Gender differences in the family-relatedness of relocation decisions

Journal of Vocational Behavior 90 (2015) 1–12

Contents lists available at ScienceDirect

Journal of Vocational Behavior journal homepage: www.elsevier.com/locate/jvb

Gender differences in the family-relatedness of relocation decisions Jan Ullrich a,⁎, Helen Pluut b, Marion Büttgen a a b

Department for Corporate Management, University of Hohenheim, Schloss Osthof/Ost, 70593 Stuttgart, Germany Department of Organisation Studies, Tilburg University, Warandelaan 2, 5000 LE Tilburg, The Netherlands

a r t i c l e

i n f o

Article history: Received 24 April 2015 Received in revised form 15 June 2015 Accepted 16 June 2015 Available online 20 June 2015 Keywords: Relocation Career Family-relatedness of work decisions Spouse Gender

a b s t r a c t Using a dyadic study design, the present study draws on research into the family-relatedness of work decisions to examine the impact of the spouse's characteristics on an employee's willingness to engage in job-related relocation. With a sample of 1234 employees, the results show that spousal willingness to relocate mediates (a) the negative effect of the spouse's community ties and (b) the positive effect of the spouse's job alternatives on an employee's willingness to relocate. The indirect effect of the spouse's job alternatives was stronger among female than among male employees. To validate employees' perceptions, this study uses matched data from a subsample of 207 employees and their spouses to test a multi-source model, which revealed several notable inconsistencies. Our exploration of discrepancies between employee and spousal ratings helps explain these differences between single-source and multi-source models. Overall, findings suggest that employees consider various spousal factors in their career decision-making, and the familyrelatedness of relocation decisions is particularly strong among female employees. © 2015 Elsevier Inc. All rights reserved.

1. Introduction People's careers rarely unfold in a single employment setting anymore. Oftentimes careers are boundaryless, thus featuring substantial interorganizational mobility (Arthur & Rousseau, 1996). Today's workforce is faced with the challenge to exhibit occupational mobility and flexibility in the work domain, yet such organizational demands are not easily reconcilable with employees' needs and their spouses' wishes and aspirations. The latter is particularly salient in the context of women's increased labor force participation, as it is not uncommon for both adult partners in a household to have career ambitions. Western societies have witnessed a steady increase in dual-earner households and non-traditional gender roles in recent decades (e.g., Haddock, Zimmerman, Lyness, & Ziemba, 2006; Marks, 2006; Masterson & Hoobler, 2015), and as the traditional single-wage earner model fades, decision-making in the work domain increasingly is affected by family factors. We therefore believe that the shift toward dual-earner families has important implications for the study of work decisions. As Greenhaus and Powell (2012) note, the “family-relatedness of work decisions is a general phenomenon that applies to a variety of work-related decisions” (p. 246). One of those decisions involves relocation, or the decision to accept a job transfer that requires a change of location. With the shift toward boundaryless careers, employees must make relocation decisions at various stages throughout their careers. The decision to relocate constitutes a role entry decision (Powell & Greenhaus, 2010) that is central for career planning but may have important implications for the (members of the) family domain. The costs associated with moving are particularly high for dual-earner couples because the spouse must quit his or her job and leave a familiar social environment behind. Thus, an employee's decision to relocate could harm the spouse's well-being. In such situations, we argue, the decision-making process depends strongly on the characteristics of the affected spouse. ⁎ Corresponding author at: Department for Corporate Management, University of Hohenheim, Schloss Osthof/Ost, 70593 Stuttgart, Germany. E-mail addresses: [email protected] (J. Ullrich), [email protected] (H. Pluut), [email protected] (M. Büttgen).

http://dx.doi.org/10.1016/j.jvb.2015.06.003 0001-8791/© 2015 Elsevier Inc. All rights reserved.

