A Role for Equity Theory in the Turnover Process: An Empirical Test1

A Role for Equity Theory in the Turnover Process: An Empirical Test1

RODGER w. GRlFFETH2 AND STEFAN GAERTNER Department of Management

Georgia State Universiw

The purpose o f the present study was to examine the role o f equity theory in the context of the contemporary turnover process. A model was developed and tested with 192 hospital employees using structural equation modeling (SEM), which placed satisfaction and intention to quit as mediators of employee turnover. The results strongly support the present model, but also suggest a role for other mediators, some of which are suggested for future research.

Equity theory (Adams, 1963, 1965) continues to be a major model stimulat- ing considerable research regarding work motivation in recent years (Bretz & Thomas, 1992; Carr, McLoughlin, Hodgson, & Maclachlan, 1996; Glass & Wood, 1996; Greenberg, 1990; Harder, 199 1, 1992; Huseman, Hatfield, & Miles, 1985, 1987; Johnson & Johnson, 1991; Joshi, 1990; King & Miles, 1994; King, Miles, & Day, 1993; Miles, Hatfield, & Huseman, 1994; Perry, 1993; Sheehan, 1993; Sweeney, 1990; Van Dierendonck, Schaufeli, & Sixma, 1994). Briefly summarized, equity theory suggests that an employee compares the ratio of his or her outcomes to inputs to the ratio of outcomes to inputs of some referent other. Employees who perceive themselves in an inequitable situation will be dissatis- fied and will try to reduce the inequity. Although Adams (1963,1965) proposed a number of ways that employees might reduce inequity, research concerned with organizations has tended to focus on employee reactions to pay inequity, such as low performance and dissatisfaction (Greenberg, 1990). Leaving the situation (i.e., voluntary turnover) was also postulated as a tension reaction mode by Adams ( 1 963, 1965), but only a few studies have examined the impact of ineq- uity perceptions on turnover in the context of contemporary ideas regarding a turnover process (e.g., Horn & Griffeth, 1995; Mobley, 1977). The purpose of the present study is to test the relationship between equity perceptions and turnover within the context of such a contemporary turnover process model.

'The authors greatly appreciate the comments of Peter Hom, Debra Cohen, Peggy Lewis, and

2Correspondence concerning this article should be addressed to Rodger W. Gtiffeth, Department two anonymous reviewers on earlier versions of this paper.

o f Management, Georgia State University, Atlanta, GA 30303.

1017

JournalofApplied SocialPsychology, 2001, 31, 5, pp. 1017-1037. Copyright 0 2001 by V. H. Winston & Son, Inc. All rights reserved.

1018 GRIFFETH AND GAERTNER

Early research into the equity-turnover relationship, using aggregate rates of turnover, found mixed results. For example, Telly, French, and Scott (1971) found aggregated perceptions of equity within subunits of an organization to be significantly related to that subunit’s turnover rate for the previous 11 months on five of the seven dimensions (e.g., supervision, social aspects). Dittrich and Carrel1 ( 1 979) developed and tested a five-dimension measure of equity percep- tions that they called the Organizational Fairness Questionnaire (OFQ). They found that the five factors (pay rules, pay administration, work pace, pay level, and rule administration) underlying the OFQ were not predictive of turnover rates by department. However, pay rules (a factor that combines comparisons of one’s own pay to that of coworkers with the fairness of the rules for granting pay increases and promotions) and work pace (fairness of the supervisor in maintain- ing a fair pace of work activity) were predictive of absence. While not directly predictive of turnover, employee perceptions of the fairness of pay rules and equality of pay among coworkers and of supervisor control of the work pace were strongly predictive of job satisfaction ( R 2 = S8).

Equity and Individual Turnover

At the individual level, tests of a direct relationship between equity percep- tions and turnover have also had mixed results. For example, Oldham, Kulik, Ambrose, Stepina, and Brand ( 1 986) found that equity perceptions in combina- tion with job complexity descriptions were marginally predictive of turnover. Vecchio, Griffeth, and Hom’s (1986) initial findings were that perceptions of supervisor control over work-pace equity were significantly related to turnover. However, when leader-member exchange quality was added, this variable fully mediated the equity-turnover relationship.

Finally, Randall and Mueller (1995) found no significant direct relationships between turnover and distributive or procedural justice perceptions. They sug- gested two plausible explanations for this lack of significant findings. First, the effect of equity on turnover is not direct, but rather is mediated by several other variables such as job satisfaction, organizational commitment, and intention to stay. Second, a lack of alternative employment opportunities among the nurses in their sample induced them to stay despite possible inequities. This latter explana- tion seems less probable since turnover and shortage rates among nurses remained high into the mid-1990s (Hom & Griffeth, 1995).

Another set of studies has examined the relationship of equity perceptions to intention to stay. The results in these studies have also been mixed. Scholl, Cooper, and McKenna ( 1 987) found that comparisons using others outside the company in similar jobs and comparisons of one’s own pay in the past were sig- nificant predictors of turnover intentions. However, Ronen (1 986) found that nei- ther equity referent (others inside the organization or others outside the organization) nor job level (skilled workers vs. managers) was significantly

EQUITY THEORY AND TURNOVER PROCESS 1019

related to turnover intentions. Finally, Berg (1991) found that global perceptions of equity were significant predictors of intention to stay among television employ- ees. Strictly speaking, however, Berg did not actually measure equity. Instead, he asked employees to assess their perceptions of fairness without reference to some comparison other. Again, it seems plausible, especially based on Adams’ (1 963, 1965) theory, that job dissatisfaction would be the immediate result of inequity perceptions, mediating the effects on turnover intentions and actual turnover.

In conclusion, these studies show relatively weak or inconsistent support regarding the relationship between inequity perceptions and turnover. There are several reasons for this inconsistency. First, all of these studies examined the direct influence of equity on turnover, generally ignoring the mediating role of felt tension of job dissatisfaction (Adams, 1963, 1965). Some studies attempted to link equity perceptions to intention to stay (or quit), a weak test of the equity- turnover relationship. According to Steel and Ovalle (I 984), intention to quit is a relatively poor surrogate for actual turnover, typically accounting for less than 25% of turnover variance. Moreover, the earlier studies were unable to take advantage of recent advances in the study of turnover identifying the process of employee turnover (Horn & Griffeth, 1991, 1995; Hom, Griffeth, & Sellaro, 1984; Mobley, 1977). Contemporary turnover theorists posit a series of cognitive and affective linkages translating job dissatisfaction into turnover cognitions and behavior. Perhaps the major weakness of each of these studies was the omission of satisfaction as a mediating variable.

