to exhibit a stronger sense of AC to
their employing organization than their low OBSE counterparts. Irrespective of the specific
mechanisms involved, however, empirical work in the area suggests a positive link between OBSE and
AC (Hui & Lee, 2000; Pierce et al., 1989; Van Dyne & Pierce, 2004). This positive link has been
confirmed by Pierce and Gardner’s (2004) recent comprehensive review of OBSE research.
Based on the above arguments and evidence, we hypothesize that OBSE will mediate the relationship
between perceived organizational support and AC. Specifically, we hypothesize that POS will have a
positive impact on OBSE, which will, in turn, have a positive impact on AC. As noted earlier, however,
there are other factors, apart from OBSE, that are likely to help to account for the positive link between
POS and AC. These include not only feelings of reciprocity and a sense of felt obligation, but also other
non-OBSE related higher-order socio-emotional feelings and needs. To the extent that these other
mechanisms are operative and effective, perceived organizational support can be expected to have a
separate additional impact on AC which is not necessarily mediated by OBSE. On this basis, therefore,
we hypothesize that POS will have both a direct and indirect positive effect on AC. Specifically, we
propose the following overall partialmediation hypothesis, which is summarized in themodel in Figure 1.
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
PERCEIVED ORGANIZATIONAL SUPPORT AND AFFECTIVE COMMITMENT 665
Hypothesis 1: The effect of perceived organizational support on affective commitment will be
partially mediated by organization-based self-esteem, so that:
(a) POS will have a positive effect on OBSE,
(b) OBSE will have a positive effect on AC, and
(c) POS will have a direct positive effect on AC above and beyond its indirect effect through OBSE.
The moderating role of perceived job insecurity
The hypothesized links between POS, OBSE, and AC shown in Figure 1 can be expected to be
moderated by a number of factors. As noted above, research by Eisenberger and his colleagues
(Eisenberger et al., 1986, 1997, 2001) suggests, for example, that the impact of perceived organizational
support on AC is moderated by individuals’ exchange ideology and by the extent to which employees
perceive that the relevant support is provided voluntarily by the organization. Here we extend this
moderator analysis by looking at the extent to which the hypothesized relationships between POS,
OBSE, and AC are affected by employees’ perceptions of their level of job insecurity. We focus on the
potential moderator role of perceived job insecurity for two reasons. First, as we explain more fully
below, both of the banks covered in the present study experienced significant turbulence and uncertainty
as a result of the 1997 Korean financial crisis. In response to its more precarious economic and financial
situation, one of the banks instituted a more extensive program of downsizing. Although after the
downsizing programs, at the time the research was conducted, levels of job insecurity varied
considerably across both banks, they continued to remain high, especially in the organization that had
experienced the more drastic job cuts. In other words, job uncertainty formed an important part of the
broader organizational context experienced by employees at the two case study banks.
Second, research on downsizing suggests that the turbulence and uncertainty associated with major
programs of job cuts tend, by and large, to have a negative effect on employee work attitudes and behavior
(Luthans & Sommer, 1999; Mone, 1994; Wagar, 1998). On this basis, job insecurity can be expected to be
negatively related to all three of the main variables in our model, including POS, OBSE, and AC. More
fundamentally, however, in line with the social exchange and socio-emotional arguments outlined above,
job insecurity can be expected tomoderate both the POS–AC relationship and the POS–OBSE relationship
in the model. Specifically, in terms of the POS–OBSE relationship, the symbolic interactionist arguments
outlined above suggest that job insecurity is likely to undermine and counteract the positive effect of POS
on individuals’ sense of OBSE. Employees who are unsure of their future in the organization may, in fact,
find it more difficult than employees who enjoy a greater sense of job security to form a clear view not only
of the extent to which the organization does indeed care for their well-being, but also of the reasons why
they may be receiving preferential treatment, and whether this treatment is likely to continue in the future
(Luthans & Sommer, 1999; Mone, 1994). Hence our first moderator argument suggests that perceived job
insecurity is likely to attenuate (negatively moderate) the POS–OBSE relationship.
H1(c): (+)
H1(a): (+) H1(b): (+) PerceivedOrganizationalSupport
Organization-basedSelf-esteem
AffectiveCommitment
Figure 1. Hypothesized partial mediation model of the POS–OBSE-AC relationship
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
666 J. LEE AND R. PECCEI
Hypothesis 2: The greater the sense of job insecurity, the less pronounced the positive association
between perceived organizational support and organization-based self-esteem will be.
In contrast, the moderating effect of job insecurity on the POS–AC relationship in our basic research
model is less clear since it is possible to identify two contrasting ways in which job insecurity may affect
the impact of perceived organizational support on AC. The first, more straightforward possibility is that
job insecurity has a negative moderating effect on (i.e., attenuates) the direct relationship between POS
and AC. In linewith social exchange arguments, individuals whose job in the organization is insecure and
uncertain, are less likely to feel a sense of obligation to the organization itself. As such, job insecurity is
likely to undermine any sense of POS–related felt obligation that individuals may have toward the
organization and, consequently, can be expected to attenuate the positive effect of POS on AC.
Hypothesis 3a: The greater the sense of job insecurity, the less pronounced the positive association
between perceived organizational support and affective commitment will be.
The second, more complex possibility is that job insecurity has a positive, rather than a negative,
moderating effect on (i.e., augments) the direct relationship between POS and AC in our basic research
model. Specifically, job insecurity can be expected to affect the very nature of any social exchange that
may underpin the POS–AC relationship. To an important extent, the terms of this exchange depend on
the position of the individual in relation to the organization. Employees who face greater job insecurity
and, therefore, a more uncertain future in the organization, tend, on the whole, to be in a weaker
position vis-a-vis the organization than employees whose jobs are more secure. As a result, employees
whose jobs are more insecure are likely to ascribe greater importance to any additional signs of support
they receive from the organization than are employees who enjoy greater security of employment. They
are also likely to be more ready to reciprocate any act of goodwill from the organization as a basis for
improving their employment chances and reducing their uncertainty. Consequently, employees who are
more insecure in their jobs are likely to react more positively to positive treatment by the organization
than employees who enjoy a more secure position. Hence, our alternative moderator argument
suggests that, to the extent that the relationship between POS and AC is underpinned by a sense of
felt obligations, then job insecurity can be expected to augment (positively moderate) this
relationship.
Hypothesis 3b: The greater the sense of job insecurity, the more pronounced the positive association
between perceived organizational support and affective commitment will be.
The above moderator hypotheses1 are shown schematically in Figure 2. These hypo-
theses, together with the main partial mediation hypothesis outlined above, were
1Although not of direct concern in the present analysis, it is worth noting that job insecurity is not expected to have a significanteffect on the third main relationship in our research model, namely that between OBSE and AC. This expectation is in line withrecent research by Hui and Lee (2000). As part of their study, Hui and Lee (2000) examined the moderating effect of OBSE on therelationship between job insecurity and organizational commitment. They found a significant negative relationship between jobinsecurity and commitment, as well as a significant positive relationship between OBSE and commitment. The job insecurity xOBSE interaction, however, was not significant in their analysis, thereby suggesting that, consonant with our own expectations,OBSE tends to have a consistent positive effect on organizational commitment that does not vary depending on employees’perceived level of job insecurity. This conclusion found confirmation in our own data. As expected, in additional analysisperceived job insecurity was not found to significantly moderate the relationship between OBSE and AC in either bank. Thus, amodel that included the OBSE x job insecurity interaction term did not yield a better fit to the data in either sample than a modelthat excluded the interaction term (bank A: D x2¼ 0.35, D df¼ 1, p> 0.10; bank B: D x2¼ 1.17, D df¼ 1, p> 0.10). Moreover,the OBSE x perceived job insecurity term was not significantly related to AC in either sample (bank A: b¼ 0.01, p> 0.10; bankB: b¼ 0 .05, p> 0.10).
