Turnover and Retention in High Tech Industry


These days, employee turnover rate is growing, especially in the high tech industries. There are both internal and external causes that affect such high turnover rate. I will discuss these phenomena and necessary prescriptions using (1) traditional turnover models, (2) Lee & Mitchell's theory as an alternative approach, and (3) various selection literatures. Note that I consider only voluntary turnover, not involuntary one (e.g., lay-off).

Cognitive, affective and behavioral antecedents in traditional models

To understand cognitive, affective and behavioral antecedents in traditional models, I will briefly summarize the development of traditional modals first, then discuss each antecedent that related to the models.

The historical development of traditional models

The first model of employee turnover might be March & Simon's (1958) organizational equilibrium model. This model considers both internal and external factors that affect employee turnover. Perceived desirability of leave is the internal factor. People think whether they should leave the organizations based on the various aspects that include job satisfaction. Ease of moving refers to the external factor such as job market conditions. People assess the ease of moving though such factors as number of alternatives, costs, and so on. March & Simon's model proposes that both of perceived desirability of leave and easy of moving affect employee voluntary turnover. Subsequent research focused on job satisfaction and number of alternatives as the major antecedents of employee turnover.

Porter & Steers (1973) proposed the met expectation model. In this model, people have some expectations about the organizations in which they will join, and if the expectations are unmet, people are likely to leave the organization. This model suggests that different people may have different expectations about the organizations and this difference may affect individual differences about whether people leave organizations or not.

Mobley (1977) extended the path from job dissatisfaction to employee turnover. This model is called intermediate linkage model that is one of the dominant perspectives in traditional turnover models. This model proposes that if a person feels dissatisfied with organizations, he or she will think about quitting. Next, he or she will search jobs in order to find alternatives and calculate expected utility of each alternative that he or she has. Then, if the alternative is superior to staying the organization, it leads to the intention to quit and actual turnover. This model was empirically tested by Hom & Griffeth and couple of other researchers, and was found to be valid.

Hulin (1995) focuses on the attitude-withdrawal behavior relationship in understanding employee turnover. That is, employee turnover might be one of the withdrawal behaviors such as absenteeism, and thus it is useful to include other withdrawal behaviors as well as other attitudinal variables (e.g., commitment) in the comprehensive model.

Overall, some researchers such as Price and Hom & Griffeth try to develop comprehensive models of voluntary turnover (e.g., using structural models) based on the Mobley's intermediate linkage model and other related perspectives (e.g., Hom & Griffith, 1995; Hom et al. 2000).

Cognitive, affective and behavioral antecedents

With regard to the various antecedents of employee voluntary turnover, individual attitudes are among the traditional, well-studied constructs. Attitudes such as job satisfaction and commitment includes both cognitive and affective components (e.g., Weiss & Cropanzano, 1996). Especially, Allen & Meyer (1990) show that the construct of organizational commitment includes affective, continuance and normative components. Research shows that attitudinal variables predict turnover although the validity coefficients are not so high (e.g., Hom & Griffith, 1995).

Affective states such as mood and emotion may also influence employee turnover and thus these variables can be the antecedents. However, George (1996) reports that the results of empirical studies on positive and negative affect and withdrawal behaviors are mixed. That is, positive affect was found to have some relationships with withdrawal behaviors but negative affect was not. There might be asymmetric effects of positive affect and negative affect, considering those two constructs are independent of each other rather than unidimensional.

With respect to the behavioral antecedents of employee turnover, job search behavior is possible antecedents of turnover, considering Mobley's model. However, it might be methodologically difficult to be empirically tested because people will not say they are looking for other jobs. Other withdrawal behaviors such as absenteeism can be behavioral antecedents of turnover. However, sometimes these behaviors are less discretionary and thus predictive validates of these behaviors may not be high (e.g., Chen et al. 1998).

Recently, Chen et al. (1998) found that organizational citizenship behaviors (OCBs) (e.g., Organ, 1988) are the valid predictors of employee turnover. Because OCBs are discretionary behaviors that are not recognized by the formal reward systems, the decrease of OCBs can be the signal of thought of quitting.

Lee & Mitchell's Unfolding Model of Employee Voluntary Turnover

Lee & Mitchell (1996) propose an alternative approach for employee turnover. The antecedents of this model might be the Steers and Mowday's (1988) model and Beach and Mitchell's image theory on decision making. I will first describe the development of Lee & Mitchell's model and then discuss the recent high turnover rates using their model.