2

J. Ullrich et al. / Journal of Vocational Behavior 90 (2015) 1–12

The present study tests a mediational model of relocation willingness that is based on the reasoning of the family-relatedness of work decisions (FRWD) framework (Greenhaus & Powell, 2012). As “gender differences … are essential to consider to fully understand the work–family interface” (Eby, Casper, Lockwood, Bordeaux, & Brinley, 2005, p. 181), our main goal here is to examine whether men and women equally consider their spouses' needs in deciding on relocation. A recent review on the role of gender in work–family decisions (Powell & Greenhaus, 2010) suggested that gender moderates the influence of family-related factors on work-domain factors. Hence, we set out to provide an in-depth examination of gender in the family-relatedness of relocation decisions. 2. Contributions of the present study Despite abundant research on relocation and its impact on employees and their spouses, we still lack understanding about how and when the spouse's attributes are taken into account during relocation decision-making. This study aims to make a three-fold contribution to theory on work and family in general and to research on relocation decision-making more specifically. First, we examine which factors affect people's willingness to make a so-called joint move when one spouse has received a relocation offer. Numerous studies have investigated determinants of the willingness to relocate among employees, proposing and finding that the spouse's willingness to follow the employee is a key factor in influencing the employee's intentions. We intend to build on this stream of research by examining determinants of a spouse's willingness to move with the focal employee. Specifically, we put forward and test a model of relocation willingness that conceptualizes spousal willingness as a mediating construct between spouse attributes and employee willingness. Second, we pursue a contribution to research on relocation decision-making by incorporating the perspective of the spouse through the use of a dyadic study design. Most studies regarding relocation decisions are characterized by an exclusive focus on the employee's perspective, thus relying on employee perceptions of spouse attributes (for an exception, see Eby & Russell, 2000). Although decisions ultimately are based on employees' perceptions of their spouses, it is imperative to validate perceptual self-reports and investigate the correspondence between employee perceptions and spousal ratings. With a dyadic survey design, we sample both employees and their spouses, which helps to address the criticism that previous research has adopted a one-sided perspective. We set out to test our proposed model not only based on employee perceptions but also based on multi-source data (i.e., from both employees and their spouses). Third, and this perhaps constitutes our most important contribution, we aim to provide an in-depth examination of gender differences in the relocation decision-making process. We will examine whether men and women differ in the degree to which they consider their spouses while making this work-related decision. To further capitalize on the dyadic nature of the data, we intend to explore potential differences in perceptual biases between men and women when it comes to evaluating their spouses' characteristics. We believe that through the examination of both employees' and spouses' perceptions, the present study provides unique insights into the family-relatedness of work decisions and relevant gender differences therein. 3. Theoretical framework Career decisions are influenced by various personal and family factors, resulting in so-called customized careers (Valcour, Bailyn, & Quijada, 2007). Customized careers involve a series of choices through which the employee attempts to construct a professional identity and creates arrangements that allow him or her to have a satisfying family life. Scholars thus need to look “at careers not only as the evolving sequence of paid work but also including how the employment cycle intersects with the family cycle and, in general, how it fits into a person's life experience” (Valcour et al., 2007, p. 194). The influence of family factors on career decisions has attracted considerable scientific interest (Powell & Greenhaus, 2010, 2012; Tharenou, 2008). Especially for relocation decisions, family structure and the spouse's characteristics have fundamental roles (Baldridge, Eddleston, & Veiga, 2006; Bielby & Bielby, 1992). In building our conceptual model (see Fig. 1), we therefore draw on a framework proposed by Greenhaus and Powell (2012) that emphasizes the family-relatedness of work decisions (FRWD). In this framework, work-related decisions are influenced by the family situation, which is defined as “a bundle of family-related pressures, demands, responsibilities, or needs that call for the attention of a focal individual and have potential implications for the well-being of the individual's family” (Greenhaus & Powell, 2012, p. 248). The decision to relocate is an impactful decision that affects the well-being of other family members, in particular the spouse. Relocation poses a threat to the social needs of accompanying spouses and interferes with their job or career aspirations. Indeed, spouse loss of social ties and problems with spouse employment are prominent stressors for relocating families (Eby et al., 2005). In this context, we expect that feelings of responsibility for their spouses' well-being shape employees' decision-making process, such that they will consider their spouses' needs in order to benefit the family or at least prevent harm to it. As such, the FRWD framework informs our focus on spousal factors as determinants of an employee's relocation willingness. When it comes to spouse-related variables that could influence an individual to accept a relocation offer, spousal attitudes toward relocation are a critical factor for employees to consider, and such attitudes strongly depend on the expected costs of moving. Relocation is associated with two inevitable consequences for the accompanying spouse, at least on the short term. First, relocation typically involves a career interruption for the spouse who accompanies his or her partner in a job-related move (Eby, 2001). That is, relocation constitutes a role entry decision for one person (i.e., to accept a job that entails a change of geographic location) yet a role exit decision for the other person (i.e., to quit one's current job). Second, an accompanying spouse (and other family members) is confronted with having to leave the social environment behind (Rodgers & Rodgers, 2000). We expect these consequences to be

J. Ullrich et al. / Journal of Vocational Behavior 90 (2015) 1–12

3

Spouse’s career salience H2a (-)

Spouse’s job alternatives

H2b (+)

Spouse’s willingness to relocate

H1 (+)

Employee’s willingness to relocate

H2c (-)

Spouse’s community ties

Fig. 1. Conceptual model.

strongly linked with relocation willingness. We therefore set out to examine the spouse's career salience and job alternatives (related to the consequence of quitting one's job) and the spouse's community ties (related to the social cost of relocation), and the FRWD framework would predict that these family-related variables (‘family situation’) influence a focal individual's willingness to relocate for work (‘work-domain decision’).

3.1. Hypotheses According to behavioral decision-making theory (March, 1994), employees weigh the costs and benefits of each career decision, and they do this in the broader work and family context, thus acknowledging that their decision may entail costs and benefits for the spouse. Most if not all career decisions (e.g., working longer hours, starting a new business, increasing job responsibilities) have potential implications for the well-being of the employee's family and thus influence work–family balance (De Hauw & Greenhaus, 2014). Jobrelated relocation is a career decision with major consequences for the family, because either the employee leaves the family behind or the family accompanies the employee (i.e., a “tied move”). We assume that in most cases, the employee prefers the second scenario, so she or he likely takes the spouse's willingness to relocate into account when making the decision. Many studies already have documented the relationship between spousal attitudes toward a potential move and the employee's willingness to relocate (Brett & Reilly, 1988; Challiol & Mignonac, 2005; Eby & Russell, 2000). Accordingly, we expect that the spouse's encouragement and support (manifested in willingness to accompany the employee) is important in influencing the employee's relocation decision processes. Hypothesis 1. The spouse's willingness to relocate is positively associated with the employee's willingness to relocate. Yet what influences an individual's willingness to move for the other person's job? The decision to relocate is associated with potential costs (or losses) for the spouse, in both the work and social domains, which shape the spouse's willingness to move for the employee's job. These losses, however, are not equally perceived across individuals. The prospect of relocating becomes more viable when the weight of losses is reduced. Below, we propose three factors that may influence how costly relocation would be for the accompanying spouse. First, moving implies that the spouse must quit her or his job, which typically represents a career interruption. Accompanying the employee in a job-related relocation could lower pay and promotion opportunities (Eby, 2001; Mincer, 1978). Spouses are less willing to incur such costs when they score higher on career salience, which indicates the overall importance that people attach to pursuing a (successful) career and the extent to which their career is central to their self-concept (Sekaran, 1982). The detrimental consequences of quitting one's job and starting over in a new place should be perceived as particularly strong deterrents among career-oriented spouses. In the more specific context of international assignments, prior research also reveals that career salience is linked negatively to willingness to follow a partner abroad (Van der Velde, Bossink, & Jansen, 2005).