More recent models that propose linkages between equity and turnover explic- itly incorporate perception of equity as an exogenous variable that has an impact on turnover via job satisfaction and quit intentions (Hulin, Roznowski, & Hachiya, 1985; Price & Mueller, 198 1). Two studies have placed the equity-tum- over relationship within the context of such a process model. Summers and Hendrix (1991) included perceptions of pay equity for comparisons with a gener- alized other (someone of similar knowledge, skills, and abilities), self (past), oth- ers inside the company, and others outside the company. Respondents were then asked to select the three most important referents. Only the equity perceptions with regard to the single most important referent were used as a measure of pay equity. Of respondents, 6% chose others outside of the company as their most important referent. A generalized other was the most frequently selected, followed closely by self (past). The model test revealed a significant mediated relationship between pay equity and turnover. The significant and hypothesized intervening variables were pay satisfaction, overall job satisfaction, and intention to leave.

Iverson and Roy (1994) performed a relatively comprehensive test of the Price and Mueller (1981, 1986) turnover model, which specifically includes ele- ments of equity perceptions of various benefits and pay in relationship to coworkers. Although the correlation coefficient for the relationship between equity and job satisfaction was positive and significant, the hypothesized and

1020 GRIFFETH AND GAERTNER

revised model yielded a negative (also significant) relationship between the two variables. Such a reversal may be indicative of multicollinearity (Neter, Wasserman, & Kutner, 1990) and suggests that a more parsimonious model might be practical. A second weakness of this test of the equity-turnover rela- tionship is the use of behavioral commitment, or intention to stay, rather than actual turnover as the ultimate dependent variable (Steel & Ovalle, 1984). How- ever, both the hypothesized and the revised model confirm affect Cjob satisfac- tion) as a mediator of the relationship between equity and intention to stay.

For both of the process model tests, the conceptualization of equity was nar- rowly limited to distributional outcomes, such as pay and benefits (Iverson & Roy, 1994; Summers & Hendrix, 1991) or to comparison others inside the orga- nization (Iverson & Roy, 1994). Prior research has found multiple comparison others to be important and, although pay seems to dominate perceptions of equity (Berg, 1991; Scholl et al., 1987; Summers & Hendrix, 1991), other facets are also important (Oldham et al., 1986; Telly et al., 1971; Vecchio et al., 1986; Wilhelm, Herd, & Steiner, 1993). It is the primary purpose of the present investigation to reexamine the role of equity perceptions within the context of contemporary turn- over theory and empirical research at the individual level of analysis. Toward this end, a model of this process is developed and tested based on equity theory. Both the predictive and nomological validity of this model are examined using struc- tural equation modeling (SEM). This model and the hypothesized relationships among the model constructs are shown in Figure 1.

Model Development

The present model posits a turnover process initiated by perceptions of ineq- uity in relationship to three key determinants of job satisfaction: pay satisfaction, satisfaction with one’s supervisor, and satisfaction with the work itself. We postu- late perceptions of inequity as determinants of these facets of job satisfaction, and, in doing so, we expand on two components of Adams’ (1 963, 1965) equity theory. First, Adams’ conceptualization of equity as concerned with the ratio of inputs to outcomes only (distributive justice) reflects the contemporaneous understanding of justice. More recently, justice has been characterized as a pro- cess, and researchers have delineated two meaningful elements in addition to dis- tributive justice as important to explaining perceptions of justice (Bies & Moag, 1986; Greenberg, 1990). These two elements are procedural justice, which refers to the rules and procedures used to arrive at the distribution of outcomes, and interactional justice, which refers to the way those who carry out the process relate to the recipient party (Bies & Moag, 1986; Greenberg, 1990). While these elements tend to be highly correlated, each uniquely adds to our understanding of the justice perceptions and reactions to those perceptions (Folger & Konovsky, 1989; McFarlin & Sweeney, 1992).

EQUITY THEORY AND TURNOVER PROCESS 1021 a Pay rules

%ion A , .

drninistratio

Work pace a’ Figure 1. Theoretical model and hypothesized relationships.

Second, research has shown that employees use more than one comparison other in developing their perceptions of inequity and the resulting dissatisfaction (Oldham et al., 1986; Rice, Phillips, & McFarlin, 1990). Commonly used refer- ents are internal comparison others or others in similar positions within the same company (Summers & DeNisi, 1990; Summers & Hendrix, 1991), and external comparison others or others in similar positions in other companies (Scholl et al., 1987; Singh, 1994; Summers & DeNisi, 1990).

For pay satisfaction, we suggest that perceptions of inequity arise from employee comparisons with external and internal referents, which include dis- tributive justice (pay level) and procedural justice (pay rules) as important to pay satisfaction (Huber, Seybolt, & Venemon, 1992; Konovsky, Folger, & Cropanzano, 1987; McFarlin & Sweeney, 1992). Note that our model distin- guishes between two aspects of procedural justice. One aspect refers to the struc- tural characteristics of the procedures (e.g., pay rules), while the other aspect refers to the administration of those structural characteristics by supervisors (e.g., pay administration and rule administration). The former aspect is considered pro- cedural justice while the latter aspect is interactional justice. Satisfaction with one’s supervisor emphasizes interactional justice (Bies & Shapiro, 1987). The concern here is how the organization’s designated procedures are administered by the supervisor. Employees who perceive their supervisors as fair in adminis- tering rules related to pay and promotional opportunities (pay administration) and in general supervision of the work environment (rule administration) will be more satisfied with their supervisors.

1022 GRIFFETH AND GAERTNER

Satisfaction with the work itself has been found to be an important determi- nant of overall job satisfaction. Using an internal standard, we measure percep- tions of equity relative to work pace maintained in the organization (work pace) among the respondent’s coworkers. Here, satisfaction will include elements of distributive and procedural justice; that is, fair pace fairly governed.

