Introduction

One of the constant challenges faced by managers is the development and retention of talent (Mitchell, Holtom, & Lee, 2001). Organizations intent on developing competitive advantage in virtually any economic or geographic context will rely on talented and dedicated employees (Pfeffer, 1994). Moreover, the personal and organizational costs of voluntary employee separation are high. Thus, employee retention merits the attention of top-level managers in today’s organizations.

Compounding the retention challenge is the fact that the composition of the U.S. work force is undergoing fundamental change. The number of Hispanics in the U.S. surged by 60% from 1990 to 2000 (U.S. Census Bureau, The Hispanic Population in the United States, 2001a). Hispanics are now the largest minority in the U.S. Given these changes, there is a clear need to consider how well organizational theories developed among relatively homogeneous U.S. populations predict important organizational outcomes in more diverse cultural contexts. This paper seeks to address one of these outcomes directly. The purpose of this paper is to develop and test the application of job embeddedness—a new theory of employee retention—in the Hispanic cultural context.

Voluntary turnover research

Voluntary turnover research has traditionally followed two main paths: (1) the study of turnover from the attitudinal perspective, considering job satisfaction and organizational commitment, and (2) the influence of the labor market on turnover through the perceived ease of movement and availability of alternative employment. The models developed along these paths, although successful in identifying turnover predictors, explain only a limited percentage of quits (Griffeth, Hom, & Gaertner, 2000). More recently, researchers considered instances in which turnover may occur in spite of high job satisfaction or labor market conditions (Lee, Mitchell, Holtom, McDaniel, & Hill, 1999) and may be moderated by variables other than those strictly related to the employee–job or employee–employer relationships (Griffeth et al., 2000).

Over the years, researchers have determined that given alternatives, people stay if they are satisfied with their job and committed to their organization and leave if they are not. However, the research reports that work attitudes play only a relatively small role overall in employee retention and leaving (Griffeth et al., 2000).

One possible explanation for the limited explanatory power of organizationally focused turnover theories is the emerging body of empirical research that suggests that many off-the-job factors are important for attachment (Maertz & Campion, 1998). The original turnover models of Price & Mueller (1981) and Mobley (1982) mention “non-work” influences and they include family attachments or conflicts between work and family roles. Research on spillover models explains how family and work lives are related (Marshall, Chadwick, & Marshall, 1992). Cohen (1995), for example, shows how non-work commitments like family, hobbies and church influence job attitudes and attachment. Lee and Maurer (1999), moreover, found that having children at home and a spouse were better predictors of leaving a job than organizational commitment.

Additionally, there are now a variety of factors that have been empirically associated with retention that are not attitudes but are organizational in nature. Inducements to stay can derive from working with groups or on certain projects that create types of commitment other than the attraction one has for his or her job or organization. For example, many companies use teams to induce attachments (Cohen & Bailey, 1997). A new theory of turnover—job embeddedness—explicitly includes both on- and off-the-job factors that influence staying.

Job embeddedness

Job embeddedness represents a broad array of influences on employee retention (Mitchell, Holtom, Lee, Sablynski, & Erez, 2001). The critical aspects of job embeddedness are the links an employee has to other people or the community, how he or she fits in the organization or environment and, lastly, what the employee would sacrifice upon leaving the organization. These three dimensions are called links, fit, and sacrifice; they are relevant in both the organization and the community (i.e., two sub-dimensions for each dimension). Job embeddedness is a composite construct formed from the six sub-dimensions that result from the 3 × 2 matrix suggested above.

Links are characterized as formal or informal connections between a person, and institutions or other people. Embeddedness suggests that a number of strands connect an employee and his or her family in a social, psychological, and financial web that includes work and non-work friends, groups, the community, and the physical environment in which he or she lives. The higher the number of links between the person and the web, the more an employee is bound to the job and the organization. We believe certain links may be more important than others and that these differences may be population specific as will be discussed later. A variety of research streams suggest that there is normative pressure to stay on a job, which derives from family, team members, and other colleagues (Maertz, Stevens, & Campion, 2003). O′Reilly, Caldwell, and Barnett (1989) use the term social integration to describe the at-work part of the link process. Furthermore, a study by Abelson (1987) assessed variables related to both on and off-the-job links. He found that people who are older, are married, have more tenure and/or have children requiring care are more likely to stay than to leave. Leaving their job and perhaps their home can sever or require the rearrangement of some of these links. In short, job embeddedness integrates these important social influences on employee retention.