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
PERCEIVED ORGANIZATIONAL SUPPORT AND AFFECTIVE COMMITMENT 667
tested using data from two independent samples of employees from two Korean banking
organizations.
Organizational Context
The study was carried out in two large Korean banks, referred to here as bank A and bank B,
respectively. Bank A was primarily a retail bank. Bank B, on the other hand, was a mixed
wholesale-retail organization that, after the 1997 financial crisis, refocused primarily on retail
activities. The two banks also shared a number of other similarities. They both were among the top
10 largest banks in Korea with a nation-wide network of local branches. Prior to the 1997 crisis,
bank A had approximately 13 500 employees and 504 branches, while bank B had 8000 employees
and 306 branches. The two banks also had similar job design and human resource practices, such as
salary and promotion systems, as well as performance appraisal and employee development
policies. Moreover, prior to the 1997 crisis, both banks were financially highly successful.
Following the 1997 financial crisis the position of the two banks became more differentiated.
Bank B was severely affected by the crisis and became insolvent in 1998 due to the successive
bankruptcies of several big Korean companies which were its main customers. In response, the
Korean government injected 4820 billion Korean Won of fresh capital into the bank in 1998. In
return, the bank had to carry out massive job cuts and salary reductions. Between 1998 and 1999 it
reduced its workforce by about 40 per cent, from 8067 to 4809 employees. The branch network was
also reduced from 306 to 291 sites. At the same time, salaries were cut by an average of 20–30 per
cent depending on employees’ position in the organization. In contrast, bank Awas not significantly
affected by the financial crisis. However, bank A also engaged in downsizing in order to strengthen
its competitiveness and reduced its workforce by about 15 per cent between 1998 and 1999, from
13 519 to 10 212 employees. But in bank A, unlike in bank B, this process of downsizing was not
accompanied by salary cuts.
H3(a): (-) H3(b): (+)
H2: (-)
PerceivedOrganizationalSupport
Organization-based Self-esteem
AffectiveCommitment
Perceived Job Insecurity
Perceived Job Insecurity
Figure 2. Hypothesized moderation of the POS–OBSE and POS–AC relationships by perceived job insecurity
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
668 J. LEE AND R. PECCEI
In both banks, these major programs of rationalization and restructuring were
accompanied by significant changes in the work environment, as well as in existing
systems of employment. In terms of the work environment, in-depth interviews with
survivors indicated that, in both banks, job pressure and workloads increased after the job
cuts since fewer employees were expected to perform the same or a greater amount of work.
At the same time, survivors also reported greater autonomy and responsibility at work since,
due to increased spans of control, supervisors could no longer monitor and supervise staff as
closely as before.
The key features of the employment system of both banks before the financial crisis were
lifetime employment, seniority-based pay and promotion. With the advent of the crisis, the
lifetime employment system broke down. In both banks, the main criterion used for making
people redundant was their merit rating, irrespective of whether they worked in headquarters or
branches, and employees perceived that future potential redundancies would also be based on
individual performance criteria. Moreover, at the time the research was conducted, both banks
planned to change their salary and promotion systems to performance-related pay and
capability-based promotion. These major prospective changes in established pay and promotion
systems added a further element of turbulence and insecurity for the workforce in both
organizations, but especially for employees in bank B due to the more precarious economic
position of this organization and the consequently higher levels of uncertainty surrounding its
future.
Method
Samples and procedures
The data used in the present study were collected as part of a larger research project designed to
examine the impact of the 1997 financial crisis and of downsizing on a range of employee attitudes
and behaviors in the two case study banks described above (Lee, 2003a,b; Lee & Corbett, 2006)2.
Separate employee surveys were conducted in the two banks between November 1999 and March
2000, after the implementation of the main programs of restructuring and downsizing in the two
organizations. Questionnaires were distributed to a sample of between 6 and 25 employees in 58
branches of bank A and 50 of bank B, as well as in the regional headquarters and in selected
departments at head office of each of the two organizations. Both banks comprised three
organizational levels—headquarters, regional headquarters, and branches. In both banks,
employment at headquarters constituted about 16 per cent of total employment; employment at
regional headquarters, about 4 per cent; and employment in the branches, about 80 per cent.
2Lee’s (2003a) first study examined the antecedents of organization-based self-esteem and tested the cross-validation of theconcept of OBSE in the Korean context, while his second study (Lee, 2003b) examined the extent to which OBSE actuallymediated the relationship between a range of positive employee work experiences and organizational citizenship behavior in thetwo banks. Finally, the third study (Lee & Corbett, 2006) examined both the direct and indirect impact of downsizing on affectivecommitment in the two banks through its impact on a range of day-to-day employee work experiences in the two case studyorganizations.
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
PERCEIVED ORGANIZATIONAL SUPPORT AND AFFECTIVE COMMITMENT 669
Questionnaires were distributed proportionately to the three levels in each bank, with the
participating sites/branches being selected randomly. The questionnaires themselves were
personally distributed at each bank site by the first author who visited each site and asked a
member of staff to distribute an appropriate number of questionnaires randomly to colleagues and
then collect them. Jaewon Lee then collected each batch of completed questionnaire a week later
from each of the participating sites.
In bank A, 545 out of 770 questionnaires were returned, making for a response rate of around 71 per
cent. The number of usable questionnaire was 4563. The number of returned questionnaires at each
level corresponded quite closely to the distribution of the population (i.e., for headquarters, 13.8 per
cent; for regional headquarters, 3.9 per cent; for branches, 82.3 per cent). Seventy one per cent of the
sample was comprised of men (population¼ 71.9 per cent), the average age of respondents was
33.9 years (population¼ 34.3) and their average level of education was 14 years. On all three
dimensions (i.e., age, gender, and the number of returned questionnaire at each level) the sample was
representative of the broader employee population in bank A.
In bank B, 480 out of 660 questionnaires were returned, equivalent to a response rate of about
72 per cent. The number of usable questionnaire was 454. Here too the number of returned
questionnaire at each level corresponded quite closely to the distribution of the population (i.e., for
headquarters, 16.3 per cent; for regional headquarters, 3.8 per cent; for branches, 80 per cent).
Seventy-two per cent of respondents in bank B were men (population¼ 73.4 per cent), the average
age of the sample was 34 years (population¼ 35.7) and the average level of education was 15 years.
On all three counts this sample also was representative of the broader employee population in
bank B.
Measures
All the main constructs included in the analysis were assessed with perceptual self-report measures
based on multi-item scales whose psychometric properties are well established. Responses to all items
were made on 7-point Likert-type scales (1¼ strongly disagree, 7¼ strongly agree). To ensure that
there was linguistic and psychometric equivalence between the English and Korean versions of the
questionnaire (Hulin & Mayer, 1986), the English version of the instrument was first translated by
Jaewon Lee and then checked independently by three Korean social scientist with a detailed knowledge
of English. To ensure that there was contextual equivalence between the Korean and English versions of
the instrument, the questionnaire was then piloted with a group of employees from the two banks. The
specific measures used in the analysis, along with sample items of the relevant constructs, are outlined
below.
Perceived organizational support (POS) was measured using Eisenberger et al.’s (1986) 17-item
scale. Much previous research (e.g., Eigenberger et al., 1986, 1990) has provided validation evidence
for this scale. A sample item was ‘The organization values my contribution to its well-being.’ OBSE
(OBSE) was assessed using 8-items from Pierce et al.’s (1989) 10-item scale (e.g., ‘I am an important
3Employees in all Korean banks periodically rotate internally to other sites/locations (for example, from departments atheadquarters to branches and from branches to departments at headquarters). Especially in bank A, many employees rotated toother branches in the interval between the distribution and collection of the questionnaires in late January 2000 and, among thequestionnaires that were distributed at that time, there were a number that were returned but left blank. Thus, some of the returnedquestionnaires were not usable, especially in bank A.
