ABSTRACT—Most research on voluntary turnover has fo- cused on dissatisfaction-induced and rational decision- making processes, with some attention paid to external market influences. This focus leaves unexplained a large portion of the variance in why people choose to quit a job. Recently, however, researchers are considering the alter- native ways that the turnover process is enacted, as well as what businesses can do to prevent turnover.

KEYWORDS—turnover; embeddedness

Turnover has been and remains an active area of theorizing and

empirical research within the domain of organizational psy- chology. Older theories are being retested and modified, new

theorizing is being introduced, and current research is providing better answers to the main research question: Why do people

voluntarily leave a job? In this article, we selectively summarize the huge body of work on the psychology of voluntary turnover and present the current directions in turnover theory and re-

search.

WHY PEOPLE LEAVE

Traditional research on turnover focuses on negative job atti- tudes (e.g., low levels of job satisfaction) as the cause of leaving. In their seminal work, March and Simon (1958) proposed a

psychological explanation of turnover that is based on individ- uals’ utility functions: When outcomes (such as pay or promotion

opportunities) are too low relative to the employee’s expect- ations, an employee becomes dissatisfied and motivated to leave

the organization, increasing his or her ‘‘desirability of move- ment.’’ Turnover, then, becomes a function of the extent of this desirability combined with the perceived ease of movement (i.e.,

number of perceived job alternatives). Much research and the- orizing followed. Griffeth, Hom, and Gaertner (2000) reported

that across the many intellectually rigorous models focusing on

job satisfaction and perceived alternatives, the ability to predict voluntary turnover remained remarkably weak, with variance

explained hovering around 5%. In response to these disappointing findings, Lee and Mitchell

(1994) used image theory from psychology (Beach & Mitchell, 1998) to develop their unfolding model of voluntary turnover. The unfolding model describes different psychological paths that people follow when they decide to leave an organization. The central contribution of this approach is that it shifted the-

orizing from an assumption that turnover is always an evaluative and rational process to a broader model of how decisions are

actually made. This analysis revealed that although some de- cisions to quit are probably quite consistent with standards of

expected-value rationality, a great many others are driven by more intuitive or routinized decision processes.

IMAGE THEORY AND THE UNFOLDING MODEL

Image theory is perhaps the most viable and elaborate alterna-

tive to the notion of decisions based on subjective expected utility. In contrast to the idea that people generally do or should

make calculations that maximize utility, image theory suggests that some types of incoming information (e.g., a job offer) prompt a comparison of that information to three job-related images. As

an automatic, though conscious, process, the individual first will compare the information withhis or her value image (the person’s

set of important values regarding his or her job). Next, the individual will compare the incoming information with the

trajectory image (the person’s set of goals that motivate job behavior). Last, the individual will compare the information with the strategic image (the behavioral tactics and strategies that the

person believes are effective in attaining job-related goals). Should the incoming information contain an alternative that

passes this screening process (i.e., being compatible with the three images), the individual then compares the alternative with the status quo. Typically, the status quo wins and the person does

nothing. Should the person have more than one alternative that passes the screening process, he or she will then conduct more

rational cost/benefit analyses among alternatives. Address correspondence to Wendy S. Harman, Truman State University, 100 E. Normal St., Kirksville, MO 63501; e-mail: [email protected].

CURRENT DIRECTIONS IN PSYCHOLOGICAL SCIENCE

Volume 16—Number 1 51Copyright r 2007 Association for Psychological Science at Erasmus Univ Rotterdam on July 26, 2010cdp.sagepub.comDownloaded from

http://cdp.sagepub.com/

The unfolding model labels incoming information that leads to

image considerations as shocks. Shocks lead the person to consider leaving his or her job. The shock can be internal or external to the

individual, and it can be negative (e.g., a fight with the boss), positive (e.g., winning the lottery), or neutral (e.g., an unanticipated

job offer). Additionally, the social and cognitive context in which the shock occurs provides the decision frame or frame of reference within which the employee interprets the meaning of the shock.

