1. Introduction

How to attract, develop and retain employees is an important issue for organizations. At the same time, employee mobility is a key element of organizational strategy. Organizations seek to minimize the tangible and intangible costs associated with employee mobility and retain employees. The reason of employee turnover is an important topic for scholars to study at present.

The concept of job embeddedness has existed for a relatively short time, based on the fact that existing theories were not sufficient to explain the reasons for turnover.

Job embeddedness is based on three studies. First, job-related factors do not explain and predict turnover well, and many nonwork factors have a great impact on employee turnover; second, according to previous experience, the triggers for employees to choose to stay come from the organization rather than the individual. Third, employees may leave because of some unexpected event, so attitude and whether they looked for another job in advance cannot accurately predict turnover.

This study uses literature analysis method to review the research on job embeddedness in the past 22 years. First of all, a profound description of the concept of job embeddedness will be given, especially components of it and core constructs. Secondly, this study reviews the two current methods of measuring job embeddedness and their differences. Then, by analyzing the past literature, this study summarizes the general research direction on job embeddedness, including the research on turnover, job performance, and gender. Finally, the shortcomings of past research and the expectation of future research are put forward.

2. Concept of Job Embeddedness

The concept of job embeddness was put forward byMitchell et al. in 2001, the core of which is embedded figures and field theory [1]. “Embedded” means that people and their surroundings are a whole and cannot be separated. Field theory means that everyone has their own perception of the living space, in the space of various things have a certain relationship. Based on this, Mitchell et al. graphically describe work as embedded in a web in which people may be trapped [1].

The composition of job embeddedness is varied, which has three dimensions from two scenarios [1]. Firstly, job embeddedness includes off-the-job embeddedness as well as on-the-job embeddedness, based on individual’s connection to the community and work. Each of these two scenarios has three dimensions, which are “links”, “fit”, and “prescription”.

“Links” refer to identifiable connections between individuals and community or organization, and making changes to existing connections may affect turnover intention. “Fit” means the extent to which people fit into the community and organization, and the perceived fit can affect employees’ willingness to resign. “Sacrifice” is the perceived cost of turnover on employees, including both material and psychological costs, which affect employees’ decision.

Through independent survey experiments, it has been shown that job embeddedness can help forecast turnover caused by job satisfaction together with organizational commitment and other traditional voluntary turnover reasons. It is also a more effective and innovative measure. Mitchell et al.’s empirical research also confirm that job embeddedness supplements previous research on the reasons for employee turnover [1].

3. Measurement of Job Embeddedness

There are currently two main methods for measuring job embeddedness, the composite measure and the global measure. The composite measure of job embeddedness was created by Mitchell et al. in 2001 and was the dominant approach in the early years of research [1]. Researchers generated a Likert Five-Point Scale with 40 items based on six dimensions of links, fit, and sentence in community and organization scenarios, as well as improvements from experimental feedback. The scale has also been proven to have good measurement and prediction abilities through experiments.

However, the composite measure of job embeddedness also has certain limitations. Firstly, due to theoretical limitations, the measurement ignores the different tendencies of individuals. Secondly, some of the items may infringe on personal privacy, further leading to problems about the authenticity of the participants’ answers. In addition, there are also measurement errors caused by excessive measurement length and some items that have not been recognized by the public.

Afterwards, Crossley et al. developed a new scale based on the precious measure——the global measure of job embeddedness [2]. Unlike previous scales, the global measure balances various subjective aspects of the respondents while taking into account individual tendencies. In terms of items, the potential invasiveness of the questions is avoided, and all issues are recognized by the public. With regard to statistics, the composite measure presumes that each dimension is equally significant to each individual. While global measure sets different subjective weights to highlight the value of different dimensions.

The researchers used Hinkin’s productive item-generation strategy, they obtained the completed and ongoing research projects from authors who were known to be studying the job embeddedness, and conducted checks to obtain a large number of clear examples. At the same time, they constructed the definition of items while ensuring the limited quantiy of items, to cover all content as much as possible. Afterwards, seven original projects were generated through modifications based on the opinions of colleagues. Finally they conducted statistical analysis of the experimental results. The measurement reliability is 0.88.

Thus, the current global measure has been formed. The results of empirical research indicate that this measurement method can predict employees’ intentions to search as well as seasonal turnover, and it also enriches the traditional turnover model.