In summary, perceptions of unjust distribution of pay (pay level) and unjust rules governing those allocation decisions (pay rules) would lead to pay dissatis- faction. Perceptions of unjust administration by one’s supervisor in terms of rules governing pay (pay administration) and in terms of one’s work environment (rule administration) would lead to dissatisfaction with one’s supervisor. Finally, per- ceptions of unjust management of the work itself (work pace) would lead to dis- satisfaction with the work itself. Following the widely supported view of turnover process theory, those three facets of job satisfaction are hypothesized to affect turnover indirectly via the intermediate variable, intention to quit (e.g., Mobley, 1977; Mobley, Griffeth, Hand, & Meglino, 1979; Price & Mueller, 1986).

In the present model, we also predict that the structural error terms of the three facets of overall job satisfaction are correlated. Structural error of an endog- enous construct represents that part of the variance of a construct that is not explained by its antecedents (Bollen, 1989). In models that include all anteced- ents of a given construct, this structural error term will be random and generally not correlated with any other construct. In the present study, however, we do not specify all antecedents of the three facets of overall job satisfaction since we are merely interested in the effects of equity concepts on turnover. As a result, the structural error term of each job satisfaction facet will not merely consist ofran- dom error. There will rather be a systematic error component in each of those job satisfaction facets, which represents that part of the job satisfaction facet that is not explained by its model antecedents. In other words, each structural error term of each of those job satisfaction facets will contain some systematic variance of its respective job satisfaction facet (e.g., the structural error term of pay satisfac- tion contains pay satisfaction to some degree). Consequently, the structural error terms of the facets of job satisfaction are predicted to be correlated because prior research suggests that facets of job satisfaction are correlated (e.g., Smith, Kendall, & H u h , 1969; Spector, 1997).

Method

Procedure

We administered a survey to 192 employees of a large hospital in a major city in the northeastern United States. Small groups were surveyed on hospital time. We asked respondents to identify themselves in order to match survey data to

EQUITY THEORY AND TURNOVER PROCESS 1023

subsequent turnover. Researchers guaranteed the confidentiality of individual responses and stressed that participation was voluntary. Usable data were obtained from 180 respondents, which is equivalent to a response rate of 94%.

Respondent Characteristics

The sample included 171 (95%) Caucasian and 142 (79%) female employ- eess. Average age was 3 1 years, and average tenure was 5.1 years; 94 (52%) were married, and the average number of dependents was nearly two; 119 (66%) had formal education beyond high school, and 162 (90%) were full-time employees. Of the sample, 63 (35%) were registered nurses, and the remaining sample con- sisted of diverse white-collar hospital employees, such as medical technicians and licensed practical nurses.

Measures

Equiw. Dittrich and Carrell’s (1979) OFQ was used to assess five dimensions of perceived unfairness: pay rules, or the fairness of one’s pay relative to one’s coworkers and the fairness of granting pay increases and promotions (a = .83, 10 items); p q administration, or the perceptions of the fairness of the supervisor in administering the rules for pay raises and promotions (a = .84, 5 items); pay level, or the fairness of one’s pay relative to others’ pay outside of the employing organization (a = .66, 4 items); work pace, or the fairness of the supervisor in maintaining a fair pace of work activity (a = .82, 6 items); and rule administra- tion, or the fairness of supervisors in maintaining acceptable forms of acceptable conduct in the workplace (a = .83, 4 items). Responses were made on a 5-point Likert-type scale ranging from 1 (strongly agree) to 5 (strongly disagree). High scores on the equity constructs indicate low perceptions of distributional, inter- personal, and procedural equity.

Satisfaction. Work satisfaction was measured using the average of three bipo- lar semantic differential scales (awful-nice, bakgood, unpleasant-pleasant; a = .9 1). Pay satisfaction and supervisor satisfaction were measured with the same semantic-differential scales (a = .98 and .95, respectively).

Intention to quit. We used a two-item index measuring the likelihood that respondents would leave the hospital within 1 year (a = .87). A similar measure was validated in earlier research (Hom et al., 1984).

Turnover: We collected voluntary turnover data approximately 1 year after the survey data were collected. Determination of whether the turnover was voluntary or involuntary was made from confidential exit interviews conducted by the per- sonnel department. Turnover was coded 1 for quitters and 0 for stayers. Of the sampled employees, 5% left the hospital voluntarily within the study period. Involuntary turnover did not occur.

1024 GRIFFETH AND GAERTNER

Srarisrical Analysis

We used LISREL 8 (Joreskog & Sorbom, 1993) to evaluate the model’s causal structure, analyzing the covariance matrix based on listwise deletion of missing data (Bentler & Chou, 1987; Bollen, 1989; Hayduk, 1987). Following a procedure described by Williams and Hazer (1986) and others (e.g., Settoon, Bennett, & Liden, 1996), the items representing each model construct were aver- aged. Measurement error was accounted for in the empirical analysis by setting the factor loadings of the manifest indicators to the square root of coefficient alpha. The error variance of each manifest indicator was set to be the product of the variance of the manifest indicator multiplied by the quantity “one minus coef- ficient alpha.”

To evaluate the empirical fit of the model, we followed Bollen’s (1989) rec- ommendations to inspect multiple indicators. We reviewed overall fit indexes such as the chi-square test, the root mean square error of approximation (RMSEA; Steiger & Lind, 1980), the comparative fit index (CFI; Bentler, 1990), the normed fit index (NFI; Bentler & Bonnett, 1980), and the goodness of fit index (GFl; Jbreskog & Sbrbom, 1984). We also assessed the fit of the model by examining individual parameter estimates. We screened these estimates for improper solutions (e.g., negative variances), counterintuitive signs, and prohibi- tively large standard errors. In addition, we checked whether the iterative proce- dure of LISREL 8 converged to a solution after relatively few iterations. Finally, we screened the modification indexes to determine whether the empirical data suggest respecification of the model.

Results

The correlations between the model constructs, along with their variances, are shown in Table 1. As past research has reported (e.g., Hom & Griffeth, 1995), the best single predictor of turnover was intention to quit (r = .20, p < .05).