The unfolding model introduces five paths (i.e., distinct sets of psychological processes) that lead to voluntary turnover (see

Table 1). In the first path, the shock triggers a preexisting script. The person then engages the preexisting script and leaves

without considering alternatives or his or her attachment to the organization. Suppose, for example, an employee has the fol- lowing pregnancy script: If I become pregnant, then I will quit

work to stay home with the new baby. Once the information about being pregnant is known, the script is enacted without further

consideration. The decision is automatic because there is an existing script that dictates the decision. The second path also contains a shock that triggers leaving without the consideration

of job alternatives. The difference is that, in this second case, there is no prior action script. Instead, incoming information is

interpreted as violating a person’s value, goal, or strategic im- ages, and leaving occurs without further deliberation. This

process is fairly automatic; the question of leaving is decided once the individual decides that one or more image violations have occurred. An example of this path would be a situation in

which a strongly disliked coworker gets promoted to be an in- dividual’s boss, violating that individual’s belief that enjoyable

work involves having a friendly relationship with the supervisor (strategy image violation), and the individual leaves without considering other job alternatives.

The third path in the unfolding model contains a shock that triggers an evaluation via the three images (values, trajectory,

strategic) of the current job; if the information contained in the

shock is not compatible with the images, the individual then considers leaving. This path can lead to a deliberate search for

job alternatives and certainly involves an evaluation of at least one alternative. For example, an unexpected job offer may

prompt someone to compare the current job to the unexpected job offer. Paths 4a and 4b do not contain shocks. Path 4a de- scribes a situation in which a person’s job satisfaction becomes

so low that he or she leaves without considering job alternatives. Path 4b is the more traditional view of turnover in which low

satisfaction leads to a deliberate search, evaluation of alterna- tives, an intention to leave, and subsequent turnover.

Several recent turnover studies have provided support for the unfolding model. To test the unfolding model, Lee, Mitchell, Wise, and Fireman (1996) used multiple case studies of a sample

of nurses who had voluntarily quit their jobs. They found that of those sampled, 14% reported following the first decision path,

14% followed the second, 32% followed the third, 18% followed 4a, and 23% followed 4b (percentages are rounded). These findings provided preliminary evidence for the model and for the

contention that people use these four distinct psychological processes when engaged in a turnover decision. Lee, Mitchell,

Holtom, McDaniel, and Hill (1999) conducted a replication and extension using a sample of accountants from the Big 6 ac-

counting firms. They reported substantial support for a slightly modified model, with 93% of the subjects fitting clearly into one of the four paths. Differences with respect to job type were un-

covered when comparing the accountant sample with the sample of nurses from Lee et al. (1996). Nurses left significantly more

often via paths 1, 2, and 4a than did the accountants, who left significantly more often by path 3. Other recent studies support the importance of job type in determining turnover path (Kam-

meyer-Mueller, Wanberg, Glomb, & Ahlburg, 2005; Maertz, Stevens, & Campion, 2003; Sims, Drasgow, & Fitzgerald, 2005).

TABLE 1

Paths by Which the Turnover Decision May Unfold Over Time

Initiating event ! Cognitive / emotional

process ! Search behavior ! Quit decision

Path 1 Shock (e.g., pregnancy) ! Prompts quitting script enactment

! None ! Automatic

Path 2 Shock (e.g., unpleasant new boss)

! Comparison of shock to images leads to high dissatisfaction

! None ! Fairly automatic

Path 3 Shock (e.g., unexpected job offer)

! Comparison of shock to images leads to relative dissatisfaction

! Search for alternatives

! Deliberate

Path 4a No shock ! Accumulating dissatisfaction

! None ! Fairly automatic

Path 4b No shock ! Accumulating dissatisfaction

! Search for alternatives

! Deliberate

Note. Moving from left to right, paths 1–3 include an initiating event (shock) that leads to the quit decision. Paths 4a and 4b do not contain shocks; instead the process begins with accumulating dissatisfaction that leads to the turnover decision.

52 Volume 16—Number 1

Employee Turnover

at Erasmus Univ Rotterdam on July 26, 2010cdp.sagepub.comDownloaded from

http://cdp.sagepub.com/

Specific Contributions The unfolding model moves traditional turnover theory away from a reliance on the rationalistic approaches of previous the-

ories to a more descriptively accurate model grounded in current psychological thinking (e.g., image theory). As a consequence,

the unfolding model makes a number of specific contributions. First, it introduces the notion of automated scripts into the turnover process (path 1). Second, the model incorporates

shocks (positive, neutral, or negative) as catalysts to turnover (paths 1, 2, and 3). Third, in contrast to the traditional single

sequential process, the multiple paths of the unfolding model allow for greater explanatory power with distinctive psycholog-

ical processes in each path. Fourth, the model highlights the possibility that job satisfaction may have no influence on the decision to quit—that is, people who are satisfied still may leave

(paths 1 and 3). Fifth, the relative speed of the quit decisions— immediate in paths 1 and 2 but slower in paths 3 and 4b—is

introduced to the turnover literature. Sixth, and finally, the un- folding model allows for the possibility that turnover could happen even in the absence of job alternatives (paths 1, 2, and

4a). In sum, the unfolding model represents a major alternative to traditional theories of leaving.