4

J. Ullrich et al. / Journal of Vocational Behavior 90 (2015) 1–12

Hypothesis 2a. The spouse's career salience is negatively associated with his or her willingness to relocate. Faced with job loss, the spouse likely considers prospects for reemployment, as job searches can be highly stressful endeavors (Barber, Daly, Giannantonio, & Phillips, 1994). Those who perceive strong job opportunities for themselves (e.g., due to highly marketable skills) generally fare better following a tied move, because of the greater reemployment quality available (Eby, 2001). Despite the fact that many modern organizations increasingly focus on the spouses of their employees in human resource policies (e.g., partner employment assistance) (Brett & Stroh, 1995; Eby & Allen, 1998), accompanying spouses run the risk that they do not find an equivalent job in the new location. We predict that spouses may therefore base their willingness to relocate on their perception of job alternatives in the destination. Hypothesis 2b. The spouse's job alternatives are positively associated with his or her willingness to relocate. Second, strong attachments to a current location may influence the spouse's willingness to relocate. Eby and Russell (2000) note that it may be daunting to leave a current location if family members have strong attachments to the community. Consistent with this notion, prior research reveals a negative relationship between an employee's community ties and his or her willingness to relocate (e.g., Baldridge et al., 2006; Gould & Penley, 1985). The spouse's community ties should exert a similar influence on spousal attitudes toward relocation. Relocation invokes social costs, in the form of the loss of social and family support groups (Rodgers & Rodgers, 2000), so we predict that strong attachments to the community lower the spouse's willingness to move. Hypothesis 2c. The spouse's community ties are negatively associated with his or her willingness to relocate. These combined arguments imply that spousal willingness to relocate mediates the relationships between the spouse's career salience, job alternatives, and community ties on the one hand and the employee's willingness to relocate on the other hand. Formally stated, Hypothesis 3a. The negative influence of the spouse's career salience on the employee's willingness to relocate is mediated by the spouse's willingness to relocate. Hypothesis 3b. The positive influence of the spouse's job alternatives on the employee's willingness to relocate is mediated by the spouse's willingness to relocate. Hypothesis 3c. The negative influence of the spouse's community ties on the employee's willingness to relocate is mediated by the spouse's willingness to relocate. This career decision-making process also may be influenced by gender, such that the family-relatedness of the work decision is stronger for female than for male employees. Women often appear less willing to accept job transfers than men (Stroh, 1999). Eby (2001) finds in her study sample that most accompanying spouses are women and that women are more likely to move repeatedly for their husbands' careers than vice versa. Such gender differences in career decision-making can be explained with gender role theory, which states that women traditionally are socialized to be considerate and understanding and to exhibit compassionate behavior (Deaux & LaFrance, 1998). Traditional gender roles and stereotypes emphasize feminine traits—which are particularly important in family domains—as appropriate for women (Powell & Greenhaus, 2010). Women are expected to demonstrate concern for others and selflessness; they are encouraged to make career decisions that enhance the well-being of the family. Women are not expected to prioritize their own careers over their husbands' or to choose work over family (Baldridge et al., 2006). Thus, a woman who considers relocating for her own career violates traditional gender role expectations. Not surprisingly, then, women report family considerations more often as a reason for their unwillingness to accept a job transfer than their male counterparts do (Bielby & Bielby, 1992). We conclude that the family-relatedness of work decisions is stronger for women than for men. Specifically, we predict that women are more likely to subjugate their career opportunities for their spouses and are more sensitive to the socioemotional costs of relocation. Therefore, the effect of the spouse's characteristics (i.e., career salience, job alternatives, and community ties) on an employee's willingness to relocate should be stronger for female employees with male spouses than for male employees with female spouses. Hypothesis 4a. The relationship between the spouse's career salience and the employee's willingness to relocate is stronger for female than for male employees. Hypothesis 4b. The relationship between the spouse's job alternatives and the employee's willingness to relocate is stronger for female than for male employees. Hypothesis 4c. The relationship between the spouse's community ties and the employee's willingness to relocate is stronger for female than for male employees.

4. Methods 4.1. Sample and procedure This study was part of a larger data collection in Germany among global companies whose employees are typically faced with mobility issues and career path decisions. Respondents were recruited from eight participating companies (although two companies