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
670 J. LEE AND R. PECCEI
part of my workplace,’ and ‘I am trusted in my workplace’)4. Validation evidence for the OBSE scale is
found in Pierce et al. (1989). Organizational affective commitment (AC) was measured using Meyer
and Allen’s (1997) 6-item Revised Affective Commitment Scale. Much previous research (e.g., Ko,
Price, & Mueller, 1997) provides validation evidence for this scale. Sample items included ‘I would be
very happy to spend the rest of my career in this bank’ and ‘I really feel as if this bank’s problems are
my own.’ Perceived job insecurity refers to the perceived likelihood of becoming unemployed (Garst,
Frese, & Molenaar, 2000). In order to measure this variable, Price’s (1991) three-item scale (that
measures perceived job security) was reversed and this reversed scale was used. A sample item
included ‘I am confident that I will be able to work for this bank as long as I wish.’
No controls were used in the main analysis. However, to check for the sensitivity of results to
different model specifications, we also ran some subsidiary analyses using a number of controls. These
controls were designed to capture basic demographic factors and individual dispositions that might
have an impact on AC as well as on perceived organizational support and OBSE (Meyer & Allen, 1997;
Pierce & Gardner, 2004; Rhoades & Eisenberger, 2002). The controls used included respondents’ age,
organizational and job tenure, gender, marital status, years of education, and positive affectivity and
negative affectivity, defined as ‘the tendency to experience pleasant and unpleasant emotions
respectively’ (Price, 1997: 435). Positive affectivity and negative affectivity were measured using
Watson, Clark, and Tellegen’s (1988) PANAS scales.
Analysis procedures
The examination of the hypothesized effects was conducted using latent variable analysis and structural
equations. The analyses (employing LISREL 8.30) were conducted in line with major
recommendations proposed in the literature (e.g., Anderson & Gerbing, 1988; Bagozzi & Yi,
1988; Hair, Anderson, Tatham, & Black, 1998; Harris & Schaubroeck, 1990; Joreskog & Sorbom,
1993). First, missing data were dealt with using mean substitution. Second, to prevent problems with
parameter estimation, the POS, OBSE, and AC scales were arbitrarily trichotomized into three
non-overlapping subscales and these subscales were then used as the manifest indicators in the analysis.
Third, in order to assign a unit of measurement for the latent variables, all latent variables were scaled
with one loading, in each case, set to 1.00. Finally, multiple goodness-of-fit indices were employed,
including the chi-square test, goodness-of-fit index (GFI), non-normed fit index (NNFI), root mean
4Organization-based self-esteem was originally assessed using Pierce et al.’s (1989) 10-item scale. However, the results ofpreliminary exploratory factor analysis showed that two items loaded on a distinct factor in both samples. As a further check,therefore, a one-factor confirmatory model where all 10 original OBSE items were constrained to load onto a single overall latentvariable was compared to a two-factor model where the two items that loaded separately in the exploratory factor analysis wereconstrained to load onto a separate latent construct from the remaining eight items. In both banks the two-factor model provided asignificantly better fit to the data than did the one-factor model (for bank A: D x2¼ 139.96, df¼ 1, p< 0.005), for bank B: Dx2¼ 168.12, df¼ 1, p< 0.005), thereby providing additional support for using the 8-item scale in the present study. The twoitems that were excluded were, ‘I can make a difference in my workplace’ and ‘I am a valuable part of my workplace’. Koreans’construal of the self may be helpful in interpreting the above results. As Inumiya, Choi, Yoon, Seo, and Han’s (1999) study hasshown, compared to Westerners, Koreans have a more interdependent construal of the self which becomes most meaningful andcomplete within the context of social relationships with others. In countries where people have such an interdependent construalof the self, emphasizing one’s own individuality and trying to stand out are regarded as inappropriate (Markus&Kitayama, 1991).Hence, statements such as ‘I can make a difference in my workplace’ and ‘I am a valuable part of my workplace,’ may well beinterpreted differently from the other OBSE items in a Korean than in a Western context where individuals have a moreindependent construal of the self. In a Korean context these items may well be interpreted as referring to a more individualisticand, therefore, potentially less harmonious notion of the self, rather than to a more integrative sense of whether the respondentsfeel that they are meaningful and important within their organization. Hence the tendency for these two items to factor analyzeseparately from the other OBSE items in both banks, and our decision, therefore, to drop them from the final OBSE scale used inthe analysis.
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
PERCEIVED ORGANIZATIONAL SUPPORT AND AFFECTIVE COMMITMENT 671
square error of approximation (RMSEA), incremental fit index (IFI), comparative fit index (CFI), and
the relative fit index (RFI).
Results
Preliminary analysis
Table 1 reports descriptive statistics, reliabilities, and zero-order correlations for the main variables
used in the analysis for banks A and B separately.
The psychometric properties of the main measures used in the analysis were examined separately in
the two banks. According to the results of the confirmatory factor analysis, a four-factor model where
the items used to assess POS, OBSE, AC, and perceived job insecurity were assumed to load onto their
corresponding latent constructs, yielded a good fit to the data in each sample [for bank A: x2 (with
48 degree of freedom)¼ 136.92 (p< 0.001), GFI¼ 0.95, NFI¼ 0.98, NNFI¼ 0.98, CFI¼ 0.99,
IFI¼ 0.99, RMSEA¼ 0.064; for bank B: x2 (with 48 degree of freedom)¼ 117.40 (p< 0.001),
GFI¼ 0.96, NFI¼ 0.98, NNFI¼ 0.98, CFI¼ 0.99, IFI¼ 0.99, RMSEA¼ 0.056]. In additional
analysis, a one-factor confirmatory model where all the items were constrained to load onto a
single overall latent variable was also tested [for bank A: x2 (with 54 degree of freedom)¼ 2392.68
(p< 0.001); for bank B: x2 (with 54 degree of freedom)¼ 2754.46 (p< 0.001)]. In both banks the
four-factor model provided a significantly better fit to the data than did the one-factor model [for bank
A: D x2¼ 2255.76, df¼ 6 (p< 0.005); for bank B: D x2¼ 2637.06, df¼ 6 (p< 0.005)]. Multi-group
analysis was also conducted to compare the equivalence of the measurement models across the two
banks. The results yielded an acceptable model fit to the data [x2 (with 126 degree of
freedom)¼ 456.92 (p< 0.001), GFI¼ 0.92, NFI¼ 0.96, NNFI¼ 0.97, CFI¼ 0.97, IFI¼ 0.97,
RMSEA¼ 0.076], thereby indicating that the measurement models across the two banks were
equivalent. In addition, as can be seen from Table 1, when scaled, all four multi-item measures
exhibited adequate levels of internal reliability in both banks, with the relevant alpha coefficients
ranging from 0.82 to 0.96.
Because this study uses self-report measures, there exists the risk of common method variance that
may inflate or deflate the strength of the associations between constructs. To check for the potential
impact of common method variance, the procedure for controlling for the effects of an unmeasured
latent methods factor recommended by Podsakoff, MacKenzie, and Podsakoff (2003) was used. That
is, the model in Figure 1 was re-estimated with all the indicators from the same source (i.e., OBSE,
POS, AC, and perceived job insecurity) loading on a general methods factor. In both banks, when
commonmethod variance was controlled for, the overall pattern of relationships between the constructs
was not affected (results available from the authors). In turn, this suggests the absence of substantial
commonmethod variance problems capable of confounding the interpretation of the substantive results
of the main statistical analysis.