Overall Fit of the Model

The LISREL 8 procedure converged to an acceptable solution rapidly after seven iterations. No improper solutions were detected in the LlSREL 8 output. An inspection of the overall fit indexes revealed an excellent fit of the model, ~ ~ ( 2 3 , N = 180) = 29.88, p = .15, ns (RMSEA = .041, NFI = .94, GFI := .97, CFI = .99). Using an SAS program provided by MacCallum, Browne, and Sug- awara (1993), we calculated the power of our statistical tests to reject a structural model that is inaccurate in the population. These calculations are based on the RMSEA fit index and revealed that the power of our tests was relatively small (.452) for rejecting a model with fair fi t in the population (RMSEA == .08;

Tabl

e 1

Mea

ns, S

tand

ard

Dev

iatio

ns, a

nd C

orre

latio

ns of t

he C

onst

ruct

s Use

d in

This

Stud

y

M

SD

1 2

3 4

5 6

7 8

9

1. P

ay ru

les

3.52

0.

66 -

2. P

ay le

vel

3.38

0.

67

.32

-

rn

3. P

ay a

dmin

istra

tion

3.04

0.

71

.52

.24

-

4. R

ule

adm

inis

tratio

n 2.

76

0.87

.1

7 .0

7 .4

1 -

5. W

ork

pace

3.

48

0.71

.3

3 .2

0 .5

3 .6

4 -

>

z 0

c

;o z 0

;o

6. P

ay s

atis

fact

ion

3.88

1.

82

-.45

-.53

-.30

-.08

-.22

-

7. S

uper

viso

r sat

isfa

ctio

n 4.

31

1.55

-.2

7 -.

I9

-SO

-.34

-.40

.32

-

-I s

10. T

urno

ver

0.05

0.

22

.06

.03

.10

.08

.15

-.09

.01

-.04

.20

m 8 om

8. W

ork

satis

fact

ion

4.38

1.

39

-.23

-.I1

-.23

-.21

-.34

.32

.40

-

9. I

nten

tion

to q

uit

2.48

0.

96

.32

.19

.37

.17

.22

-.25

-.39

-.38 -

Not

e. N

= 1

80. C

orre

latio

ns g

reat

er th

an .I

4 ar

e si

gnifi

cant

atp

< .0

5.

cn

0)

1026 GRIFFETH AND GAERTNER

Pay rules

a’ Work pace

Figure 2. Completely standardized parameter estimates among the model constructs. Links between exogenous variables were omitted for clarity. *p < .05.

Browne & Cudeck, 1993), but relatively large (.837) for rejecting a model with poor fit in the population (RMSEA = . lo; Browne & Cudeck, 1993).

Despite the good overall fit of the model, we detected two relatively large modification indexes. Those modification indexes revealed that an additional direct path between pay rules and intention to quit reduces the chi-square statistic by approximately 11.95, while an additional direct path between pay administra- tion and intention to quit reduces the chi-square statistic by approximately 10.33. The detection of those two significant modification indexes suggests that there may be further constructs, other than the three facets of job satisfaction that were incorporated in the model, that mediate the effect of equity on intention to quit. This possibility is discussed in a later section of this paper.

Parameter Estimates

Figure 2 illustrates the parameter estimates for the model. All of the predicted parameter estimates, including those between the structural error terms of the job satisfaction facets, were significant and had the theoretically predicted signs. The model explained 6% of the turnover variance, 26% of the variance of intention to quit, 49% of the variance in pay satisfaction, 34% of the variance in supervisor satisfaction, and 14% of the variance in work satisfaction, respectively.

In order to highlight the relative importance of each of the determinants of employee turnover, we also calculated the total direct and indirect effects of each

EQUITY THEORY AND TURNOVER PROCESS 1027

Table 2

Completely Standardized Indirect and Total Effects of Variables on Intention to Quit and Turnover

Total indirect effect on Total effect

intention to on intention Total effect on Antecedent of turnover quit to quit turnover

Pay rules Pay level Pay administration Pay rules Work pace Pay satisfaction Supervisor satisfaction Work satisfaction Intention to quit

.04

.08

.14*

.05

.09*

.04

.08

.14*

.05

.09* -.14* -.29* -.23 * -

.o 1

.02

.03*

.o 1

.02 -.04 -.07* -.06* .25*

variable on turnover and intention to quit. These results are shown in Table 2. Table 2 shows that the completely standardized total effect of the equity con- structs on turnover are small in magnitude (i.e., .03 or smaller). While those equity constructs clearly have an appreciable effect on the satisfaction facets (i.e., -. 14 to -.29), much of their effect on actual turnover dissipates in the turnover process.

Discussion

All told, multiple criteria resulting from the SEM analysis provide strong sup- port for the model proposed in this study. The predictive validity of the model falls within the range reported in previous research (Hom et al., 1984) or its evo- lutionary predecessor, the Mobley (1977) model (Lee, 1988; Steel, Lounsbury, & Horst, 1981). More importantly, the results also support Adams’ (1963, 1965) notion that inequity perceptions do influence turnover via job satisfaction. Addi- tionally, a role for intention to quit, another mediator extending Adams’ theory, was also supported.

The findings of this study contribute to turnover research in at least two respects. First, and most notably, we suggested and detected variables with an impact on job satisfaction facets that are not yet incorporated in comprehensive

1028 GRIFFETH AND GAERTNER

turnover models (e.g., Hom & Griffeth, 1995; Kim, Price, Mueller, & Watson, 1996; Price & Mueller, 1986). These turnover models presently include distribu- tive justice (Kim et al., 1996; Price & Mueller, 1986) or procedural justice (Hom & Griffeth, 1995) as the only variable relating equity theory to turnover out- comes. The results of this study suggest that turnover research should pay attention to additional variables suggested by equity theory, such as pay adminis- tration, rule administration, and work pace. Of course, whether those or similar constructs (e.g., interactional justice, fairness of work allocation) should be added to contemporary turnover models depends on empirical confirmation in further tests that establish the generalizability of the findings obtained here. It should be kept in mind that a single study conducted in one setting with one sam- ple (n = 180) is not sufficient to establish a generalizable empirical pattern.