J. Ullrich et al. / Journal of Vocational Behavior 90 (2015) 1–12

5

were overrepresented), which were active in various industries, such as automotive, consumer goods, building services engineering, and pharmaceuticals. We targeted professional staff (thus excluding line workers) and ultimately recruited a convenience sample of 4491 employees. This set of employees was sent digital links to the survey. At the end of the survey, employees were asked to involve their spouses in the study as well (assuming they were in a relationship). They could either enter the e-mail address of their spouse or copy the spouse survey link and send it themselves. Employee and spouse survey links contained an identification number that enabled us to match dyadic data. We received (usable) surveys from 1359 employees, for a response rate of 30%, but we retained only those employees that were in a relationship. These 1234 employee surveys could be matched to 211 (usable) spouseprovided surveys. Analysis of the descriptive statistics revealed that, on average, couples had been in a relationship for 13.4 years (range 1–44), most (62.4%) had at least one child living at home, and 6.2% indicated that they cared for other relatives. The employee sample consisted of 686 men and 548 women. The average age of the employees was 40.6 years (range 20–58), their average job tenure was 12.1 years, 51.9% of the participants had managerial responsibility, and the mean relative contribution to household income was 62%. In the matched employee–spouse sample, we had 127 male and 84 female employees, while the spouse sample included 88 men and 123 women. Accordingly, we identified four same-sex and 207 opposite-sex couples, resulting in percentages of 51% men and 49% women in total. With our focus on gender differences, we excluded the four same-sex couples, resulting in a final sample size of 207 couples. The average age of this set of employees was 39.7 years (range 16–58) and that of the spouses was 39.1 years (range 16–63). In the employee subsample, the average job tenure was 10.6 years, 58% of the participants had managerial responsibility, and the mean relative contribution to household income was 63%. In the spouse sample, 89% were employed, the average length of their job tenure was 8.2 years, 21% of the participants had managerial responsibility, and the mean relative contribution to household income was 33%. 4.2. Measures Each of the spousal variables in our model was rated by both employees and their spouses. We conducted exploratory factor analysis (principal component analysis [PCA] using Varimax rotation) and confirmatory factor analysis on all items (from both the employee and the spouse survey) to assess measurement validity of each construct. In all instances, the PCA extracted two components that corresponded to the employee and spousal ratings. Confirmatory factor analyses further confirmed the two-factor structure for the constructs and showed that all items had significant loadings above .70 on their respective factors. 4.2.1. Willingness to relocate We measured employees' willingness to relocate with a four-item scale developed by Landau, Shamir, and Arthur (1992), including “I would accept a geographic relocation for my job.” The seven-point Likert scale ranged from 1 = I totally disagree to 7 = I totally agree, and scale reliability was α = .91. Landau et al. (1992) describe the multidimensional structure of willingness to relocate—differentiating between readiness to relocate to remain employed and for career enhancement—but we did not find such a structure. Instead, the results supported a unidimensional pattern (one factor with an eigenvalue greater than 1; 78% of the variance explained by the main component). We measured the employee's perception of the spouse's willingness to relocate with one item developed by Challiol and Mignonac (2005): “If I had a proposition to relocate, my partner would be willing to follow me” (1 = not at all, 7 = for sure), which we modified slightly for the spouse survey (“If my partner had a proposition to relocate, I would be willing to follow him/her”). 4.2.2. Community ties We measured the spouse's community ties (in the spouse survey) with three items from a scale developed by Lee, Mitchell, Sablynski, Burton, and Holtom (2004), such as “I think of the community where I live as home.” In addition, to refer to employees' perceptions of their spouses' community ties, we slightly modified the items (e.g., “My partner thinks of the community where he/ she lives as home”). The seven-point Likert scale ranged from 1 = I totally disagree to 7 = I totally agree. The internal consistency was adequate (α = .90 for spouse's community ties, α = .89 for employee's perception of spouse's community ties). 4.2.3. Career salience We used three items from the Life Role Salience Scales (Amatea, Cross, Clark, & Bobby, 1986) to measure the spouse's career salience (in the spouse survey). Items included statements such as “It is important for me to make a career.” To refer to employees' perceptions of their spouses' career salience, we slightly modified the items (e.g., “It is important for my partner to make a career”). The seven-point Likert scale ranged from 1 = I totally disagree to 7 = I totally agree. The internal consistency of the scales was good (α = .93 for spouse's career salience, α = .94 for employee's perception of spouse's career salience). 4.2.4. Job alternatives for the spouse We used a one-item measure developed by Rusbult, Farrell, Rogers, and Mainous (1988) and adapted it to fit the two perspectives (i.e., employee's and spouse's perception of job alternatives for the spouse) as follows: “In case of a relocation for my job, my partner would be sure of finding work that would satisfy him (her) if he (she) was forced to quit his (her) work” and “In case of a relocation for my partner's job, I would have no problem of finding new work that satisfies me.” Again, the seven-point Likert scale ranged from 1 = I totally disagree to 7 = I totally agree.