Fit is defined as an employee’s perceived compatibility or comfort with an organization and with his or her environment. According to job embeddedness theory, an employee’s personal values, career goals and plans for the future must “fit” with the larger corporate culture and the demands of his or her immediate job (e.g., job knowledge, skills, and abilities). In addition, a person will consider how well he or she fits the community and surrounding environment. Job embeddedness posits that the better the fit, the stronger the ties to the organization. In studying voluntary turnover, for example, O′Reilly, Chatman, and Caldwell (1991) found that misfits with the organization’s values terminated slightly faster than fits. Chatman (1991) reported that when organizational entry produces poor person-organization fit, employees are likely to leave the organization. Chan (1996) suggests that having one’s personal attributes fit with one’s job may decrease turnover, and Villanova, Bernardin, Johnson, and Dahmus (1994) found that lack of job compatibility predicted turnover. Thus, a person’s fit with the job and organization relates to attachments to the organization. Mitchell et al. (2001) believe that there is a community dimension of fit as well. A person may or may not like the weather, amenities and general culture of the location in which he or she resides. In addition, outdoor activities (e.g., fishing, skiing), political and religious climates, and entertainment activities (college or professional sports, music, theater) vary dramatically by region and location. Importantly, these assessments of fit may be independent of job or organization fit (I like GM but I dislike Michigan winters). Relocation would obviously require a recalibration of fit, but even a new job without relocation could disturb one’s general patterns with new hours of work or a different commute.

Sacrifice captures the perceived cost of material or psychological benefits that may be forfeited by leaving one’s job. For example, leaving an organization likely promises personal losses (e.g., giving up colleagues, interesting projects or desirable perks). The more an employee gives up when leaving, the more difficult it is to sever employment with the organization (Shaw, Delery, Jenkins, & Gupta, 1998). Non-portable benefits like stock options or defined benefit pensions may involve sacrifices. These latter factors have been shown to be related to turnover (Gupta & Jenkins, 1980). Other potential sacrifices incurred by leaving an organization include opportunities for job stability and advancement (Shaw et al., 1998). Community sacrifices (as well as links and fit to some extent) are mostly an issue if one has to relocate. Leaving a community that is attractive and safe or where you are close to your family can be hard. Of course, one can change jobs but stay in the same home. But even then, various conveniences like an easy commute or the ability to be at home during certain times due to flextime (e.g., when kids come home from school) may be lost by changing jobs.

With a sample of retail employees and another sample of hospital employees, Mitchell et al. (2001) reported the following. First, job embeddedness was reliably measured as an aggregated score across items for fit in the organization, fit in the community, links to the organization, links to the community, sacrifice in leaving the organization, and sacrifice in leaving the community. Second, aggregated job embeddedness was correlated with intention to leave and predicted subsequent voluntary turnover. Third, job embeddedness significantly predicted subsequent voluntary turnover after controlling for gender, job satisfaction, organizational commitment, job search, and perceived alternatives. A subsequent study (Lee, Mitchell, Sablynski, Burton, & Holtom, 2004) extended the theory and research on job embeddedness. One contribution was to disaggregate job embeddedness into its two major sub-dimensions, namely, on- and off-the-job embeddedness. Using a large sample of bank employees, regression analyses revealed that off-the-job embeddedness was significantly predictive of subsequent voluntary employee turnover and volitional absences, whereas on-the-job embeddedness was non-significant. However, to date no research has been published that examines the predictive value of job embeddedness in a culturally diverse population. The purpose of this paper is to address this question.

Hispanic culture

The term “Hispanic” is used to refer to those people who originate from Mexico, Puerto Rico, Cuba, or Central or South America. The incorporation of different Hispanic groups in the American society has been different from the experience of earlier immigrants from European countries. While there are considerable differences in the geographic, economic, and historical contexts of these countries, the people from these countries frequently self-identify as Hispanic—at least one indication of shared cultural values—which, contrary to the case of European immigrants, stems from a shared language (Melendez, Rodriguez, & Figueroa, 1991). For example, on the 2000 U.S. Census a respondent could indicate that he or she was Hispanic either because of foreign birth or heritage (though born in the U.S.). Diverse national origin is but one of the differences among Hispanics. Research has shown differences depending on the time of arrival to the United States, number of generations in the country and socio-economic status (Valenzuela & Dornbusch, 1994). So, while many important differences between people characterized as Hispanic may exist, we believe that there are also substantial similarities and, thus, it is appropriate to follow the U.S. Census categorization for purposes of this research.