Second, the present findings might also contribute to the development of the Lee and Mitchell turnover model (Lee & Mitchell, 1994; Lee, Mitchell, Holtom, McDaniel, & Hill, 1999; Lee, Mitchell, Wise, & Fireman, 1996). That model includes shocks (defined as “particular, jarring events that initiate the psycholog- ical analyses involved in quitting a job”; Lee et al., 1999, p. 6) as one of its major components. An anonymous reviewer suggested that some form of inequity could be considered as such a shock. While generally agreeing with this possibil- ity, we need to add cautiously that this particular study was not concerned with particular events, but rather with general perceptions of organizational circum- stances. Surely, shocks might jar inequity perceptions, which in turn might trig- ger turnover; however, this and other possible relationships linking shocks with equity and turnover is beyond the scope of our investigation. Yet it is noteworthy that the lack of studies relating particular events to employee turnover is a com- mon feature of past turnover research. The identification of this lack and some attempts to overcome it might well be regarded as one of the most outstanding research contributions of Lee, Mitchell, and their associates.

We suggest that our conclusions hold up despite several shortcomings of the present study. In all, we identified six such shortcomings. First, we found that the relationship between equity perceptions with respect to pay rules, pay adminis- tration, and intention to quit is not completely mediated by the three job satisfac- tion facets incorporated in our model. This, in turn, raises the question as to whether other mediators exist that were not incorporated in the present model. Future research may address this question; several possibilities exist, however. For example, future research directed at expanding the number of equity dimen- sions to include coworkers, promotions, working conditions, and organizational communication might be a valuable direction. Perceived inequity regarding these dimensions could then be hypothesized to lead to corresponding dimensions of job dissatisfaction.

Moreover, we omitted organizational commitment from the present model because Adams’ (1963, 1965) theory said little about inequity leading to lowered

EQUITYTHEORYANDTURNOVERPROCESS 1029

commitment. However, an extension of this model could easily accommodate the commitment construct. Similarly, some research has begun to examine inequity effects on organizational citizenship behaviors (OCBs; Harder, 199 1). An expanded model of inequity’s effects on the turnover process may include a mea- sure of OCBs. Finally, research has experimented with the construct of perceived organizational support (POS; Eisenberger, Huntington, Hutchinson, & Sowa, 1986; McFarlane-Shore & Tetrick, 1991). POS is concerned with employees’ perceptions of the organization’s attitude toward them and may well be another potential intermediate linkage between inequity perceptions and intention to quit. By including these constructs in future research, a more comprehensive model of equity and turnover could be developed and tested.

Second, the turnover process outlined in the model is parsimonious, lacking inclusion of variables such as job search, expected utility of withdrawal, and comparison of alternatives with the present job (Hom & Griffeth, 1991; Mobley, 1977). However, omitting mediators does not change estimation of the overall effect of an antecedent turnover cognition on consequent turnover (Bollen, 1989). But, we do include intention to quit in the model-a key variable in the turnover process that is widely regarded as being the most immediate antecedent of turnover (Hom & Griffeth, 1995; Steel & Ovalle, 1984). It is noteworthy that some research recently challenged this view, showing that job search (Kopelman, Rovenpor, & Millsap, 1992) and active job search (Blau, 1993) are more highly related to turnover than are turnover intentions. Models, which include intention to search (rather than intentions to quit) as the immediate precursor of turnover, provide a better fit to cross-sectional data (Sager, Griffeth, & Hom, 1998). Nevertheless, recent comprehensive meta-analyses still show that intention to quit is the best predictor of turnover (Griffeth, Hom, & Gaertner, 2000; Hom & Griffeth, 1995). Thus, more evidence, particularly from longitudinal studies, is needed to effectively invalidate the traditional claim that the most immediate pre- cursor of turnover behavior is turnover intention.

Third, the measurement of voluntary turnover using results from exit inter- views includes error to some extent (Campion, 1991). Some of this error might be a result of inaccuracies in company records. In addition, some of the error is certainly a result of the inherent vagueness of the definitions underlying the mea- surement of turnover types, such as voluntary turnover (Gaertner, Griffeth, & Hagtvedt, 1999; Horn & Griffeth, 1995). Either way, we want to stress that the measurement of voluntary turnover in this study (as in all other studies of employee turnover) includes error that was not specified in the analysis proce- dure since the extent of the error is unknown.

Fourth, following past research, voluntary turnover was measured as a binary variable (see Method section). Aside from the question as to whether voluntary turnover is truly binary (Williams, 1990) or not (Hunter & Schmidt, 1990), it is clear that binary measurement of a dependent variable violates assumptions

1030 GRIFFETH AND GAERTNER

underlying the data-analysis strategy that we chose in this study (e.g., Bollen, 1989). Hence, the question arises as to why we did not use a data-analysis strat- egy that is more attuned to binary dependent variables, such as logistic regression (Huselid & Day, 199 1) or survival analysis (Morita, Lee, & Mowday, 1989).

Our argument is threefold. First, SEM has several advantages over the other two methods, particularly when an investigator is interested in testing entire models consisting of latent variables. Those advantages include the flexibility to specify links and equations freely as the theory suggests, the possibility to explic- itly specify measurement error, the possibility to assess the fit and parameters of all structural equations simultaneously, and many more ( e g , Bollen, 1989; Hayduk, 1987; Rigdon, 1998). Second, prior simulations conducted in turnover research found that the overall fit of the model as well as most of the parameter estimates are relatively unaffected by violations of statistical assumptions through the inclusion of a dummy-coded dependent variable (Aquino, Criffeth, Gaertner, & Lee, 1998; Mueller, Boyer, Price, & Iverson, 1994). The parameter estimate linking intention to quit to actual turnover is, of course, affected by this violation of assumptions, supposedly in that this parameter estimate is biased downward (Kemery, Dunlap, & Griffeth, 1988). In order to provide further evi- dence, we reran the same model (Figure 2), with turnover excluded, using LISREL 8. The resulting overall fit of the model was very similar to the overall fit of the model with turnover included, x2( 15, N = 180) = 2 1.9 1 , p = .1 I , ns (RMSEA = .049, NFI = .96, GFI = .97, CFI = .98), and would not suggest any other conclusions than those that are outlined. Moreover, the unstandardized and completely standardized parameter estimates obtained were almost identical in terms of magnitude (maximum difference was .01 for the completely standard- ized parameter estimate linking supervisory satisfaction and intention to quit). The t values were, for all intents and purposes, identical, too (maximum differ- ence was .06 for the t value of the parameter estimate linking supervisory satis- faction and intention to quit). Third, traditional SEM procedures for the modeling of binary dependent variables without violating statistical assumptions ( leg , weighted least square procedures) have very large sample-size requirements (Joreskog & Silrbom, 1989), which exceed the sample size available for this study. Other alternatives with less prohibitive sample-size requirements are pres- ently being developed ( e g , the robust weighted least square procedure underly- ing the Mplus computer program; Muthen & Muthen, 1999). However, they have not been tested in published Monte Carlo simulations yet.