6

J. Ullrich et al. / Journal of Vocational Behavior 90 (2015) 1–12

4.2.5. Control variables The following were considered as control variables: presence of children, if spouse is currently employed, if spouse holds managerial position, spouse's job tenure, if employee holds managerial position, employee's job tenure, and employee's contribution to household income (%). Separate analyses with and without control variables (to examine their effect on the observed relationships; Spector & Brannick, 2011) revealed that none of these variables influenced the results in a meaningful way. Thus, we did not include control variables in the analyses reported below. 4.3. Analyses Our dyadic study design enables us to test our hypotheses in two distinct ways. First, we test our proposed model on the basis of ratings provided by employees (n = 1234). All variables thus stem from the employee survey, and the measurements are based on a single source. We believe it is interesting to focus on the employee's perception of the spouse's characteristics and attitudes because individuals ultimately act on perceptions (i.e., to what extent does the employee think that her or his spouse is willing to accompany her or him on a relocation, and to what extent does this perception influence the employee's own willingness to relocate?). However, because same-source relationships are subject to common method bias that could account for significant effects, we also test our proposed model using multi-source data, stemming from both the employee and the spouse survey (n = 207). That is, we rely on spouses' own ratings of their characteristics (community ties, job alternatives, career salience) and their willingness to relocate. With this approach, we address the issue of same-source bias while simultaneously extending our examination of FRWD by investigating spousal ratings of spouse factors (cf. employee perceptions of such factors). To test our hypothesized model, we performed structural equation modeling using Mplus software (version 7). We ran the model using maximum likelihood estimation. To test the indirect effects of the spouse's characteristics on the employee's willingness to relocate (Hypotheses 3a, 3b and 3c), we conducted a bootstrapping analysis (n = 5000). To test for the moderating role of gender (Hypotheses 4a, 4b and 4c), we conducted a difference test across the two groups (men versus women). 5. Results Table 1 presents the descriptive statistics and the correlational matrix. The composite reliabilities for all constructs were greater than the desired minimum of .70 (Bagozzi & Yi, 2012). We found evidence of convergent validity in the satisfactory average variance extracted (AVE) values. The minimum AVE was .71, implying that more variance was explained than unexplained by the indicators linked with the respective constructs. We examined the discriminant validity of the constructs using the Fornell-Larcker criterion, according to which the correlation of a construct with its indicators (i.e., the square root of the AVE) must be greater than the correlation between the construct and any other construct (Fornell & Larcker, 1981). The results in Table 1 satisfied this condition. We set out to test our hypothesized model using both single-source (employee) data and multi-source (employee and spouse) data. We first report the results from the single-source data. As a first step, we tested our mediation model (see Fig. 1) in full, which generated good overall fit indices, such that the data supported the modeled relationships. The root mean square error of approximation (RMSEA) was .041, the chi-square/df statistic (χ2/df) was 1.41, and the comparative fit index (CFI) was .98. These values exceed the suggested thresholds (Bagozzi & Yi, 2012). We depict the results in Fig. 2. In support of Hypothesis 1, employees were more willing to relocate when they perceived their spouses to be willing to accompany them (β = .44, p b .001). Moreover, employees perceived higher spousal willingness to relocate when they believed that their spouses held weak community ties (β = − .12, p b .001) and would have many job alternatives (β = .41, p b .001). However, Table 1 Descriptive statistics and correlations.

1. Employee's willingness to relocateb 2. Spouse's willingness to relocate perceivedb 3. Spouse's willingness to relocatea 4. Spouse's community ties perceivedb 5. Spouse's community tiesa 6. Spouse's job alternatives perceivedb 7. Spouse's job alternativesa 8. Spouse's career salience perceivedb 9. Spouse's career saliencea M SD CR AVE

1

2

.84 .43⁎⁎ .26⁎⁎ −.20⁎⁎

− .39⁎⁎ −.09⁎

−.04 .16⁎ .08 .16⁎ .18⁎ 4.04 1.78 .91 .71

−.10 .41⁎⁎ .03 −.04 .01 4.22 1.82 − −

3

− −.24⁎⁎ −.02 .13⁎ .38⁎⁎ −.04 .09 4.91 1.79 − −

4

.86 .39⁎⁎ .06 −.12 −.02 −.20⁎⁎ 5.83 1.24 .89 .73

5

.88 −.01 −.03 −.01 .04 5.73 1.46 .91 .77

6

7

8

9

.04 .22⁎⁎ 4.19 1.78 − −

.92 .60⁎⁎ 3.30 1.83 .94 −.84

.90 3.58 1.71 .93 .82

−

.23⁎⁎ .04 −.04 3.63 1.90 − −

−

Note. Ns = 187 to 1234 (due to nonworking spouses). CR = composite reliability. AVE = average variance extracted. The square root of the AVE is on the diagonal. a Variables are rated by spouses. b Variables are rated by employees. ⁎ p b .05. ⁎⁎ p b .01.

J. Ullrich et al. / Journal of Vocational Behavior 90 (2015) 1–12

7

Spouse’s career salience perceived -.02

Spouse’s job alternatives perceived

.41***

Employee’s willingness to relocate

.44***

Spouse’s willingness to relocate perceived

-.12***

Spouse’s community ties perceived

= Variables of Employee Survey

= Variables of Spouse Survey

Fig. 2. Mplus results for the single-source model. Note. Standardized coefficients are shown. Ns = 1093 to 1234 (due to nonworking spouses). We replicated the model controlling for several variables (see section ‘control variables’). Children at home and employee's job tenure negatively predicted employee's willingness to relocate and spouse's job tenure was a significant negative predictor of spouse's willingness to relocate perceived. Inclusion of these variables did not change the coefficients in a meaningful way. *p b .05. **p b .01. ***p b .001.

perceptions of the spouse's career salience did not predict perceptions of spousal willingness to relocate (β = −.02, p = .50). Thus, we found support for Hypotheses 2b and 2c but not for Hypothesis 2a. Table 2 contains the results of the bootstrapping procedure used to test the mediation hypothesis (Hypotheses 3a, 3b and 3c). Indirect effects are significant when their respective confidence intervals do not contain zero (Lau & Cheung, 2012). We estimated the

Table 2 Test of indirect effects for the single-source model. Indirect effect

Spouse's career salience → Employee's willingness to relocate

Spouse's job alternatives → Employee's willingness to relocate

Spouse's community ties → Employee's willingness to relocate

Confidence interval

All Male employees Female employees Difference test All Male employees Female employees Difference test All Male employees Female employees Difference test

Lower 0.5%

Lower 2.5%

Lower 5%

Estimate

Upper 5%

Upper 2.5%

Upper 0.5%

−0.041 −0.026 −0.049 −0.043 0.127 0.086 0.132 −0.162 −0.137 −0.140 −0.184 −0.068

−0.032 −0.015 −0.038 −0.026 0.137 0.096 0.148 −0.140 −0.123 −0.121 −0.161 −0.044

−0.029 −0.009 −0.032 −0.018 0.142 0.103 0.155 −0.128 −0.115 −0.111 −0.149 −0.031

−0.008 0.020 −0.004 0.025 0.170⁎⁎ 0.136⁎⁎ 0.201⁎⁎ −0.070⁎ −0.080⁎⁎ −0.063⁎ −0.098⁎⁎