WHY PEOPLE STAY

As we have reviewed, traditional turnover models hold that

employees leave because of negative job attitudes and stay be- cause of positive job attitudes. Accordingly, the vast majority of

scholarly efforts presume that staying is simply the opposite of leaving (e.g., increase job satisfaction to retain employees),

though this is not necessarily the case. Research by Rusbult, Farrell, and colleagues (1983, 1988) showed that turnover de- cisions are influenced by people’s comparisons between the

investments made in their job or organization, the rewards they receive, the quality of alternatives, and the costs associated with

working for a particular organization—and these comparisons change over time. As this research and the unfolding model suggest, employee retention issues are not simplistic. Mitchell,

Holtom, Lee, Sablynski, and Erez (2001) proposed a construct called job embeddedness to explain why employees remain in an organization. Job embeddedness describes a web of forces that cause people to feel they cannot leave their job. The critical

components to job embeddedness include the extent to which people are linked with others or to activities, the extent to which their jobs and communities fit with other aspects of their lives,

and the ease with which their respective links can be broken— that is, what they would sacrifice if they left. These three di-

mensions are identified as links, fit, and sacrifice, respectively, and are concerned with both on-job and off-the-job experiences. Links refers to formal and informal connections that a person

has with other individuals and institutions. The more connected an individual is with the organization (e.g., belongs to work-

groups) and with the community (e.g., affiliated with local clubs,

interest groups, or churches), the more embedded he or she is.

Fit is the individual’s perceived compatibility with the organ- ization and with the community. The employee’s personal values

and career goals need to be congruent with the larger organ- izational culture; this congruence allows the employee to feel

tied personally and professionally to the organization. In addi- tion, individuals need to feel as though they and their family also fit with the community in which they live. Again, the better the

fit, the more likely the person is to stay. Sacrifice refers to the perceived cost of leaving. These costs may be material or psy-

chological. Leaving may entail giving up the advantages asso- ciated with tenure (e.g., big office, vacation time), as well as the

personal losses such as companionship with colleagues or perks unique to the organization. Community sacrifices are relevant only if the individual needs to move to a new location. The loss of

the sense of belonging to a community (including giving up such things as ticketsto the local football team, the home inhabited for

20 years) can influence the community sacrifice dimension. The empirical evidence for job embeddedness is substantial.

In a two-sample test of grocery employees and hospital workers,

Mitchell et al. (2001) found that employees’ self-rated embed- dedness scores at time 1 predicted which employees had left the

organization a year later at time 2. Additionally, the researchers found that overall job embeddedness predicts variance in turn-

over over and above that explained by the major variables in most turnover research (i.e., job satisfaction, organizational commitment, job search, and job alternatives). Further, Lee,

Mitchell, Sablynski, Burton, and Holtom (2004) sought to extend the original formulation by considering on- and off-the-job

embeddedness separately. These scholars reported that off-the- job embeddedness predicted absences and turnover (over and above that of job satisfaction and organizational commitment),

whereas on-the-job embeddedness did not. Conversely, they also reported that on-the-job embeddedness predicted organiza-

tional citizenship behavior and in-role job performance (again, over and above satisfaction and commitment), whereas off-the-

job embeddedness did not.

Specific Contributions The job embeddedness construct makes two main contributions to the turnover literature. First, it significantly expands the

scope of variables researchers consider when trying to under- stand why people remain in a job, such as the inclusion of nonattitudinal determinants of turnover like structural links to

other people. Second, embeddedness includes consideration of off-the-job factors like fit with one’s neighbors and community—

a contribution that is more subtle and theoretical in nature. Job embeddedness is premised on the notion that many people rarely consider leaving their jobs because they are so immersed in their

environments. This realistic understanding of human psychol- ogy can guide continued efforts to understand why people leave

and why they stay.