As the fifth shortcoming, we must reemphasize that the internal consistency of pay level (a = .66) fell marginally below the Nunnally and Bernstein ( 1 994) standard (a = .70). Note that the SEM procedure controls for unreliability, which limits the biasing effect of unreliability on the resulting parameter estimates (Bollen, 1989). It might be that some researchers take this marginal internal con- sistency as indicating bias in measurement. Unfortunately, we cannot confirm or

EQUITY THEORY AND TURNOVER PROCESS 1031

refute such a claim. Note, however, that the measure of pay level is taken from a scale that was tested in prior research (OFQ; Dittrich & Carrell, 1979) and that its internal consistency in fact falls only marginally short of the Nunnally and Bernstein standard (i.e., A a = .04).

A final shortcoming of this study refers to the unmeasured variable problem. It is well known that unmeasured variables that are related to both the dependent and independent variable can have a distorting effect on parameter estimates (e.g., Neter et al., 1990). Unfortunately, our model does not include or control for any of the dozens of environmental, structural, individual, and demographic vari- ables that are hypothesized or found to be related to turnover, job satisfaction, and equity perceptions (see Kim et al., 1996, for a list of some of those vari- ables). Accordingly, the results obtained here must be interpreted cautiously. We conclude that the results of this study suggest that turnover researchers should pay more intention to the multifaceted nature of equity. Future research should attempt to replicate the findings obtained here in other settings and incorporate a multifaceted view of equity in future tests of comprehensive turnover models, such as the Price and Mueller model (e.g., Kim et al., 1996) or the Hom and Griffeth (1995) model.

In sum, the results of this study highlight the impact of several aspects of equity perceptions on satisfaction, intention to quit, and turnover. Organizations that attempt to keep the turnover among their work force low are advised to pay close attention to the perceived fairness of their human resource management systems. Based on the results of this study, we conclude that the perceived fair- ness of the pay level alone is not sufficient to reduce employee turnover. The per- ceived fairness of other aspects of the human resource management system, such as pay rules, treatment by supervisors, and work pace appears to be no less important than perceived pay-level fairness for predicting employee turnover. Yet, consistent with earlier turnover research, the present model explained only 6% of the variance in turnover. Much research remains to be done before volun- tary employee turnover is understood.

References

Adams, J. S. (1963). Toward an understanding of inequity. Journal of Abnormal and Social Psychology, 67,422-436.

Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.), Advances in experimental socialpsychology (Vol. 2, pp. 267-299). New York, NY: Academic Press.

Aquino, K., Griffeth, R. W., Gaertner, S., & Lee, M. (1998, March). Test of the structural validity and generalitability of a referent cognitions model: A cross-cultural replication. San Diego, CA: Proceedings of the Southwestern Academy of Management.

1032 GRIFFETH AND GAERTNER

Bentler, P. M. (1 990). Comparative fit indexes in structural models. Psychologi- cal Bulletin, 107,238-264.

Bentler, P. M., & Bonnett, D. G. ( 1 980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin, 88, 588-606.

Bentler, P. M., & Chou, C. (1987). Practical issues in structural modeling. Socio- logical Methods and Research, 16, 78- 1 17.

Berg, T. R. (1991). The importance of equity perception and job satisfaction in predicting employee intent to stay at television stations. Group and Organiza- tion Studies, 16, 268-284.

Bies, R. J., & Moag, J. S. (1986). lnteractional justice: Communication criteria of fairness. Research on Negotiation in Organizations, 1,43-55.

Bies, R. J . , & Shapiro, D. L. (1 987). Interactional fairness judgments: The influ- ence of causal accounts. Social Justice Research, 1, 199-2 18.

Blau, G. (1993). Further exploring the relationship between job search and vol- untary individual turnover. Personnel Psychology, 46,3 13-330.

Bollen, K. A. (1989). Structural equations with latent variables. New York, NY: John Wiley & Sons.

Bretz, R. D., & Thomas, S. L. (1992). Perceived equity, motivation, and final- offer arbitration in major league baseball. Journal of Applied Psychology, 77,

Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 136- 162). Newbury Park, CA: Sage.

Campion, M. A. (1991). Meaning and measurement of turnover. Comparison of alternative measues and recommendations for research. Journal of Applied

Can; S. C., McLoughlin, D., Hodgson, M., & Maclachlan, M. (1996). Effects of unreasonable pay discrepancies for under- and overpayment on double demo- tivation. Genetic, Social, and General Psychology Monographs, 122, 475- 494.

Dittrich, J . E., & Carrell, M. R. (1979). Organizational equity perceptions, employee job satisfaction, and departmental absence and turnover rates. Organizational Behavior and Human Performance, 24,29-40.

Eisenberger, R., Huntington, R., Hutchinson, S., & Sowa, D. (1986). Perceived organizational support. Journal of Applied Psychology, 71, 500-507.

Folger, R., & Konovsky, M. A. (1989). Effects of procedural and distributivejus- tice on reactions to pay raise decisions. Academy of Management Journal, 32,

Gaertner, S., Griffeth, R. W., & Hagtvedt, R. (1999, August). The conceptual underpinnings of continuous turnover type measurement: Appking fuzzy logic. Paper presented at the annual meeting of the Academy of Management, Chicago, IL.

280-287.

Psychology, 76, 199-2 12.

1 1 5-1 30.

EQUITY THEORY AND TURNOVER PROCESS 1033

Glass, R. S., & Wood, W. A. (1 996). Situational determinants of software piracy: An equity theory perspective. Journal of Business Ethics, 15, 1 189-1 198.