0.012 0.051 0.024 0.067 0.203 0.177 0.254 −0.013 −0.048 −0.020 −0.054 0.109

0.015 0.058 0.029 0.074 0.209 0.184 0.263 −0.003 −0.042 −0.011 −0.045 0.122

0.022 0.068 0.041 0.093 0.221 0.198 0.287 0.016 −0.028 0.002 0.028 0.151

0.038

Note. Bias-corrected confidence intervals for the indirect effects and the difference test. Bootstrap sample size = 5000. Indirect effects are mediated by the spouse's willingness to relocate. All variables are rated by employees. ⁎ p b .05. ⁎⁎ p b .01.

8

J. Ullrich et al. / Journal of Vocational Behavior 90 (2015) 1–12

confidence intervals for the 90%, 95%, and 99% levels. The indirect effects of community ties and job alternatives on the employee's willingness to relocate, through spousal willingness to relocate, were statistically significant (perceived community ties ab = −.08, p b .01; perceived job alternatives ab = .17, p b .01). The relationship between the spouse's job alternatives and the employee's willingness to relocate was fully mediated by perceptions of spousal willingness. The negative indirect effect from the spouse's career salience to the employee's willingness to relocate was not significant (not surprising, considering the lack of support for Hypothesis 2a). Thus, we found support for only Hypotheses 3b and 3c among the mediation hypotheses. To test for gender differences in the FRWD (Hypotheses 4a, 4b and 4c), we examined whether the hypothesized mediated relationships differed among men and women, using the model constraint command in Mplus. This test of difference resulted in a z-score, as reported in Table 2. First, we investigated whether the indirect effect of the spouse's career salience perceived on the employee's willingness to relocate was moderated by gender (Hypothesis 4a). The z-score was not significant (z = .025, p N .10), so the two groups (men vs. women) did not differ in the examined relationship (i.e., no support for Hypothesis 4a). We proceeded in a similar way for Hypotheses 4b and 4c. For the path starting with the spouse's job alternatives perceived, the analysis revealed that this relationship was moderated by gender (z = −.07, p b .05); the positive association between a spouse's job alternatives and an employee's willingness to relocate was stronger for female than for male employees (women ab = .20, p b .01; men ab = .14, p b .01), in support of Hypothesis 4b. The test of difference for the path from the spouse's community ties perceived did not provide significant results (z = .04, p N .10), resulting in the rejection of Hypothesis 4c.

5.1. Additional analyses We replicated the former series of analyses using both employee and spousal ratings, to address the frequently raised criticism that empirical findings are often based on an employee's perception of spouse characteristics only. The results of the mediation model based on the multi-source data are presented in Fig. 3. The RMSEA was .040, the chi-square/df statistic (χ2/df) was 1.34, and the CFI was .99.

Spouse’s career salience .02

Spouse’s job alternatives

.35***

Spouse’s willingness to relocate

.25***

Employee’s willingness to relocate

.05

Spouse’s community ties

= Variables of Employee Survey

= Variables of Spouse Survey

Fig. 3. Mplus results for the multi-source model. Note. Standardized coefficients are shown. Ns = 187 to 207 (due to nonworking spouses). We replicated the model controlling for several variables (see section ‘control variables’). Children at home negatively predicted employee's willingness to relocate and spouse's job tenure was a significant negative predictor of both employee's and spouse's willingness to relocate. Inclusion of these variables did not change the coefficients in a meaningful way. *p b .05. **p b .01. ***p b .001.