Given the relatively recent emergence of Hispanics as a dominant minority, very little research is available that examines systematic differences between Hispanics and other races with respect to organization relevant attitudes (e.g., job satisfaction, organizational commitment) and behaviors (e.g., performance, organizational citizenship behaviors, absenteeism, turnover). However, we believe that a number of differences between the cultures may suggest the potential for higher scores among Hispanics than North Americans.

Given the dearth of management research on the cultural variables that deal specifically with Hispanics in the U.S., collectivism (Hofstede, 1996) may constitute an acceptable proxy and may suggest indirect evidence for the value of job embeddedness. Because national cultures are the categories for comparison in Hofstede’s research, the degree of collectivism recorded in Latin American countries from which Hispanics come to the U.S. provides the basis for an estimate. Latin American cultures cover most of the lower half of Hofstede’s 0–100 scale for the individualism-collectivism dimension, ranging from a low of six (Guatemala) to a high of 46 (Argentina). Inspite of this ‘spread’, all the Latin American countries reported by Hofstede (1996) can be defined as having relatively collectivistic cultures (i.e., their scores are below 50). Moreover, because the theory emphasizes the importance of comparisons, the cultural distance between those countries (average score of 20) and the U.S. score (91) becomes particularly relevant. Because job embeddedness places significant importance on connections to groups, teams families and communities, we believe it will be higher in the Hispanic than majority member cultures.

While individualistic societies tend to base much of their social structure on the concept of the nuclear family, collectivistic cultures often favor extended family arrangements that condition members to perceive themselves as members of a “we” that transcends the “I” (Hofstede, 1996). Since collectivism is inversely correlated with wealth, the extended family doubles as an economic structure to which members often have a financial obligation.

Collectivist family arrangements are characteristic of the Latino population and while the body of research is biased toward low-income families of Mexican ancestry (Velez-Ibañez, 1996) suggesting kinship networks predominantly based on economic needs, more recent studies indicated kinship networks and extended families are not rooted only in economic reasons (Valenzuela & Dornbusch, 1994). Moreover, studies have shown that extended families become larger and stronger with generational advancement, and acculturation (Velez-Ibanez, 1996). Not surprisingly, family ties are considered to exert a strong influence on Hispanic decision making with regard to relocation (National Survey of Latinos, 2002). Anecdotal evidence in the popular press suggests that because of close familial ties, many Hispanic employees decline relocations that would provide upward mobility (Joinson, 2000).

One of the significant challenges faced by the maquiladora industries located along the U.S.–Mexico border is the exceptionally high turnover among workers, which can run higher than 100% annually (Crispin, 1990). In a study of maquiladoras, Maertz et al. (2003) found that when employees were asked to discuss to what or whom they are loyal, more than 50% of the references were to family, friends, and neighbors. In contrast, less than 2% mentioned their employer. The degree to which workers in the maquilas express loyalty to family and friends suggests the potential importance of the “links in the community” dimension of job embeddedness.

Regarding trends among Hispanics living in the U.S., there is evidence from the 2000 U.S. Census (U.S. Census Bureau, 2001b) to support the notion that Hispanics are less likely to permanently relocate long distances. While the overall rate of moving is highest among Hispanics (over White, Black, and Asian Americans), moves outside the county or state are lowest among Hispanics. Moreover, the same report indicates that overall long-distance moves are more likely to be made for work-related reasons, while short distance moves are more likely to be made for housing-related reasons (U.S. Census Bureau, Geographic Mobility, 2001b). The combination of these data seems to confirm the evidence offered earlier regarding potential hesitancy among Hispanics to move away from family for work-related reasons.