Volume 16—Number 1 53

Wendy S. Harman et al.

at Erasmus Univ Rotterdam on July 26, 2010cdp.sagepub.comDownloaded from

http://cdp.sagepub.com/

FUTURE RESEARCH AND CONCLUSIONS

The unfolding model and the job embeddedness construct are significant advances in the turnover literature, but researchers

still need to determine the basic science behind the models. It would be beneficial to integrate this work more fully with

judgment and decision-making research regarding deliberate and automatic processes. For example, are there situational

triggers that make people aware of quit options (e.g., other people leaving, changes outside of work)? This sort of basic perceptual research could greatly increase understanding

of the actual phenomenon of quitting. In addition, research that addresses the intersection of the unfolding model and

job embeddedness construct would be exceptionally helpful. For example, does job embeddedness buffer the effects of shocks from the unfolding model? Additionally, research look-

ing into practical interventions that are informed by the theories would be valuable. For example, which types of

shocks in the unfolding model are controllable by the organ- ization, and how might organizations work to embed employees

in the organization or community? Because many organiza- tional psychologists work in applied settings, endeavoring to develop applications of this research could be particularly

beneficial. In closing, premised on psychologically grounded models of

cognition and decision-making, both the unfolding model and the job embeddedness construct provide means of looking be- yond job satisfaction to other factors important to the voluntary

turnover process. We look forward to new work that builds on these tools.

Recommended Reading Lee, T.W., & Mitchell, T.R. (1994). (See References)

Lee, T.W., Mitchell, T.R., Sablynski, C.J., Burton, J.P., & Holtom, B.C. (2004). (See References)

Mitchell, T.R., Holtom, B.C., Lee, T.W., Sablynski, C.J., & Erez, M. (2001). (See References)

REFERENCES

Beach, L.R., & Mitchell, T.R. (1998). The basics of image theory. In L.R. Beach (Ed.), Image theory: Theoretical and empirical foundations (pp. 3–18). Hillsdale, NJ: Erlbaum.

Griffeth, R.W., Hom, P.W., & Gaertner, S. (2000). A meta-analysis of antecedents and correlates of employee turnover: Update, mod- erator tests, and research implications for the next millennium. Journal of Management, 26, 463–488.

Kammeyer-Mueller, J.D., Wanberg, C.R., Glomb, T.M., & Ahlburg, D. (2005). The role of temporal shifts in turnover processes: It’s about time. Journal of Applied Psychology, 90, 644–658.

Lee, T.W., & Mitchell, T.R. (1994). An alternative approach: The un- folding model of voluntary employee turnover. Academy of Man- agement Review, 19, 51–89.

Lee, T.W., Mitchell, T.R., Holtom, B.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., Sablynski, C.J., Burton, J.P., & Holtom, B.C. (2004). The effects of job embeddedness on organizational citi- zenship, job performance, volitional absences, and voluntary turnover. Academy of Management Journal, 47, 711–722.

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.

Maertz, C.P., Stevens, M.J., & Campion, M.A. (2003). A turnover model for the Mexican maquiladoras. Journal of Vocational Behavior, 63, 111–135.

March, J.G., & Simon, H.A. (1958). Organizations. New York: Wiley. Mitchell, T.R., Holtom, B.C., Lee, T.W., Sablynski, C.J., & Erez, M.

(2001). Why people stay: Using job embeddedness to predict voluntary turnover. Academy of Management Journal, 44, 1102– 1121.

Rusbult, C.E., & Farrell, D. (1983). A longitudinal test of the invest- ments model: The impact on job satisfaction, job commitment, and turnover variations in rewards, costs, alternatives, and invest- ments. Journal of Applied Psychology, 68, 429–438.

Rusbult, C.E., Farrell, D., Rogers, G., & Mainous, A.G. (1988). Impact of exchange variables on exit, voice, loyalty, and neglect: An in- tegrative model of responses to declining job satisfaction. Academy of Management Journal, 31, 599–627.

Sims, C.S., Drasgow, F., & Fitzgerald, L.F. (2005). The effects of sexual harassment on turnover in the military: Time-dependent modeling. Journal of Applied Psychology, 90, 1141–1152.

54 Volume 16—Number 1

Employee Turnover

at Erasmus Univ Rotterdam on July 26, 2010cdp.sagepub.comDownloaded from

View publication statsView publication stats

https://www.researchgate.net/publication/263333287