J. Ullrich et al. / Journal of Vocational Behavior 90 (2015) 1–12

9

We again confirmed Hypothesis 1, because we found in the multi-source data analysis that employees were more willing to relocate when their spouses were willing to accompany them (β = .25; p b .001). As we show in Fig. 3, we found support for the spouse's job alternatives as a determinant of spousal willingness to relocate (β = .35, p b .001), in line with Hypothesis 2b, but not for the spouse's career salience (Hypothesis 2a; β = .02, p = .81) or the spouse's community ties (Hypothesis 2c; β = .05, p = .51). The latter finding is not consistent with our results from the analysis based on employee perceptions alone. The bootstrapping analysis to test for the indirect effects in the multi-source model revealed that the mediated effect from the spouse's job alternatives to the employee's willingness to relocate was statistically significant (ab = .09, p b .01), but the mediated effects of community ties (ab = .003, p N .10) and career salience (ab = .005, p N .10) were not. Thus, we found support for Hypothesis 3b but not Hypothesis 3a or Hypothesis 3c, as we summarize in Table 3. In line with the procedure we described previously, we addressed Hypotheses 4a–4c with a test of difference for the indirect effects for men versus women. In line with results from the single-source model, we did not find significant gender differences regarding the effect of the spouse's career salience (z = −.05, p N .10), so we reject Hypothesis 4a. Nor can we accept Hypothesis 4b, which suggested gender differences for the relationship between the spouse's job alternatives and the employee's willingness to relocate, on the basis of results from the multi-source model (z = −.07, p N .10). With regard to gender differences for the mediated relationship between the spouse's community ties and the employee's willingness to relocate, Hypothesis 4c received support in the multi-source model (z = .20, p b .05). The indirect effect of the spouse's community ties was significant only for female employees (ab = −.17, p b .05), not for male employees (ab = .02, p N .10). The partially inconsistent findings between the single- and multi-source models prompted us to explore discrepancies between employees' perceptions and their spouses' own evaluations more closely. Although employees' decision-making ultimately depends on their perceptions, it is important to examine whether those perceptions are in line with the spouse's actual attitudes (rated by the spouse). We aimed to examine possible distortion effects, due to perceptual biases, in the context of relocation decision-making. Here, we focused on the drivers of spousal willingness to relocate. In addition to examining the relationship between employee perceptions and spousal ratings, we aimed to explore how perceptions form, by investigating whether employees base their perceptions of the spouse's characteristics on their own characteristics. We related the employee's perception of the spouse's characteristics to the employee's own characteristics, as well as to the spousal ratings, with the results in Fig. 4. The employee's perception of the spouse's career salience represented a relatively accurate estimate of spousal ratings (β = .65, p b .001). Yet this perception was also associated with the employee's own career salience (β = .16. p = .01). We found a different pattern of results for perceptions of the spouse's community ties, which were explained to a large extent by the employee's own community ties (β = .57, p b .001) and therefore formed a somewhat less accurate estimate of the spousal ratings of community ties (β = .32, p b .001). Employees' perceptions of the spouse's job alternatives after a relocation formed an approximate estimation of how spouses assessed their own potential job perspectives (β = .26, p b .001). We also explored to what extent gender influenced such discrepancies by conducting chi-square difference tests (unconstrained models versus equality constrained models). The results suggested that men and women differ in the extent to which they perceive their spouses' characteristics accurately (as indicated by a match with the spousal rating). We found significant (Δχ2(1) = 6.18, p b .025) differences between men and women regarding their evaluation of the spouse's community ties, such that women were more accurate in perceiving their spouses' community ties (male employees β = .28, p b .001; female employees β = .47, p = .001). Instead, male employees drew more strongly (Δχ2(1) = 3.05, p b .10) on their own community ties (β = .65, p b .001) than did female employees (β = .46, p b .001). A similar pattern of results emerged for perceptions of the spouse's career salience. Again, these results pointed to gender differences such that men, more than women (Δχ2(1) = 2.73, p b .10), based their perceptions of their spouses' career salience on their own career salience (male employees β = .23, p = .003; female Table 3 Test of indirect effects for the multi-source model. Indirect effect

Spouse's career saliencea → Employee's willingness to relocateb

Spouse's job alternativesa → Employee's willingness to relocateb

Spouse's community tiesa → Employee's willingness to relocateb

Confidence interval

All Male employees Female employees Difference test All Male employees Female employees Difference test All Male employees Female employees Difference test

Lower 0.5%

Lower 2.5%

Lower 5%

Estimate

Upper 5%

Upper 2.5%

Upper 0.5%

−0.050 −0.070 −0.061 −0.237 0.023 −0.034 −0.002 −0.274 −0.096 −0.036 −0.650 −0.024

−0.032 −0.054 −0.029 −0.195 0.037 −0.009 0.028 −0.217 −0.070 −0.016 −0.503 0.016

−0.025 −0.044 −0.013 −0.165 0.044 0.001 0.044 −0.193 −0.057 −0.008 −0.438 0.037

0.005 −0.004 0.046 −0.048 0.086⁎⁎ 0.049 0.124⁎ −0.069 0.003 0.024 −0.173⁎ −0.198⁎

0.043 0.015 0.152 0.014 0.152 0.126 0.255 0.027 0.055 0.092 −0.032 0.423

0.053 0.023 0.176 0.028 0.166 0.141 0.283 0.046 0.066 0.105 −0.008 0.471

0.075 0.045 0.227 0.060 0.198 0.192 0.349 0.083 0.090 0.138 0.031 0.600

Note. Bias-corrected confidence intervals for the indirect effects and the difference test. Bootstrap sample size = 5000. Indirect effects are mediated by the spouse's willingness to relocatea. a Variables are rated by spouses. b Variables are rated by employees. ⁎ p b .05. ⁎⁎ p b .01.

10

Employee’s career salience

.16**

Spouse’s career salience

.65***

Spouse’s job alternatives Employee’s community ties

.26***

Spouse’s job alternatives perceived

.57***

Spouse’s community ties perceived Spouse’s community ties

.32***

= Variables of Employee Survey

= Variables of Spouse Survey

Fig. 4. Results of distortion analysis. Note. Standardized coefficients are shown. *p b .05. **p b .01. ***p b .001.

J. Ullrich et al. / Journal of Vocational Behavior 90 (2015) 1–12

Spouse’s career salience perceived

J. Ullrich et al. / Journal of Vocational Behavior 90 (2015) 1–12

11

employees β = .02; p = .80). The relationship between spousal ratings of job alternatives and employee perceptions also was significantly (Δχ2(1) = 5.16, p b .05) moderated by gender, such that men were more accurate in assessing their spouses' job alternatives than women were (male employees β = .40, p b .001; female employees β = .08, p = .48).