In the actual use of measuring tools, whether it is a composite or global measure of job embeddedness, the selection needs to be determined based on the experimental background. If researchers want to reduce concerns of concept–concept impact in self-report, they can choose composite measure. When individual tendencies need to be considered, they can choose the global measure.

4. Practical Application of Job Embeddedness

4.1. Job Embeddedness and Turnover

Embeddedness itself is put forward due to the lack of explanation of turnover, so the research on job embeddedness along with turnover is a major area in this field. Even in the case of different variables participating in the regulation, a large amount of experiments have confirmed that when compared with low job embeddedness, high job embeddedness is significantly correlated with low turnover intention. In the meantiome, it’s well known that job embeddedness has two sides, although employees with high job embeddedness have lower turnover intention, which enables organizations to retain employees and reduce the cost caused by staff turnover. However, for employees, they may passively stay in their positions because of job embeddedness, which will have certain negative effects on their mental and physical health. Another significant finding is that job embeddedness can help predict employee exit behavior, taking into account more comprehensive factors than traditional voluntary turnover models.

Proposer of job embeddedness, Mitchell et al.'s initial experiment demonstrated that people with high job embeddedness have lower turnover intention, along with perceived job alternatives, job search behaviors, job satisfaction, organizational behavior, and job embeddedness all contribute to predicting turnover [1].

Subsequent studies also prove this point. Even when abusive supervision causes negative effects on employees’ physical and mental health or they are exposed to adverse work environments, employees who are deeply embedded have lower turnover rate than those with low job embeddedness. Because they are more loyal to their organization, which means it is hard to leave [3].

On this basis, there is a significant difference in prediction direction between on-the-job and off-the-job embeddedness, they have obviously disparate prediction direction. Off-the-job embeddedness can predict turnovers and omissions, but it cannot involve whether organizational citizenship or job performance, while on-the-job embeddedness is completely the opposite. However, on-the-job embeddedness moderates the effects of volitional absences along with job performent on voluntary turnover [4]. On-the-job embeddedness moderates the effects of volitional absences along with job performent on voluntary turnover.

Job embeddedness, especially on-the-job embeddedness, can make moderating effect on the positive relationship between voluntary turnover and work efficiency. The link between work efficiency and turnover is modest when high job embeddedness, and when at low levels, the opposite is true [5].

When researchers use the Social comparison theory to explore the reasons for turnover, an interesting finding is that colleagues’ job embeddedness also affects individuals’ turnover intention. When colleagues’ job embeddedness increase, individuals’ turnover intention decrease [6].

When it comes to the reasons for turnover in the psychological field, a recent study in China conducted a questionnaire survey on nearly 400 Chinese intellectuals and found that knowledge workers’ psychological contract violations have a significant direct impact on turnover intention, while job embeddedness moderating the effect on turnover intention which caused by psychological contract violations [7].

4.2. Job Embeddedness and Job Performance

Job embeddedness is crucial for organizations as it not only plays a significant role in turnover, but also has a unique impact on employee performance. In general, there is a positive correlation between job embeddedness and job performance. When job embeddedness plays a role of moderating or mediating variable, the impact on job performance is multifaceted.

In an experiment, the researcher conducted 839 effective questionnaires to five independent organizations in 1999, and asked the superior to rate the performance of subordinates. Through statistical analysis, after controlling job satisfaction and other two factors, job performance were connected with on-the-job embeddedness in a positive way [4].

The direct impact of job embeddedness on employee performance is relatively weak, but it has an important moderating effect on leader-member exchange together with organizational-based dignity quality on employee performance. When there is high quality of leader-member exchange is, high job embeddedness has a positive impact on employee performance, such as providing additional resources for employees. When the leader-member exchange is low, high job embeddedness has the opposite effect on employee performance [8].

Engagement is often associated with job performance, and previous researchers have not formally distinguished between engagement and job embeddedness. Through surveys of employees in different industries in the United States, as well as their supervisors and colleagues, Jonathan et al. found that engagement and job embeddedness are two independent structures. This has a unique role for further in-depth research on the prediction of employee turnover intention by both [9].