Greenberg, J. (1990). Employee theft as a reaction to underpayment inequity: The hidden cost of pay cuts. Journal ofAppliedPsychology, 75,561-568.

Griffeth, R. W., Hom, P. W., & Gaertner, S. (2000). A meta-analysis of anteced- ents and correlates of employee turnover: Update, moderator tests, and research implications for the next millennium. Journal of Management, 26,

Harder, J. W. (1991). Equity theory versus expectancy theory: The case of major league baseball free agents. Journal of Applied Psychology, 76,458-464.

Harder, J. W. (1992). Play for pay: Effects of inequity in a pay-for-performance context. Administrative Science Quarterly, 37, 32 1-335.

Hayduk, L. (1987). Structural equation modeling with LISREL: Essentials and advances. Baltimore, MD: Johns Hopkins University Press.

Hom, P. W., & Griffeth, R. W. (1991). Structural equations modeling test of a turnover theory: Cross-sectional and longitudinal analysis. Journal of Applied

Hom, P. W., & Griffeth, R. W. (1995). Employee turnover. Cincinnati, OH: South- Western College Publishing.

Hom, P. W., Griffeth, R. W., & Sellaro, L. (1984). The validity of Mobley’s ( 1 977) model of employee turnover. Organizational Behavior and Human Performance, 34, 14 1 - 174.

Huber, V. L., Seybolt, P. M., & Venemon, K. (1992). The relationships between individual inputs, perceptions, and multidimensional pay satisfaction. Jour- nal of Applied Social Psychology, 22, 1356- 1373.

H u h , C . L., Ronowski, M., & Hachiya, D. (1985). Alternative opportunities and withdrawal decisions: Empirical and theoretical discrepancies and an integration. Psychological Bulletin, 97,233-250.

Hunter, J. E., & Schmidt, F. L. (1990). Dichotomization of continuous variables: The implications for meta-analysis. Journal ofApplied Psychology, 75,334-349.

Huselid, M. A., & Day, N. E. (1991). Organizational commitment, job involve- ment, and turnover: A substantive and methodological analysis. Journal of Applied Psychology, 76, 380-391.

Huseman, R. C., Hatfield, J. D., & Miles, E. W. (1985). Test for individual per- ceptions of job equity: Some preliminary findings. Perceptual and Motor Skills, 61, 1055-1064.

Huseman, R. C., Hatfield, J. D., & Miles, E. W. (1987). A new perspective on equity theory: The equity sensitivity construct. Academy of Management Review, 12,222-234.

Iverson, R. D., & Roy, P. (1994). A causal model of behavioral commitment: Evi- dence from a study of Australian blue-collar employees. Journal of Manuge- ment, 20, 15-4 1.

463-488.

Psychology, 76, 350-366.

1034 GRIFFETH AND GAERTNER

Johnson, W. R., & Johnson, G. J. (1991). The effects of equity perceptions on union and company commitment. Journal of Collective Negotiations in the Public Sector, 20,235-244.

Joreskog, K. G., & Sorbom, D. (1984). LISREL 6 user’s guide. Moresville, IN: Scientific Software.

Joreskog, K. G., & Sorbom, D. (1989). LISREL 7: A guide to the program and applications (2nd ed.). Chicago, IL: SPSS.

Joreskog, K. G., & Sbrbom, D. ( I 993). LISREL 8: Structural equation modeling with the SIMPLlS command language. Hillsdale, NJ: Scientific Software.

Joshi, K. (1990). An investigation of equity as a determinant of user information satisfaction. Decision Sciences, 21,786-807.

Kemery, E. R., Dunlap, W. P., & Griffeth, R. W. (1988). Correction for range restrictions in point-biserial correlations. Journal of Applied PsychoEogy, 73,

Kim, S., Price, J. L., Mueller, C. W., & Watson, T. W. (1996). The determinants of career intent among physicians at a U.S. Air Force hospital. Human Rela- tions, 49,947-975.

King, W. C., & Miles, E. W. (1994). The measurement of equity sensitivity. Jour- nal of Occupational and Organizational Psychology, 67, 133-142.

King, W. C., Miles, E. W., & Day, D. D. (1993). A test and refinement of the equity sensitivity construct. Journal of Organizational Behavior, 14,301 - 317.

Konovsky, M. A., Folger, R., & Cropanzano, R. (1987). Relative effects of proce- dural and distributive justice on employee attitudes. Representative Research in Social Psychology, 17, 15-24.

Kopelman, R. E., Rovenpor, J. L., & Millsap, R. E. (1 992). Rationale and con- struct validity evidence for the Job Search Behavior Index: Because inten- tions (and New Year’s resolutions) often come to naught. Journal of Vocational Behavior, 40,269-287.

Lee, T. W. (1988). How job satisfaction leads to employee turnover. JournaZ of Business and Psychology, 2,263-271.

Lee, T. W., & Mitchell, T. R. (1994). An alternative approach: The unfolding model of voluntary employee turnover. Academy of Management Review, 19,

Lee, T. W., Mitchell, T. R., Holtom, 9. C., McDaniel, L. S., & Hill, J. W. (1999). The unfolding model of voluntary turnover: A replication and extension. Academy of Management Journal, 42,450-462.

Lee, T. W., Mitchell, T. R., Wise, L., & Fireman, S. (1996). An unfolding model of voluntary employee turnover. Academy of Management Journal, 39, 5-36.

MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1993). Power analysis and determination of sample size for covariance structure modeling. Psycho- logical Metho& 1, 130-149.

688-69 1 .

5 1-89.

EQUITY THEORY AND TURNOVER PROCESS 1035

McFarlane-Shore, L., & Tetrick, L. E. (1991). A construct validity study of the Survey of Perceived Organizational Support. Journal of Applied Psychology,

McFarlin, D. B., & Sweeney, P. D. (1992). Distributive and procedural justice as predictors of satisfaction with personal and organizational outcomes. Acad- emy of Management Journal, 35,626-637.

Miles, E. W., Hatfield, J. D., & Huseman, R. C. (1994). Equity sensitivity and outcome importance. Journal of Organizational Behavior, 15, 585-596.