At a descriptive level the results in Table 1 show that average levels of POS, OBSE, and AC differed
significantly between the two organizations. As expected, the organization that experienced a more
stable environment (i.e., bank A) exhibited significantly higher scores on all three variables than the
organization that experienced higher levels of uncertainty (i.e., bank B) (mean POS: bank A¼ 4.49,
bank B¼ 3.96, difference p< 0.001; mean OBSE: bank A¼ 4.99, bank B¼ 4.61, difference p< 0.001;
mean AC: bank A¼ 5.02, bank B¼ 4.58, difference p< 0.001). On the other hand, also as expected,
employees in bank B exhibited a significantly higher average level of perceived job insecurity than
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
672 J. LEE AND R. PECCEI
Table
1.Descriptivestatistics
andzero-order
correlationsofstudyvariables
Variables
Mean
SD
NCronbacha
Correlations
BankA
BankB
BankA
BankB
BankA
BankB
BankA
BankB
12
34
Perceived
jobinsecurity
3.73
4.18
1.41
1.29
455
454
0.88
0.82
—�0
.29��
��0
.30��
��0
.09�
Perceived
organizational
support
4.49
3.96
0.86
0.84
456
454
0.95
0.96
�0.32��
�—
0.43��
�0.40��
�
Organization-based
self-esteem
(OBSE)
4.99
4.61
0.85
0.80
456
454
0.94
0.93
�0.34��
�0.53��
�—
0.46��
�
Affectivecommitment(A
C)
5.02
4.58
1.04
1.24
456
454
0.90
0.92
�0.22��
�0.54��
�0.55��
�—
BankA
correlationsarebelow
thediagonal.
BankB
correlationsareabovethediagonal.
� p�0.05;��p<0.01;��� p
<0.001.
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
PERCEIVED ORGANIZATIONAL SUPPORT AND AFFECTIVE COMMITMENT 673
employees in bank A (mean perceived job insecurity: bank A¼ 3.73, bank B¼ 4.18, difference
p< 0.001). In this context it is also worth noting that POS, OBSE AC, and perceived job insecurity
were moderately intercorrelated in both samples. In bank A the correlations ranged between�0.22 and
0.55, while in bank B they ranged from �0.29 to 0.46.
Main analysis: mediating role of OBSE
The main hypothesized partial mediation model shown in Figure 1 yielded an acceptable fit to the data
in each organization [for bank A: x2 (with 24 degree of freedom)¼ 106.57 (p< 0.001), GFI¼ 0.95,
NFI¼ 0.98, NNFI¼ 0.98, CFI¼ 0.98, IFI¼ 0.98, RMSEA¼ 0.087; for bank B: x2 (with 24 degree offreedom)¼ 61.80 (p< 0.001), GFI¼ 0.97, NFI¼ 0.99, NNFI¼ 0.99, CFI¼ 0.99, IFI¼ 0.99,
RMSEA¼ 0.059]. However, in line with Holmbeck’s (1997) recommendations, in order to determine
whether OBSE mediated the relationship between POS and AC, we compared the fit of the
hypothesized partial mediation model to that of a reduced model where the path between OBSE and AC
was deleted. In bank A, the partial mediation model (x2¼ 106.57 with df¼ 24, p< 0.001) provided a
significantly better fit to the data than the reduced model (x2¼ 132.74 with df¼ 25, p< 0.001).
Specifically, the difference in chi-square between the partial mediation and the reduced model was
statistically significant (D x2¼ 26.17, df¼ 1, p< 0.005), indicating that the path between OBSE and
AC should be included in the model. The size of the direct effect of POS on AC in the partial mediation
model was 0.40, while the size of the total effect of POS on AC in the reduced model was 0.58. In other
words, the results indicated that the direct effect was reduced by about 31 per cent [i.e., (0.58–0.40)/
0.58] when OBSE was included as a mediating variable. Thus, in bank A OBSE played a partial
mediator role between POS and AC, given that 69 per cent of the direct effect was maintained in the
relationship.
In bank B, the partial mediation model (x2¼ 61.80 with df¼ 24, p< 0.001) also provided a
significantly better fit to the data than the reduced model (x2¼ 109.86 with df¼ 25, p< 0.001).
Specifically, the difference in chi-square between the partial mediation and the reduced model was
statistically significant (D x2¼ 48.06, D df¼ 1, p< 0.005), indicating that the path between OBSE and
AC should be included in the model. The size of the direct effect of POS on AC in the partial mediation
model was 0.23, while the size of the total effect of POS on AC in the reduced model was 0.40. The
results indicated that the direct effect was reduced by about 42.5 per cent [(i.e., (0.40–0.23)/0.40] when
OBSE was included as a mediating variable. Thus, OBSE played a partial mediator role between POS
and AC also in bank B, given that 57.5 per cent of the direct effect was maintained in the relationship.
Therefore, in line with hypothesis 1, POS in both banks was found to have a significant direct positive
effect on AC, as well as a significant indirect positive effect on AC mediated through OBSE.
LISREL estimates (completely standardized coefficients) of the relationship between POS, OBSE,
and AC from the best-fitting partial mediation model for each bank are summarized in Figure 3. As can
be seen, in both organizations perceived organizational support was significantly positively related to
OBSE (bank A: b¼ 0.57, p< 0.001; bank B: b¼ 0.46, p< 0.001) which, in turn, was significantly
positively associated to AC (bank A: b¼ 0.31; p< 0.001; bank B: b¼ 0 .36, p< 0.001). In addition, in
both organizations perceived organizational support had a direct positive impact on AC (bank A:
b¼ 0.40, p< 0.001; bank B: b¼ 0.23, p< 0.001).
The direct, indirect and total effects of perceived organizational support and OBSE on AC in each
organization are summarized in Table 2. The table shows that POS had a significant positive total effect
on AC in both organizations. In both cases, however, the direct effect was considerably stronger than
the indirect one. Overall, however, the results of the analysis provide clear support to mediation
hypotheses 1 (a), 1 (b), and 1 (c).
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
674 J. LEE AND R. PECCEI
Main analysis: moderating role of perceived job insecurity
Although several procedures for investigating moderator effects have been proposed for use with
structural equation models, these various approaches appear to produce similar results (Joreskog &Yang,
1996; Mathieu, Tannenbaum, & Salas, 1992; Ping, 1995). Here we adopted the procedure proposed by
Ping (1995) since, both conceptually and operationally, this is one of the most straightforward approaches
to the analysis of moderator effects. Before conducting the moderator analysis, all observed variables
used to obtain the interaction (product) terms were centered and a residual centering approach was
adopted in order to minimize potential problems of multicollinearity and identification (Cortina, Chen, &
Dunlap, 2001; Lang, 1988). The significance of the moderator effects was tested by comparing the
chi-square values for models including and excluding the interaction terms.
We start with the analysis of the moderating effect of perceived job insecurity on the relationship
between POS and OBSE (see Hypothesis 2). The model that included the POS x perceived job
insecurity interaction term yielded a better fit to the data than did the model that excluded the
interaction term only in the bank A sample (bank A: D x2¼ 4.36, D df¼ 1, p< 0.05; bank B: Dx2¼ 0.59, D df¼ 1, p> 0.05). Moreover, as can be seen from the POS x perceived job insecurity
coefficients reported in Figure 4, the effect of the interaction between these two variables on OBSE was
statistically significant and negative only in bank A (bank A: b¼�0.09, p< 0.05; bank B: b¼�0.06,
p> 0.05). The precise form of the interactions for banks A and B are shown in Figures 5 (a) and 5 (b),
respectively. In brief, our results indicate that perceived job insecurity negatively moderated the
positive POS–OBSE relationship in bank A but not in bank B, thereby providing only partial support
for Hypothesis 2.