U.S. based research has also considered off-the-job factors in job mobility. A meta-analysis of demographic predictors (Griffeth et al., 2000) shows number of children in the household to be negatively correlated (r = −.14) with voluntary turnover. Research by Lee and Maurer (1999) also speaks to the importance of family size and household structure as influences in reducing turnover. According to the United States Census Bureau, Hispanic families have a higher than average number of children living in the household. While the national average is 1.85 children under the age of 18, families of Hispanic heritage report 2.19. Moreover, 30.6% of family households in which a Hispanic person was the householder consisted of five or more people. In contrast, only 11.8% of non-Hispanic White family households were this large (U.S. Census Bureau, The Hispanic Population in the United States, 2001a). By the year 2020 white non-Hispanic people will represent 67% of the work force, down from the current 76%, and the Hispanic presence is predicted to increase by almost 60% nationwide, and will be more accentuated in certain regions of the country.

With skills learned in the home country devalued in new labor markets, immigrants with a generally poor command of the new country’s language often seek residence in areas where significant numbers of similar immigrants live. Their economic destinies depend heavily on the social structures present in these communities and, thus, their mobility is likely to be limited (Portes & Sensenbrenner, 1993).

In sum, the above research suggests the collective importance of family and community influences on Hispanic job mobility. Further, many of the items used to measure the community aspects of job embeddedness would likely elicit higher scores from Hispanics than non-Hispanics. The foregoing discussion leads to two hypotheses:

Hypothesis 1A

The fit-community, links-community and sacrifice-community dimensions of job embeddedness will be higher among Hispanics than Caucasians in comparable jobs.

Hypothesis 1B

The overall level of job embeddedness in the community will be higher among Hispanics than Caucasians in comparable jobs.

To date, very little research exists that examines the impact of a specific cultural heritage on individual-level outcomes of interest to organizations such as turnover, absenteeism or performance. We believe that job embeddedness may help to connect culture—in this case Hispanic culture—to demographic influences in explaining additional variance in turnover. This paper replicates and extends the research by Mitchell et al. (2001) and Lee et al. (2004) by inquiring not only about the value of job embeddedness as a predictor of voluntary turnover, but also by considering its strength in a diverse population with a strong participation of employees of Hispanic heritage. Accordingly, we propose the following hypotheses:

Hypothesis 2A

After controlling for job satisfaction and organizational commitment, job embeddedness is significantly and negatively correlated to voluntary turnover of all employees regardless of race (Replication).

Hypothesis 2B

After controlling for job satisfaction and organizational commitment, job embeddedness is significantly and negatively correlated to voluntary turnover of Hispanic employees (Extension).

Given the relatively recent rise of Hispanic participation in the U.S. labor force, little literature is available to comprehensively identify the array of factors causing Hispanics to develop stronger attachment to the firms that employ them or to other forms of employment (e.g., self-employment). One research topic that has received attention recently is Hispanic entrepreneurship. Some of these studies suggest that working within a ‘Latino’ atmosphere (e.g., a firm owned or managed by Hispanics) may strengthen the embeddedness of Hispanic employees. For example, “Most U.S. employers are ill-prepared to evaluate foreign-earned human capital and thus its value is discounted. By contrast, co-ethnic employers may recognize the value of human capital earned in their home country” (Sanders & Nee, 1996).

Research conducted on entrepreneurship suggests that on-the-job and off-the-job embeddedness may have a stronger effect for those Hispanics living in “enclaves.” There is reason to believe that the same set of factors that explain sociological embeddedness in a community also explain job embeddedness for Hispanics. Aside from the obvious example of family-owned companies that tend to blur the line between community and organization, larger groups of co-ethnic but non-related immigrants also provide a base and a support structure for entrepreneurs. The latter would thrive in environments defined by immigrant communities whose size increases while they remain linguistically isolated (Evans, 1989).

Recent work by Maertz et al. (2003) develops a turnover model for the Mexican maquiladoras. A number of findings from their research inform this study. Specifically, they found that Mexican workers evaluated their jobs in terms of flexibility to accommodate non-work interests (e.g., family and education). Another was the perception that leaving the organization would bring about costs in seniority benefits. These underlying values are clearly consistent with the “sacrifice upon leaving the organization” dimension of job embeddedness. Moreover, these workers indicate a desire to remain in an organization because of their relationships with co-workers, which suggests the value placed on “links in the organization” (Maertz et al., 2003).