Mobley, W. H. (1977). Intermediate linkages in the relationship between job sat- isfaction and employee turnover. Journal ofAppIied Psychology, 63,237-240.

Mobley, W. H., Griffeth, R. W., Hand, H. H., & Meglino, B. M. (1979). Concep- tual and empirical analysis of military recruit training attrition. Psychological Bulletin, 86,493-522.

Morita, J. G., Lee, T. W., & Mowday, R. T. (1989). Introducing survival analysis to organizational researchers: A selected application to turnover research. Journal of Applied Psychology, 74,280-292.

Mueller, C. W., Boyer, E. M., Price, J. L., & Iverson, R. E. (1994). Employee attachment and noncoercive conditions of work: The case of dental hygien- ists. Work and Occupations, 21, 179-212.

Muthen, L. K., & Muthen, B. 0. (1 999). MpIus user k guide: The comprehensive modeling program for applied researchers. Los Angeles, CA: Muthen & Muthen.

Neter, J., Wasserman, W., & Kutner, M. H. (1990). Applied linear statistical models: Regression, analysis of variance, and experimental design. Home- wood, IL: Irwin.

Nunnally, J. C., & Bernstein, 1. H. (1994). Psychometric theory (3rd ed.). New York, NY: McGraw-Hill.

Oldham, G. R., Kulik, C. T., Ambrose, M. L., Stepina, L. P., & Brand, J. F. (1986). Relations between job facet comparisons and employee reactions. Organizational Behavior and Human Decision Processes, 38,28-47.

Perry, L. S. (1 993). Effects of inequity on job satisfaction and self-evaluation in a national sample of African-American workers. Journal of Social Psychology,

Price, J. L., & Mueller, C. W. (1981). A causal model of turnover for nurses. Academy of Management Journal, 24,543-565.

Price, J. L., & Mueller, C. W. (1986). Absenteeism and turnover of hospital employees. Greenwich, C T JAl.

Randall, C. S., & Mueller, C. W. (1995). Extensions of justice theory: Justice evaluations and employees’ reactions in a natural setting. Social Psychology Quarterly, 58, 178- 194.

Rice, R. W., Phillips, S. M., & McFarlin, D. B. (1990). Multiple discrepancies and pay satisfaction. Journal of Applied Psychology, 75,386-393.

76,637-643.

133,565-573.

1036 GRIFFETH AND GAERTNER

Rigdon, E. E. (1998). Structural equation modeling. In G. A. Marcoulides (Ed.), Modern methods for business research (pp. 25 1-294). Mahwah, NJ: Erlbaum.

Ronen, S. (1986). Equity perception in multiple comparisons: A field study. Human Relations, 39, 333-345.

Sager, J. K., Griffeth, R. W., & Hom, P. W. (1998). A comparison of structural models representing turnover cognitions. Journal of Vocational Behavior, 53,

Scholl, R. W., Cooper, E. A., & McKenna, J. F. (1987). Referent selection in determining equity perceptions: Differential effects on behavioral and attitu- dinal outcomes. Personnel Psychology, 40, 11 3-124.

Settoon, R. P., Bennett, N., & Liden, R. C. ( I 996). Social exchange in organiza- tions: Perceived organizational support, leader-member exchange, and employee reciprocity. Journal of Applied Psychology, 81,2 19-227.

Sheehan, E. P. (1 993). The effects of turnover on the productivity of those who stay. Journal of Social Psychology, 133, 699-706.

Singh, P. (1994). Perception and reactions to inequality as a function of social comparison referents and hierarchical levels. Journal of Applied Social Psy-

Smith, P. C., Kendall, L. M., & Hulin, C. L. (1969). The measurement ofsatisfac- tion in work and retirement: A strategy for the study of attitudes. Chicago, IL: Rand-McNally.

Spector, P. E. ( 1997). Job satisfaction: Application, assessment, causes, and con- sequences. Thousand Oaks, CA: Sage.

Steel, R. P., Lounsbury, J., & Horst, W. (1 98 1, March). A test of the internal and external validity of Mobley’s model of employee turnover. In Proceedings of the 24th annual conference of the Midwest Academy of Management. Carbondale, IL: Southern Illinois University.

Steel, R. P., & Ovalle, N. K. (1 984). A review and meta-analysis of research on the relationship between behavioral intentions and employee turnover. Jour- nal of Applied Psychology, 69,673-686.

Steiger, J. H., & Lind, J. C. (1980, May). Statistically based tests for the number of common factors. Paper presented at the annual meeting of the Psychomet- ric Society, Iowa City, IA.

Summers, T. P., & DeNisi, A. S. (1990). In search of Adams’ referent other: Reexamination of referents used in the evaluation of pay. Human Relations, 43,497-5 1 1.

Summers, T. P., & Hendrix, W. H. (1991). Modeling the role of pay equity per- ceptions: A field study. Journal of Occupational Psychology, 64, 145-1 57.

Sweeney, P. D. (1 990). Distributive justice and pay satisfaction: A field test of an equity theory prediction. Journal of Business and Psychology, 4, 329- 341.

254-273.

chology, 24,557-565.

EQUITY THEORY AND TURNOVER PROCESS 1037

Telly, C., French, W., & Scott, W. (1971). The relationship of inequity to turnover among hourly workers. Administrative Science Quarterly, 16, 164- 172.

Van Dierendonck, D., Schaufeli, W. B., & Sixma, W. B. (1994). Burnout among general practitioners: A perspective from equity theory. Journal of Social and Clinical Psychology, 13, 86-100.

Vecchio, R. P., Griffeth, R. W., & Horn, P. W. (1986). The predictive utility of the vertical dyad linkage approach. Journal of Social Psychology, 128,6 17-625.

Wilhelm, C. C., Herd, A. S., & Steiner, D. D. (1993). Attributional conflict between managers and subordinates: An investigation of leader-member exchange effects. Journal of Organizational Behavior, 14,53 1-544.

Williams, C. R. (1990). Deciding when, how, and if to correct turnover correla- tions. Journal ofApplied Psychology, 75,732-737.

Williams, L. J., & Hazer, J. T. (1986). Antecedents and consequences of satisfac- tion and commitment in turnover models: A reanalysis using latent variable structural equation methods. Journal of Applied Psychology, 71,2 19-23 1.