When companies use “forced ranking” system to manage employees, employee insecurity is greatly increased. In a recent study of employees at two large manufacturing companies in China, nearly 60 percent respondents were male and more than 59 percent had college degrees. The researchers used Richter five-component scale to measure and analyze job embeddedness, job performance, affective commitment and job insecurity. The results suggested that job insecurity has negative correlation with job performance when employees are at low job embeddedness, at the same time, job insecurity has negative correlation with job performance when employees are at high job embeddedness. Besides, affective commitment mediates the interaction of job embeddedness and job insecurity. As an emotional response to work experience, affective commitment is a very important mediator as shown in Figure 1[10].

Figure 1:Theoretical model of job embeddedness as a moderating variable and a mediating variable.

4.3. Job Embeddedness and Gender

According to the theory of gender roles, women are more entitled to “take care of family” and “weak” roles, while men are more entitled to “career-oriented” and “strong” roles. Therefore, in the society, there are many effects of gender stereotypes on employees of different genders. Even at high levels of embeddedness, women may be more vulnerable to negative influences from family or corporate culture in the workplace.

The two priorities of personal life are family and work, but the expectations of the two roles are not always compatible. When a person’s limited time and energy cannot meet the expectations of the two roles, work-family conflict (WFC) occurs. Job embeddness can regulate the positive relationship of WFC with emotional exhaustion. The higher degree of job embeddedness is at, the more obvious correlation between work efficiency and emotional exhaustion. For working mothers, while high job embeddedness means they are less likely to leave their jobs, they suffer from higher emotional exhaustion, guilt, and hostility. Which is very psychologically damaging for women who need to meet their family and work roles [5].

On the other hand, women's promotion opportunities in career are limited and less than men’s. In the process of working and promotion, women often encounter hidden gender obstacles such as “glass ceiling”. As mentioned above, job embeddness include “links”, “fit”, and “sacrifice”. Research shows that men and women have significantly different perceived organizational things in the past, with men tend to perceive more sacrifice than women. In terms of organizational links, men also have more links than women in group participation and years of service [11]. This may be due to the fact that organizational culture is more tolerant of men than women, so even women with high job embeddedness are less likely than men to get promoted.

5. Conclusion

The proposal of job embeddedness is a very good supplement to the reasons for leaving. Besides, it is not limited to the consideration of organization or community alone. Both the composite measure and the global measure have milestone significance for job embeddedness study. The two methods realize the concretization of job embeddedness from different angles. According to previous studies, on the one hand, job embeddedness has high correlation with turnover and job performance, and can help predict employee turnover or improve job performance. In terms of job embeddedness and gender research, scholars have paid proper attention to the unfair treatment of women due to cultural and social roles, which is more obvious at high embeddedness.

In the past, researchers on job embeddedness paid more attention to on-the-job embeddness. Employees’ job embeddedness can be easily measured, but the relationship between individuals and community has not gotten the same level of attention. In fact, off-the-job embeddedness has a great impact on the efficiency of employees while working, and even the retention. At the same time, the weights assigned to “links”, “fit” and “sacrifice” are also disparate because of different ages and positions in social networks. In reality, it is not uncommon for employees to stay because they have to take care of the elderly or children at home.

In terms of data analysis, the common limitation of many experiments is the data of survey is cross-sectional, which means data often come from a single point in time, so specific factors such as the background and work of the respondents cannot be taken into account, which restricts the interpretation of experimental results. Future research can use longitudinal studies to improve the rigor of the research.

For the research on the measurement of job embeddedness, although different weights are set in the global measure, such measurement method is not frequently used in subsequent studies, because there are only 7 questions in it and the questionnaire pays more attention to the on-the-job embeddedness mearing. With more and more in-depth studies on job embeddedness, future researchers should consider developing a tool that takes individual differences into account, as well as attaching equal importance to community and organization scenarios, so as to improve measurement accuracy.

Employee turnover and job performance have always been the main research directions of job embeddedness, but there is still a large gap in the impact of job embeddedness on individual life and family. Future research may attempt to study from the perspective of individual employees rather than the organization.

This study summarizes the development of job embeddedness by reviewing the past literature. On the one hand, it helps researchers find the problems and shortcomings in the field, so as to provide a basis for further research. On the other hand, people who don't know about job embeddedness can get a quick and comprehensive understanding of it with this study.