Finally, Triandis’ (1989) work implies that high collectivism should enhance the level or importance of organizational commitment where the organization is seen as an in-group. Given that Mitchell et al. (2001) and Lee et al. (2004) found high correlations between organizational commitment and “fit in the organization,” we believe that if the organization is viewed as an in-group, the generally collectivist Hispanics will exhibit higher “fit in the organization.” Thus, we propose the following hypotheses:

Hypothesis 3A

The fit-organization, links-organization and sacrifice-organization dimensions of job embeddedness will be higher among Hispanics in a predominantly Hispanic firm than in a non-Hispanic firm.

Hypothesis 3B

The overall level of job embeddedness in the organization will be higher among Hispanics in a Hispanic firm than in a non-Hispanic firm.

Methods

Overview and samples

The study was conducted at two leading southeast Florida banks. Both financial institutions are of recognized national presence and are active in the same market segment (e.g., retail banking services). In one of the banks, all of its Broward County branch offices were targeted for distribution of the survey (hereinafter BC Bank). At the second bank, all branches located in Miami-Dade County were targeted (hereinafter MD Bank). According to the United States Census Bureau, the Miami-Hialeah metropolitan area ranks third in the nation in the concentration of Hispanic households, with predominant roots in Cuba and in other Latin-American nations. Hispanics represent 13.5% of the U.S. population, while in Miami-Dade and Broward counties they represent 60.1 and 19.5%, respectively. In terms of African–American population, Miami-Dade and Broward counties are also above the national average, with 21.1 and 23.6%, respectively, compared to 12.8% nationwide (U.S. Census Bureau, American Community Survey Profile, 2002).

In the case of BC Bank, the survey instruments were sent to each branch office in sealed envelopes via inter-office mail. Branch managers were aware of the distribution via company e-mail from senior management, which also requested their cooperation with the distribution of the instruments to all branch employees. The distribution took place during May 2002. In the case of MD Bank, the survey instruments were provided in packages to branch managers for distribution to employees, at a gathering of senior personnel for the Miami-Dade County region in August 2002. It is important to note that MD Bank was previously owned by Hispanics and still has a high concentration of Hispanic employees at all levels of the organization, including top management.

In both cases, the full cooperation of the organization was indicated to the employees with a letter of introduction from management. A postage-prepaid, self-addressed envelope was included. Confidentially of responses was guaranteed by a letter on university letterhead. In order to give respondents additional comfort with the level of confidentiality of the survey, reply envelopes were addressed to the university. Management sent a reminder via e-mail to all participating employees 10 days after the distribution of the survey.

Questionnaires were distributed to 555 employees of MD Bank and to 670 employees of BC Bank. Employees returned 189 usable questionnaires from MD bank and 233 from BC bank, representing 34.1 and 34.8% response rates, respectively. Although the institutions were in agreement with providing turnover information in the future, they requested the survey be anonymous, unless the employees agreed to self-identification. Self-identification was encouraged by offering participation in a draw with a $500 cash prize to those employees that chose to provide their names. Of the 423 valid responses, 216 of the respondents provided their names and in most cases, telephone numbers. Of the 216 identified respondents, only 180 resulted in a match with bank records, mostly due to hardly legible handwriting that could not be matched to bank records. One year after the initial distribution, both banks provided actual turnover information for the list of self-identified employees, indicating date of termination of employment and whether such termination was voluntary or involuntary.

Measures

The methodology follows the model used by Mitchell et al. (2001) in the initial study of job embeddedness. The survey instrument used by Mitchell was distributed to employees of the two financial institutions. All items are provided in Appendix. The question regarding ethnicity allows for five possible answers, following the standard used by Miami-Dade County for ethnic classification.

Most questions required answers in a five-point Likert scale (strongly disagree = 1, strongly agree = 5). Following Mitchell et al. (2001), we created Z-scores for the fill-in questions (such as tenure in the position, organization and type of work or number of children) before including them in the calculation of the aggregate variables. The reliability (Cronbach’s alpha) of the 34 items measuring job embeddedness construct is measured at .89. Similar to the results found by Mitchell et al. (2001), including low scores for Sacrifice-Community (.66) and Links-Organization (.64), the reliability estimates (Cronbach’s alpha) for all other variables were higher than the recommended .70 (Hair, Anderson, Tatham, & Black, 1998).