Turning to the analysis of the moderating effect of perceived job insecurity on the relationship
between POS and AC, the model that included the POS x perceived job insecurity interaction term
provided a better fit to the data in each sample than did the model that excluded the interaction term
Bank A: 0.40*** Bank B: 0.23***
Bank A: 0.57*** Bank A: 0.31*** Bank B: 0.46*** Bank B: 0.36***
* = p < 0.05, ** = p < 0.01, *** = p < 0.001
PerceivedOrganizationalSupport
AffectiveCommitment
Organization-based Self-esteem
Figure 3. LISREL estimates (standardized coefficients) for the relationship between POS, OBSE, and AC fromthe partial mediation model in each bank (test of hypotheses 1(a), 1(b), and 1 (c))
Table 2. LISREL estimates (standardized coefficients) of the decomposed direct, indirect, and total effects of POSand OBSE on AC in each bank
Main study variables
Direct effectson AC
Indirect effectson AC Total effects on AC
Bank A Bank B Bank A Bank B Bank A Bank B
Perceived organizational support 0.40��� 0.23��� 0.17��� 0.17��� 0.57��� 0.40���
Organization-based self-esteem (OBSE) 0.31��� 0.36��� 0.31��� 0.36���
�¼p� 0.05; ��¼p< 0.01; ���¼p< 0.001.
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
PERCEIVED ORGANIZATIONAL SUPPORT AND AFFECTIVE COMMITMENT 675
(bank A:D x2¼ 6.06, D df¼ 1, p< 0.05; bank B:D x2¼ 6.79, D df¼ 1, p< 0.01). Moreover, as shown
by the POS x perceived job insecurity coefficients reported in Figure 6, the effect of the interaction
between these two variables on ACwas statistically significant and positive in both organizations (bank
A: b¼ 0.09, p< 0.05; bank B: b¼ 0.11, p< 0.05). The precise nature of the interactions for banks A
and B are shown in Figures 7 (a) and 7 (b), respectively. As can be seen, the results show that perceived
job insecurity positively moderated the positive POS–AC relationship, thereby supporting hypothesis
3(b) over hypothesis 3(a), which suggested that job insecurity has a negative moderating effect on the
POS–AC relationship.5
Finally, it is worth noting that to check for the sensitivity of the results to different model
specifications, we also re-ran the analysis adding first, just the set of demographic variables (i.e., age,
gender, marital status, years of education, and organizational and job tenure) as controls in the analysis,
and then adding the demographic and the dispositional variables (i.e., positive and negative affectivity)
together as controls. The results of these additional analyses were virtually the same as those of the
main analysis reported above (results available from the authors). The only difference was that in the
main analysis, as we have seen, the coefficient of the POS x job insecurity interaction on OBSE for bank
B was negative but non-significant at either the 0.05 or even the 0.10 level, while in the additional
analysis this coefficient was still negative but significant at just below the 0.10 level. This suggests that
the tendency observed in bank A for job insecurity to moderate the relationship between POS and
OBSE potentially also holds in bank B, but in a much weaker form. More generally, though, the results
of the additional analyses indicate that the mediator and moderator effects observed in the main
analysis are quite stable, in the sense that, by and large, they remain virtually unchanged irrespective of
the controls that are included in the regressions.
Bank A: 0.40*** Bank B: 0.23***
Bank A: 0.49*** Bank A: 0.31*** Bank B: 0.40*** Bank B: 0.36***
Bank Aa: -0.09*
Bank Ba: -0.06
a coefficient = POS x perceived job insecurity
* = p < 0.05 ** = p < 0.01 *** = p < 0.001
PerceivedOrganizationalSupport
Organization-based Self-esteem
Perceived Job Insecurity
AffectiveCommitment
Figure 4. LISREL estimates (standardized coefficients) for the moderating effect of perceived job insecurity onthe POS–OBSE relationship (test of hypothesis 2)
5As a further check on the results of the separate moderator analyses reported here, we also tested an overall model that includedboth moderator effects of interest together using, once again, the residual centering approach recommended by Lang (1988). Forboth banks the results of these additional analyses (available from the authors) were virtually the same as those reported aboveobtained with the separate models.
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
676 J. LEE AND R. PECCEI
Discussion
Theoretically, work dealing with the relationship between perceived organizational support and
affective organizational commitment has, for the most part, adopted a social exchange perspective
linked to notions of reciprocity and felt obligation. Our aim in the present study was to contribute to
understanding in this area in two ways. First, by exploring complementary socio-emotional
explanations of the POS–AC relationship, focusing, in particular, on the mediating role played by
OBSE. And second, by examining the extent to which, in line with social exchange and
socio-emotional arguments, key relationships between POS, OBSE, and AC are moderated by
perceived job insecurity. To this end, we developed a number of mediator and moderator models of the
relationship between POS, OBSE, AC and job insecurity which we then tested in separate analyses
using employee data from two Korean banks which experienced different levels of turbulence and
uncertainty following the 1997 Korean financial crisis.
The results of our analysis direct attention to five main points. The first relates to the role of OBSE as
a mediator of the relationship between perceived organizational support and AC. Our findings indicate
(a)
(b)
Low perceived job insecurity = one standard deviation below the mean High perceived job insecurity = one standard deviation above the mean
OB
SEHigh perceived job insecurity
OB
SE
POS
POS
Low perceived job insecurity
High perceived job insecurity
Low perceived job insecurity
Figure 5. (a) Interaction of POS and perceived job insecurity on OBSE (bank A). (b) Interaction of POS andperceived job insecurity on OBSE (bank B)
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
PERCEIVED ORGANIZATIONAL SUPPORT AND AFFECTIVE COMMITMENT 677
(a)
(b)
Low perceived job insecurity = one standard deviation below the mean High perceived job insecurity = one standard deviation above the mean
POS
POS
AC
High perceived job insecurity
Low perceived job insecurity
AC
High perceived job insecurity
Low perceived job insecurity
Figure 7. (a) Interaction of POS and perceived job insecurity on AC (bank A). (b) Interaction of POS andperceived job insecurity on AC (bank B)
Bank A : 0.09* a
Bank Ba: 0.11*
Bank A: 0.39*** Bank B: 0.25***
Bank A: 0.57*** Bank A: 0.32*** Bank B: 0.46*** Bank B: 0.39***
a coefficient = POS x perceived job insecurity
* = p < 0.05 ** = p < 0.01 *** = p < 0.001
PerceivedOrganizationalSupport
Organization-based Self-esteem
AffectiveCommitment
Perceived Job Insecurity
Figure 6. LISREL estimates (standardized coefficients) for the moderating effect of perceived job insecurity onthe POS–AC relationship (test of hypotheses 3(a) and 3(b))
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
678 J. LEE AND R. PECCEI
that, in line with Hypothesis 1 and the predictions of the socio-emotional perspective, OBSE is a
significant mediator of the POS–AC relationship. Consistent with symbolic interactionist arguments, in
both organizations examined, POS was found to have a positive impact on OBSE. And, in line with the
predictions of classical conditioning (Meyer & Allen, 1997), as well as with self-consistency
arguments (Korman, 1970), OBSE was, in turn, found to have a positive effect on AC.
The second point relates to the relative importance of OBSE as an explanatory mechanism in the
POS–AC relationship. In both organizations, the partial mediation model that was tested provided a
better fit to the data than did the reduced no mediation model positing a purely direct link between POS
and AC. In other words, although OBSE emerged as a significant mediator of the POS–AC relationship,
perceived organizational support was also found to have a direct effect on AC that did not go through
OBSE. This direct effect was substantial. On average, it accounted for about 58.8 per cent of the
relationship between perceived organizational support and AC across the two organizations, compared
to approximately 41.2 per cent via OBSE. More generally, therefore, our results indicate that although
OBSE is indeed an important mediator of the POS–AC link, there are other mechanisms through which
perceived organizational support has an impact on AC. These may include not only felt obligation, but
also other non-OBSE related socio-emotional factors. As we emphasize below, the systematic
identification of the full range of mechanisms involved is an important area for future research.