Summary of Lee & Mitchell model

Steers and Mowday's model suggests that there are multiple processes included in voluntary turnover process, while traditional models assume only one intermediate link from job dissatisfaction to turnover. However, they did not specify the concrete processes. Beach & Mitchell's image theory is a more naturalistic model compared with other normative and behavioral decision making models. The theory suggests that screening is more important than choice and the people's decision process is based on not only the utility maximization, but also other mechanisms that include more automatic, habitual ones. Three images (i.e., value, trajectory, and strategic image) play an important role in their theory.

The main argument of Lee & Mitchell's model is that the decision path in the traditional approach such as Mobley's model is only one of the four alternative decision paths. People may think differently about turnover. Sometimes they quit very fast or rather automatically. Sometimes they are very careful about making a decision about turnover.

Lee & Mitchell's unfolding model consists of several key constructs. They are, shock to the system, scripts or schema, image violation, dissatisfaction and job search, and actual quitting. There are four different paths and three of them start with the shock or life event (that is not necessarily negative). In decision path 1, such shock make the focal person to look back their past experiences. If script match the turnover behavior, he or she will quit. In decision path 2, script match does not occur but if the shock violates his or her image (e.g., value, trajectory and strategic), he or she will quit. In decision path 3, image violation does not lead to quitting directly but instead he or she experience dissatisfaction, that leads to the job search and comparison of alternatives before actual turnover. In decision path 4, shock does not occur but people sometimes access past script and image violation, and similar process such as path 2 and 3 can occur. Lee & Mitchell's model have been empirically tested (e.g., Lee et al. 1996, 1999)

Explanation of high turnover rates using Lee & Mitchell's model

There may be some factors that affect high turnover rates. First, the frequency and significance of "shock to the system" may be increasing. In the rapidly developing industries such as high tech sectors, quite many events can occur. These events might be negative (e.g., job failure), positive (e.g., a phone call from a headhunter) or neutral. The quality or significance of these events may affect the different decision paths.

Second, people may accumulate scripts that are related to the immediate turnover. If job-hopping become popular, people may accumulate experience of changing jobs. These experiences will contribute of developing behavioral scripts (e.g., quit soon if certain event occurs).

Third, people's images may be changing. The value, trajectory, and strategic images can be affected by recent trends, contextual factors, and so on. Therefore, it is possible that peoples values, goals, and strategies gradually changing so that people weigh more values to the changing jobs than before.

Fourth, even if shock does not occur, people can be more frequently look back their past experiences or access fit between their images and organizations. Because organizations are also changing in the turbulent environments, misfit or image violation may sometimes occur, which leads to the decision path 4.

In sum, there may be a lot of factors that affect the decision process that leads to turnover. Lee & Mitchell's model is very useful for understanding these factors.

Influence of selection practices

Literature on personnel selection provides many insights about the employee voluntary turnover. Among the alternatives, realistic job preview (RJP) may be the effective practice in order to reduce turnover. As met expectation model suggests, if people's expectancy unmet the actual organizational reality, people are likely to quit soon. Therefore, RJP decreases the likelihood of turnover because it provides real view of the organization, which changes the expectation of job candidates or eliminates the candidates whose expectations will not meet with the reality.

Next, the concept of person-environmental fit, especially person-organization fit (P-O fit; e.g., Kristof, 1996), is important for understanding employee turnover. Traditional selection practice tends to be based on the Person-Job fit. That is to find a person who has KSAOs that are required to perform jobs. However, some researchers (e.g., Bowen, 1993) argue that companies should select people who are high in person-organization fit. Research shows that high person-organization fit decrease employee turnover (e.g., Chatman). But we should be careful about the understanding of fit because it is very complex. Kristof (1996) shows that there is perceived fit as well as actual fit, and that there is also complementary fit (e.g., fit between needs and supply) as well as supplementary fit (e.g., value congruence). In sum, the increase of fit leads to the positive attitudes (e.g., job satisfaction) toward organizations. Thus people high in P-O fit may be less likely quit.

Finally, socialization practices also play an important role in voluntary turnover. Socialization may occur before people enter the organization (i.e., in personnel selection process). Research shows that effective socialization leads to the person-organization fit and it decreases the probabilities of voluntary turnover (e.g., Chatman).

In sum, some selection practices such as RJP, person-organization fit, socialization practices play an important role in reducing voluntary turnover. Organizations should make the most of these selection practices in order to reduce turnover rates.

Future research

Voluntary turnover may be functional or dysfunctional (e.g., Hom & Griffeth, 1995). Therefore, it is not always true that voluntary turnover should be avoided. Future research is needed that provides rich insights about how organizations can reduce voluntary turnover of valued employees while promotes turnover of the least valued employees in order to increase organizational effectiveness.