Results

Demographic characteristics of the respondents are presented in Table 1. Correlations between the variables are presented in Table 2; Caucasian results are presented in the lower diagonal and Hispanic results in the upper diagonal. Descriptive statistics, reliability estimates, and t-tests for other variables are presented in Table 3. To test for selection bias a Chi-square test was performed for gender (information on tenure, age and marital status was not made available by the organizations). The majority of respondents are women, consistent with the overall composition of the banks where 70–80% of employees are female. There was no evidence of gender bias among respondents (χ2 = 3.49, n.s.). Further comparisons of respondents who did not self-identify with the 180 employees who did self-identify show no significant difference between the two groups (on gender, tenure, age, or marital status).

Table 1 Demographic characteristics of respondents
Full size table
Table 2 Correlations (lower diagonal: Caucasian; upper diagonal: Hispanic)
Full size table
Table 3 Reliability, means, standard deviations, and t-tests
Full size table

As can be seen from Table 3, as predicted by Hypothesis 1A, Hispanics demonstrate higher levels of fit-community (t = 1.9, p < .03) and sacrifice-community (t = 3.7, p < .001) than Caucasians. However, contrary to Hypothesis 1A, Hispanics demonstrated a lower level of links-community (t = −2.0, p < .03) than Caucasians. Overall job embeddedness in the community was higher among Hispanics than among Caucasians (t = 1.7, p < .05) as predicted by Hypothesis 1B. Thus, H1A is partially supported and H1B is fully supported. In general, we observe support for one key idea: Hispanics tend to be more embedded in their communities than Caucasians as measured by job embeddedness.

As can be observed in Table 4, after controlling for the institution in which the respondents work, job satisfaction, and organizational commitment, job embeddedness in the organization is a statistically significant predictor of voluntary turnover among respondents of all races (Hypothesis 2A). Further, in accord with Hypothesis 2B, after controlling for the institution in which the respondents work, job satisfaction, and organizational commitment, job embeddedness in the organization is a statistically significant predictor of voluntary turnover among Hispanics (Table 4).

Table 4 The effects of job embeddedness on voluntary turnover
Full size table

Finally, from the t-tests presented in Table 3 it is also apparent that Hispanics in predominantly Hispanic firms do not differ with respect to job embeddedness in the organization from Hispanics in non-Hispanic firms. This is contrary to Hypothesis 3A and 3B.

Discussion

This study highlights the importance of job embeddedness, not only as a predictor of turnover, but also as a guide to managers on retention policies. It addresses not only organization-related issues, but also extends research into employee attachment to the community. In addition, this research provides information on job embeddedness not previously available regarding specific characteristics of Hispanic employees. With the increased participation of Hispanics in the labor force, understanding this ethnic group gains importance.

The hypotheses presented here posit that due to cultural influences Hispanics would show a higher degree of embeddedness in the community than Caucasians. Analysis of the responses of over 400 employees in two institutions performing substantially similar jobs supports the hypotheses.

Interestingly, job embeddedness in the organization was not stronger among Hispanics in Hispanic firms than in non-Hispanic firms. This finding seems to concur with the Maertz et al. (2003) study that “found no consistent evidence that maquiladora workers see the organization as an in-group.” Further, contrary to the Lee et al. (2004) findings, job embeddedness in the community did not predict subsequent voluntary turnover, though job embeddedness in the organization did. This was true for the aggregate analysis as well as the Hispanic only analysis.

There are a number of practical implications of these findings. First, organizations that require long-distance relocation for advancement may face resistance from Hispanic employees or possibly experience higher turnover in this population. Second, employers may want to consider a broad definition of family with respect to social and other activities so as to further connect employees’ family to the “organizational web.” Third, efforts to assist in integrating on- and off-the-job activities (e.g., promoting involvement in local schools by providing paid time off to do so) may provide retention benefits as well as community benefits.