The third point relates to the moderating effects of perceived job insecurity. As expected, perceived
job insecurity was not found to have a significant moderating effect on the relationship between OBSE
and AC in either of the two case study organizations. More importantly, however, our results provided
at least partial support to the idea that job insecurity attenuates the positive relationship between POS
and OBSE (Hypothesis 2). As we have seen, this negative moderating effect of job insecurity on the
POS–OBSE relationship was far clearer and more pronounced in bank A than in bank B, where the
effect was, at best, marginal and apparent only in the analysis that included the controls. We return to
this point later. In contrast, job insecurity was found to have an equally strong moderating effect on the
direct relationship between POS and AC in both banks. In this case, however, our results provided
support to the idea that job insecurity augments (Hypothesis 3b), rather than attenuates (Hypothesis 3a),
the direct positive effect of POS on AC.
This differential pattern of interactions observed in our data is important since it reinforces the
relevance of both social exchange and socio-emotional arguments for the understanding of the links
between POS, OBSE, and AC. In other words, job insecurity, treated as a moderator variable, helps to
underscore the importance and usefulness of distinguishing social exchange from socio-emotionally
based explanations of the effects of perceived organizational support on AC. Specifically, in line with
Hypothesis 2, our results suggest that job insecurity and POS provide contradictory cues to employees
about the extent to which their organization actually values their contribution and cares for their
well-being. As a result, job insecurity tends to undermine the OBSE-enhancing effects of perceived
organizational support. Consistent with symbolic interactionist arguments in fact, our results indicate
that employees who are more insecure in their jobs need to perceive a comparatively higher level of
organizational support in order to experience the same level of commitment-enhancing OBSE as
employees who enjoy a greater level of job security.
At the same time, in line with Hypothesis 3b, and consistent with social exchange arguments, our
results also suggest that employees are likely to respond differently to support from the organization
depending on their sense of job insecurity. In particular, because of their greater vulnerability and
uncertainty, employees who are less secure in their jobs are more likely to ascribe greater symbolic and
practical value to any signs of support from the organization than are employees who enjoy a greater
sense of job security. A given level of organizational support, therefore, is likely to generate a stronger
sense of felt obligation among individuals who perceive their job to be less secure than among those
who perceive their job to be more secure. Hence, as our results show, individuals who are more
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
PERCEIVED ORGANIZATIONAL SUPPORT AND AFFECTIVE COMMITMENT 679
uncertain about their future in the organization tend, on the whole, to respond more positively to
perceived organizational support and are, therefore, more prone to reciprocate POS with AC than are
employees who enjoy a greater sense of job security. More generally, therefore, our results indicate that
job insecurity may well have contradictory moderating effects on the direct and indirect (through
OBSE) relationship between POS and AC. In practice, these effects may cancel each other out.
Whether they do or not, however, is an empirical question that needs to be examined situation by
situation.
This brings us to the fourth main point to emerge from the analysis. This relates to the stability of the
results across the two organizations covered in the study. Importantly, as we have seen, the partial
mediation model provided the best fit in both organizations and, in both cases, all the relationships
involved proved significant. Given the different levels of turbulence and uncertainty experienced by the
two organizations, this suggests that the links between POS, OBSE, and AC are quite robust. The same
applies to the job insecurity moderator effects examined, although here the effects involved tended to
be generally less pronounced, as well as less consistent.
Despite the overall robustness of the results, however, some variations were also apparent across the
two organizations. In particular, although there was a clear positive link between POS and both OBSE
and AC in both organizations, this link was found to be significantly weaker in the bank that
experienced the more massive program of downsizing (bank B). To formally test whether the impact of
perceived organizational support on OBSE and on ACwas significantly stronger in bank A than in bank
B, multi-group analysis was conducted to assess the invariance of the relevant parameter estimates
across the two samples. To this end, a partial mediation model in which the POS–OBSE and the
POS–AC parameters were assumed to be the same across the two organizations was compared to a
model in which these two parameters were not assumed to be invariant. The latter model fit the data
significantly better than the former one (D x2¼ 10.03, D df¼ 2, p< 0.01), indicating that the effect of
POS on both OBSE and AC was significantly stronger in bank A than in bank B.
The differential strength of the associations between POS, OBSE, and AC in the two banks suggests
that the effects of POS on OBSE, and AC may well vary depending not only on individuals’ personal
circumstances, such as their own particular level of job insecurity, but also depending on the more
general organizational context and situation. In particular, the weaker relationship found between POS
and both OBSE and AC in bank B than in bank A, suggests that the protracted uncertainty and
turbulence associated with major programs of job cuts and rationalization may well undermine
employee confidence and trust in the organization. In such circumstances, employees may be more
likely to view any support they receive from the organization in a more instrumental light, as a response
to immediate pressures rather than as a sign of the organization’s genuine concern for and commitment
to them as individuals. The greater the level of organizational turbulence and uncertainty, therefore, the
less readily are employee perceptions of organizational support likely to translate into higher levels of
self-esteem and commitment among the workforce. More generally, therefore, the level of
organizational turbulence and uncertainty may well constitute an important organizational-level
boundary condition affecting the impact of POS on OBSE. This is clearly a speculative argument that
would need to be tested across a larger sample of organizations experiencing varying degrees of
turbulence and uncertainty.
Avariant of the above boundary condition argument might also help to explain why there was a much
clearer tendency for job insecurity to attenuate the positive effect of POS on OBSE in bank A than in
bank B. One possible explanation for this result in fact is that, as hypothesized, a greater personal sense
of job security does indeed help to reinforce the positive effect of POS on individuals’ sense of OBSE.
However, in situations characterized by high levels of general turbulence and uncertainty about the very
future of the organization, even individuals who feel that their job in the organization is secure, are
likely to wonder for how long any preferential treatment that they may receive from the organization is
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
680 J. LEE AND R. PECCEI
likely to last. In other words, irrespective of their own particular sense of job security, when the
potential survival of the organization is unclear, employees are more likely to question the extent to
which the organization can continue to value their contribution and care for their well-being in the
future. In such a situation, therefore, individual perceptions of job security or insecurity are less likely
significantly to affect the relationship between POS and OBSE. Hence, the much weaker moderating
effect of job insecurity on the POS–OBSE relationship in bank B than in bank A. Once again, however,
this argument remains speculative and needs to be tested across a larger sample of organizations.
Finally, in line with much of the recent literature on downsizing (Hui & Lee, 2000; Luthans &
Sommer, 1999; Mone, 1994; Wagar, 1998), our findings confirm the substantial negative effects that
major job cuts and a sense of job insecurity tend to have on employee attitudes, including both OBSE
and organizational commitment. In this context it is worth noting that, in our study, the negative effects
of organizational turbulence and downsizing on OBSE may well have been magnified by broader
socio-cultural factors specific to the Korean context. As a number of scholars have noted, in contrast to
Westerners’ so-called independent construal of the self which emphasizes the self as a separate being
with its own autonomy, in East Asia including Korea, the self is only meaningful in the context of
society as a whole (Inumiya, Choi, Yoon, Seo, & Han, 1999; Markus & Kitayama, 1991). Hence, in
Korea, society as a whole can be a ‘significant other’ and, in line with symbolic interactionist
arguments, the view from society at large can affect peoples’ conception of the self, including their
self-esteem. Before the 1997 financial crisis, bank B was a successful institution, which provided its
employees with high salaries, job security, and generous company-based welfare benefits, such as
family allowances, and children’s educational allowances. As a result, employment in the bank was
prized and was regarded as being highly prestigious within Korean society. However, after the financial
crisis, bank B came to be regarded as a weak institution with poor prospects. Hence, being an employee
of the bank no longer carried the same status and prestige as it had previously, thereby potentially
contributing to the significantly lower levels of OBSE observed in bank B than in bank A in our study.