Overall the findings suggest the need for new ways of thinking about the Hispanic population in this country. “It is neither monolithic nor a hodgepodge of distinct national origin groups. Rather, Latinos share a range of attitudes and experiences that set them apart from the non-Hispanic population” (National Study of Latinos, 2002). While we advance no hypotheses with regard to job satisfaction, organizational commitment, or intent to leave, we note that all three are statistically significantly higher among Hispanics than among Caucasians (Table 3). Though the results may be sample or industry specific, we believe that these findings are interesting and merit future research attention. Specifically, we note that according to the dominant turnover paradigm (Hom & Griffeth, 1995) higher levels of job satisfaction and organizational commitment in a population would lead us to predict a lower level of intent to leave. In this case it is exactly the opposite (intent to leave is also higher among Hispanics). However, actual voluntary turnover among Hispanics in the self-identified sample was 7.4% compared with 14.7% for Caucasians and 16.7% for African–Americans.

Several questions remain, particularly concerning future projections of the trends outlined so far. Given how successful the U.S. has been at assimilating and socializing immigrants, it would be reasonable to argue that Hispanics will eventually share the cultural features of their new home. If this translates into a shift from collectivism to individualism on the part of future generations of U.S. Latinos, the particularities of labor relations involving Hispanics will be transitional or short-lived.

The negative correlation between wealth and collectivism seems to support this position. Census and economic data indicate that income among Latinos is growing. This would welcome the formation of nuclear families in detriment of extended ones, which constitute one of the cornerstones of collectivism as a cultural feature (Hofstede, 1996).

However, there is also evidence that supports the opposite position and leads us to believe Hispanic culture, including collectivism, is here to stay. Hofstede detected a strong positive correlation between collectivism and number of children per family, and the aforementioned census data (U.S. Census Bureau, The Hispanic Population in the United States, 2001a) shows that Hispanic couples have 18% more children than overall non Hispanic couples in the U.S., and as mentioned before, studies have shown that extended families become larger and stronger with generational advancement, and acculturation (Velez-Ibanez, 1996). Future studies of workplace attachment clearly must measure and model collectivism.

Limitations of the study

The population studied is predominantly female. While this is a characteristic of the employee body of the host organizations, the results obtained should be taken with caution. Previous research (Griffeth et al., 2000), points to gender differences in turnover responses. Further, though the demographic characteristics of the respondents who provided self-identification on their surveys are not demonstratively different from those who did not, the logistic regression results may be subject to selection bias. However, despite low power, significant results were obtained, suggesting a potentially robust phenomenon.

Analysis of U.S. census information shows a difference in the characteristics of Hispanics in southeast Florida, when compared to national averages. Southeast Florida’s Hispanic population is above the national averages for Hispanics in terms of education and income and is predominantly of Cuban descent—not Mexican or Puerto Rican descent as it is predominant in other regions of the country. The high rate of entrepreneurship and concentration in the Miami area has resulted in a prosperous enclave of privileged migration (Portes & Truelove, 1987).

This study also has limitations to generalizability common to most geographically constrained studies. While the focus of this research is on the applicability of job embeddedness as a predictor of employee retention among Hispanics in the U.S., to best understand this phenomenon, multi-factor/multi-method research across multiple regions of the U.S. must be undertaken.

Recommendations for future studies

The demographic limitations of the population studied present an opportunity for future research that may contribute not only to further validate the job embeddedness construct, but also to provide a better understating of the response of Hispanics of different socio-economic strata. In addition, future studies should include a more balanced participation of males.

The concept of job embeddedness includes a number of variables intended to measure the employee’s attachment (link) to and fit in the community and organization and the potential “cost” (sacrifice) of leaving such organization or community. Research has already established the significant and negative relationship between children in the household and voluntary turnover (Griffeth et al., 2000). However, the present variables included it the job embeddedness concept do not include such consideration. The results of this study demonstrate a significant difference in job embeddedness in the community between Hispanic and non-Hispanic employees. While it can be argued that a number of cultural and social reasons make employees of Hispanic descent more embedded in their communities, it is also a fact that the family structure of Hispanics includes a higher than average number of children in the household. Number of children in the household was not previously considered as part of the job embeddedness construct (Mitchell et al., 2001). Including items that address the impact of family size in the job embeddedness concept seems appropriate.

In sum, it appears that while Hispanics exhibit different levels of job embeddedness from Caucasian workers, the job embeddedness model is a reliable predictor of employee retention. Because of its focus on the community as well as the organization, job embeddedness may be a robust model for future investigation of attachment among Hispanics—the largest U.S. minority.