More generally, as we have seen, the negative effects of downsizing and job insecurity can, to some
extent, be mitigated by providing effective support to employees. Such support can help to enhance
employee commitment to the organization in two main ways. First, is by creating a stronger sense of
felt obligation among the workforce. Second, and equally important, is by enhancing employees’ sense
of OBSE. Our findings suggest that such a support strategy is likely to be particular important in helping
to maintain levels of commitment among employees who feel less secure about their jobs. At the same
time, though, our findings also suggest that such a support strategy requires careful management. For it
to be effective, employees need to be convinced that the organization’s actions are motivated by good
intentions. Thus, as Meyer and Allen (1997) have argued more generally, to enhance workforce
commitment management should not only keep employees informed of its actions and intentions, but it
should also monitor their understanding of and reactions to major organizational initiatives and seek to
involve them in key decisions that directly affect their well-being. Our findings suggest that the
adoption by management of an explicitly proactive strategy of communication, involvement and
support of this kind is likely to be particularly important in helping to maintain levels of workforce
commitment during major programs of organizational restructuring and downsizing.
Study limitations and future research
As already noted, the self-report measures used in the present study did not pose significant problems of
commonmethod variance in the analysis. However, the cross-sectional nature of the data clearly did not
permit a systematic causal analysis of the links between POS, OBSE, and AC. This is an important
limitation of the study. So is the lack of data on alternative mechanisms through which perceived
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
PERCEIVED ORGANIZATIONAL SUPPORT AND AFFECTIVE COMMITMENT 681
organizational support may have an impact on commitment, including, for example, felt obligation, as
well as other non-OBSE related socio-emotional factors, such as individuals’ feelings of self-respect,
approval, growth, and affiliation. Additional data of these kind would have allowed for a better
understanding of the specific mediating role played by OBSE in the POS–AC relationship relative not
only to social exchange, but also to other possible socio-emotional mechanisms. A final limitation of
the study relates to the lack of data on other possible individual employee perceptions and orientations,
apart from perceived job insecurity, that might moderate the observed relationships between POS,
OBSE, and AC. In particular, data on employees’ exchange ideology and trust in management, as well
as data on their evaluation of management’s handling of the program of restructuring and downsizing,
would have allowed for a more fine-grained moderator analysis of the impact of POS on both OBSE
and AC in the two organizations.
Because of these various limitations, the results and conclusions of the present study need to be
treated with caution. To address these limitations, future research should seek to employ longitudinal
data designed to assess changes in employee levels of POS, OBSE, and AC over time. Ideally such data
should be collected not only before and after major programs of organizational restructuring and
downsizing, but also at various points during the implementation of such programs. This would allow
for a more systematic analysis of the way in which perceived organizational support, OBSE and AC
vary over the lifetime of such initiatives and, hence, provide a better understanding of underlying
patterns of causal relations between these variables. It would also allow for a more detailed analysis of
the way in which changes in perceived levels of job security affect the relationship between POS,
OBSE, and AC over time.
More research is clearly needed also to test the generalizability of the present findings. In particular,
further research is required to determine the extent towhich the pattern of results obtained in the present
study applies to other organizations in different societal, cultural, and economic contexts. Particularly
useful in this respect would be research designed to test whether the positive relationships between
POS, OBSE, and AC observed in the present study also hold in non-Korean organizations, as well as in
more successful organizations that are not going through major periods of turbulence and uncertainty.
Our expectation is that the present pattern of results is not only likely to apply more widely also to
non-Korean organizations, but that the positive links found between POS, OBSE, and AC are, if
anything, likely to be even more pronounced in successful organizations that are not going through
major programs of rationalization and restructuring. These expectations, however, require systematic
testing through further research.
In the process, it would also be useful to extend the analysis of the POS–OBSE-AC relationship in
two main ways. First, is by expanding the range of mediator variables included in the analysis so as to
allow for a more systematic comparison of social exchange and socio-emotional explanations of the
POS–AC relationship. And second, is by extending the analytical model to include a wider range of key
employee perceptions and orientations that may significantly moderate relationships of interest in
the area. The present study constitutes a first step in such a broader, longer-term program of
inquiry.
Acknowledgement
This research was supported by the Kyung Hee University Research Fund in 2004 (KHU-20040907).
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
682 J. LEE AND R. PECCEI
Author biographies
Jaewon Lee received his PhD in Industrial and Business Studies at the University of Warwick and is
now an assistant professor of Business Administration at Kyung Hee University in Korea. His primary
research interests include organization-based self-esteem, organizational politics, and the impact of
HRM on organizational performance and employee well-being.
Riccardo Peccei received his DPhil in sociology at Oxford University and is now a professor of Human
Resource Management and Organisational Behaviour at King’s College, London. His primary research
interests include the study of the transformation of work and employment relations in the service sector,
gender demography, the impact of HRM on organizational performance and employee well-being, and
the nature and consequences of employee empowerment and participation at work.
References
Ajzen, I., & Fishbein, M. (1977). Attitude-behavior relations: A theoretical analysis and review of empiricalresearch. Psychological Bulletin, 84, 888–918.
Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommendedtwo-step approach. Psychological Bulletin, 103, 411–423.
Armeli, S., Eisenberger, R., Fasolo, P., & Lynch, P. (1998). Perceived organizational support and policeperformance: The moderating influence of socioemotional needs. Journal of Applied Psychology, 83, 288–297.
Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy ofMarketing Science, 16, 74–94.
Baumeister, R. F. (1995). Self and identity: An introduction. In A. Tesser (Ed.), Advanced Social Psychology(pp. 51–97). New York: McGraw-Hill.
Blumer, H. (1969). Symbolic interactionism. In R. Collins (Ed.), Four sociological traditions: Selected readings(pp. 69–112). Oxford: Oxford University Press.
Brockner, J. (1988). Self-esteem at work: Research, theory, and practice. Lexington, MA: Lexington Books.Byrne, D., & Clore, G. L. (1970). A reinforcement model of evaluative responses. Personality, 1, 103–127.Cortina, J. M., Chen, G., & Dunlap, W. P. (2001). Testing interaction effects in LISREL: Examination andillustration of available procedures. Organizational Research Methods, 4, 324–360.
Eisenberger, R., Cummings, J., Armeli, S., & Lynch, P. (1997). Perceived organizational support, discretionarytreatment, and job satisfaction. Journal of Applied Psychology, 82, 812–820.
Eisenberger, R., Fasolo, P., & Davis-LaMastro, V. (1990). Perceived organizational support and employeediligence, commitment, and innovation. Journal of Applied Psychology, 75, 51–59.
Eisenberger, R., Huntington, R., Hutchison, S., & Sowa, D. (1986). Perceived organizational support. Journal ofApplied Psychology, 71, 500–507.
Eisenberger, R., Armeli, S., Rexwinkle, B., Lynch, P. D., & Rhoades, L. (2001). Reciprocation of perceivedorganizational support. Journal of Applied Psychology, 86, 45–54.
Garst, H., Frese, M., &Molenaar, P. C. M. (2000). The temporal factor of change in stressor-strain relationships: Agrowth curve model on a longitudinal study in East Germany. Journal of Applied Psychology, 85, 417–438.
Gouldner, A. W. (1960). The norm of reciprocity: A preliminary statement. American Sociological Review, 25,161–178.
Greenberg, J., Solomon, S., Pyszczynski, T., Rosenblatt, A., Burling, J., Lyon, D., Simon, L., & Pinel, E. (1999).Why do people need self-esteem? Converging evidence that self-esteem serves an anxiety-buffering function. InR. F. Baumeister (ed.), The self in social psychology (pp. 24–49). Philadelphia, PA: Psychology Press.
Hair, J. F. Jr., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis (5th ed.). UpperSaddle River, NJ: Prentice Hall.
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
PERCEIVED ORGANIZATIONAL SUPPORT AND AFFECTIVE COMMITMENT 683
Harris, M. M., & Schaubroeck, J. (1990). Confirmatory modeling in organizational behavior/human resourcemanagement: Issues and applications. Journal of Management, 16, 337–360.
Harter, S. (1993). Causes and consequences of low self-esteem in children and adolescents. In Baumeister R. F.(Ed.), Self-esteem: The puzzle of low self-regard (pp. 197–229). London: Plenum Press.
Hewitt, J. P. (1997). Self and Society: A symbolic interactionist social psychology (7th ed.). London: Allyn andBacon.
Holmbeck, G. N. (1997). Toward terminological, conceptual, and statistical clarity in the study of mediators andmoderators: Examples from child-clinical and pediatric psychology literatures. Journal of Consulting andClinical Psychology, 65, 599–610.
Hui, C., & Lee, C. (2000). Moderating effects of organization-based self-esteem on organizational uncertainty:Employee response relationships. Journal of Management, 26, 214–232.
Hulin, C. L., & Mayer, L. J. (1986). Psychometric equivalence of a translation of the job descriptive index intoHebrew. Journal of Applied Psychology, 71, 83–94.
Inumiya, Y., Choi, I., Yoon, D., Seo, D., & Han, S. (1999). The relationship between unrealistic optimism andindependent-interdependent construals of self in Korean culture. Korean Journal of Social and PersonalityPsychology, 13, 183–201.
Joreskog, K. G., & Sorbom, D. (1993). LISREL 8:Structural equation modeling with the SIMPLIS commandlanguage. Chicago. IL: Scientific Software International.
Joreskog, K. G., & Yang, F. (1996). Nonlinear structural equation models: The Kenny-Judd model with interactioneffects. In G. A. Marcoulides, & R. E. Schumacker (Eds.), Advanced Structural Equation Modeling Techniques(pp. 57–88). Hillsdale, NJ: Lawrence Erlbaum.
Ko, J., Price, J. L., & Mueller, C. W. (1997). Assessment of Meyer and Allen’s three-component model oforganizational commitment in South Korea. Journal of Applied Psychology, 82, 961–973.
Korman, A. K. (1970). Toward an hypothesis of work behavior. Journal of Applied Psychology, 54, 31–41.Korman, A. K. (1971). Organizational achievement, aggression and creativity: Some suggestions toward anintegrated theory. Organizational Behavior and Human Performance, 6, 593–613.
Lang, C. E. (1988). Residual centering, exploratory and confirmatory moderator analysis, and decomposition ofeffects in path models containing interactions. Applied Psychological Measurement, 12, 163–175.
Lee, J. (2003a). An analysis of the antecedents of organization-based self-esteem in two Korean banks.International Journal of Human Resource Management, 14, 1046–1066.
Lee, J. (2003b). An analysis of organization-based self-esteem as a mediator of the relationship between itsantecedents and consequences. The Korean personnel Administration Journal, 28, 33–60.
Lee, J., & Corbett, J. M. (2006). The impact of downsizing on employees’ affective commitment. Journal ofManagerial Psychology, 21, 176–199.
Levinson, H. (1965). Reciprocation: The relationship between man and organization. Administrative ScienceQuarterly, 9, 370–390.
Locke, E. A., McClear, K., & Knight, D. (1996). Self-esteem and work. In C. L. Cooper, & I. T. Robertson (Eds.),International review of industrial and organizational psychology (Vol. 11, pp. 1–32.). New York: John Wiley &Sons.
Luthans, B. C., & Sommer, S. M. (1999). The impact of downsizing on workplace attitudes.Group &OrganizationManagement, 24, 46–70.
Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation.Psychological Review, 98, 224–253.
Mathieu, J. E., Tannenbaum, S. I., & Salas, E. (1992). Influences of individual and situational characteristics onmeasures of training effectiveness. Academy of Management Journal, 35, 828–847.
Mathieu, J. E., & Zajac, D.M. (1990). A review andmeta-analysis of the antecedents, correlates, and consequencesof organizational commitment. Psychological Bulletin, 108, 171–194.
Meyer, J. P., & Allen, N. J. (1991). A three–component conceptualization of organizational commitment. HumanResource Management Review, 1, 61–89.
Meyer, J. P., &Allen, N. J. (1997).Commitment in the workplace: Theory, research, and application. London: SagePublications.
Meyer, J. P., Stanley, D. J., Herscovitch, L., & Topolnytsky, L. (2002). Affective, continuance, and normativecommitment to the organization: A meta-analysis of antecedents, correlates, and consequences. Journal ofVocational Behavior, 61, 20–52.
Mone, M. A. (1994). Relationships between self-concepts, aspirations, emotional responses, and intent to leave adownsizing organization. Human Resource Management, 33, 281–298.
Mottaz, C. J. (1988). Determinants of organizational commitment. Human Relations, 41, 467–482.
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
684 J. LEE AND R. PECCEI
Mowday, R. T., Porter, L. W., & Steers, R. M. (1982). Employee-organization linkages: The psychology ofcommitment, absenteeism, and turnover. London: Academic Press.
Pierce, J. L., & Gardner, D. G. (2004). Self-esteem within the work and organizational context: A review of theorganization-based self-esteem literature. Journal of Management, 30, 591–622.
Pierce, J. L., Gardner, D. G., Cummings, L. L., & Dunham, R. B. (1989). Organization-based self-esteem:Construct definition, measurement, and validation. Academy of Management Journal, 32, 622–648.
Ping, R. A. (1995). A parsimonious estimating technique for interaction and quadratic latent variables. Journal ofMarketing Research, 32, 336–347.
Podsakoff, P. M., MacKenzie, S. B., & Podsakoff, N. P. (2003). Common method biases in behavioral research: Acritical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879–903.
Price, J. L. (1991). Chicago Public Schools (CPS) Questionnaire Items Wave I.Price, J. L. (1997). Handbook of organizational measurement. International Journal of Manpower, 18, 303–558.Rhoades, L., & Eisenberger, R. (2002). Perceived organizational support: A review of the literature. Journal ofApplied Psychology, 87, 698–714.
Rhoades, L., Eisenberger, R., & Armeli, S. (2001). Affective commitment to the organization: The contribution ofperceived organizational support. Journal of Applied Psychology, 86, 825–836.
Shore, L. M., & Shore, T. H. (1995). Perceived organizational support and organizational justice. In R. S.Cropanzano, & K.M. Kacmar (Eds.),Organizational politics, justice, and support: Managing the social climateof the workplace (pp. 149–184). London: Quorum Books.
Van Dyne, L., & Pierce, J. L. (2004). Psychological ownership and feelings of possession: Three field studiespredicting employee attitudes and organizational citizenship behavior. Journal of Organizational Behavior, 25,439–459.
Van Dyne, L., Vandewalle, D., Kostova, T., Latham, M. E., & Cummings, L. L. (2000). Collectivism, propensity totrust and self-esteem as predictors of organizational citizenship in a non-work setting. Journal of OrganizationalBehavior, 21, 3–23.
Wagar, T. H. (1998). Exploring the consequences of workforce reduction. Canadian Journal of AdministrativeSciences, 15, 300–309.
Watson, D., Clark, L. A., & Carey, G. (1988). Development and validation of brief measures of positive andnegative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063–1070.
Zimbardo, P. G., & Leippe, M. R. (1991). The psychology of attitude change and social influence. New York:McGraw-Hill.
Copyright # 2006 John Wiley & Sons, Ltd. J. Organiz. Behav. 28, 661–685 (2007)
DOI: 10.1002/job
PERCEIVED ORGANIZATIONAL SUPPORT AND AFFECTIVE COMMITMENT 685