6. Discussion This study has provided an examination of the extent to which job-related relocation depends on spousal characteristics and attitudes. Building on the FRWD framework (Greenhaus & Powell, 2012), we tested an integrative model in which spouse characteristics (career salience, job alternatives, and community ties) affect an employee's willingness to relocate. We found that this willingness depends strongly on the spouse's willingness to accompany the employee on his or her move. Spousal willingness in turn is higher when the spouse has many job alternatives and weak community ties. The family-relatedness of the relocation decision also is influenced by gender, such that female employees more strongly consider their spouses' job alternatives in the decision-making process than do their male counterparts. Although modern times are characterized by an apparent decline in adherence to traditional gender norms, women thus appear more likely to sacrifice their own career in favor of their husbands'. Beliefs about how men and women should behave have become more egalitarian, but women continue to be socialized to demonstrate concern for others, selflessness, and sensitivity (Rudman & Phelan, 2008). As a consequence, it seems that women (as spouses of employees) tend to accept the subordinate role of a tied mover, whereas men perceive relocation as a career sacrifice. We believe it would be a fruitful endeavor to incorporate measurements on such (pre-move and post-move) perceptions in future research. With our dyadic study design, we extend previous research by examining the decision-making process from both employee and spouse perspectives. An exclusive focus on employees means that scholars rely on employees' perceptions of the family situation. But, as our findings confirm, employees can have distorted perceptions of their spouses. Therefore, we replicated our model with spousal ratings of the spouse variables (cf. employees' perceptions of those spouse characteristics). The results were largely consistent across the single-source and multi-source models, though only for female employees were spousal ratings of community ties related to their willingness to relocate (thus, gender moderated this indirect effect in the multi-source analysis). In light of differences between the single- and multi-source models, we examined perceptual biases in the decision-making process. Greenhaus and Powell (2012) note that “individual decision makers may exhibit cognitive biases that distort their judgments about the likely consequences of their actions in decision-making situations” (p. 251). Relocation entails a decision to prioritize one's own career over the partner's career, and it may be susceptible to cognitive (or perceptual) biases, because employees may overestimate the extent to which their spouses support them in their decisions (and underestimate the extent to which their spouses must make sacrifices). Our results suggest that such distorted perceptions appear particularly common among men. Future research should explore the role of biases in perceptual processes and gender differences therein; it seems particularly interesting to examine the extent to which employees rely on personal interest to make decisions “for the greater good” (Pronin, Lin, & Ross, 2002). That is, they claim (or even think) the choice will benefit the family, even though they mainly do it for themselves. Distorted perceptions may be explained by our finding that employees largely base their perceptions of spouse characteristics on their own characteristics. For example, employees who felt less rooted in their social environment thought that their spouses also had weak community ties, consistent with an assimilation bias. People overestimate the extent to which their partners share their attitudes and traits (Sillars & Scott, 1983). For both community ties and career salience, this bias was stronger among men than women. Interestingly, men were more accurate than women in assessing their spouses' job alternatives. Considering the spousal ratings, female employees systematically underestimated their spouses' job alternatives following a move. However, as the evaluation of job alternatives reflects an expectation about a hypothetical job search, the latter finding also might reflect a biased self-perception on the part of the spouse, in the sense that men may overestimate their job prospects to a stronger degree than women do.

6.1. Limitations and future research Our dyadic design is a significant strength of our study. Nevertheless, several study limitations need to be noted. First, we had a relatively small matched sample (n = 207) to test the multi-source model, which raises the issue whether a lack of statistical power could account for the different results we found across the two sets of analyses (single-source versus multi-source). We therefore conducted a power analysis following Muthén and Muthén (2002). The results indicated that our sample sizes were sufficiently large to detect significant effects in all models we tested. Second, for some of our study variables, we relied on single-item measures, which we derived from previous research, yet scholars might develop and use multi-item scales for these constructs. Third, our study is based on cross-sectional data, which poses several problems. Although we addressed common source bias by testing a model that is based on both employee and spousal ratings, causal interpretations for the relationships in our model are not warranted. Related to this notion, we stress the need for longitudinal research, to examine both actual relocation and how employees and their accompanying spouses fare after the move. We only captured a small portion of the decision-making process; we do not know whether respondents actually received a job transfer offer. Fourth, we can only speculate why gender differences occur in the relocation decision-making process. Additional studies of FRWD should incorporate measures of some of the explanations put forward in this and other research for gender differences, such as feminine versus masculine traits or gender role ideology.

12

J. Ullrich et al. / Journal of Vocational Behavior 90 (2015) 1–12

6.2. Practical implications For organizations, the family-relatedness of relocation decisions implies that the willingness of employees to take on international assignments or engage in expatriation is not influenced solely by work-related factors (i.e., what can be offered and influenced by organizations). Thus, when recruiting new people, organizations need to be aware of family-related factors that might affect the prospective employee's decision to accept the job offer. Our (robust) finding regarding the effect of the spouse's job alternatives implies that recruiting organizations should offer relocation services and employment assistance as a form of institutional support (Eby & Allen, 1998) to improve the spouse's job prospects. Employers' policies should be tailored to the needs of the (dual-earner) couple as a whole, to make job-related relocation more attractive and increase the couple's willingness to relocate for work. Finally, human resource departments should provide assistance to their employees and discuss the consequences of relocation for the employee's family in detail. Such conversations can encourage employees to discuss the situation with their spouses and really consider relocation as a joint decision, to help reduce any distorted views among employees regarding their spouses' situation and willingness to accept a potential relocation. References Amatea, E. S., Cross, E. G., Clark, J. E., & Bobby, C. L. (1986). Assessing the work and family role expectations of career-oriented men and women: The life role salience scales. Journal of Marriage & Family, 48, 831–838. Arthur, M. B., & Rousseau, D. M. (1996). Introduction: The boundaryless career as a new employment principle. In M. B. Arthur, & D. M. Rousseau (Eds.), The boundaryless career: A new employment principle for a new organizational era (pp. 3–20). New York: Oxford University Press. Bagozzi, R., & Yi, Y. (2012). Specification, evaluation, and interpretation of structural equation models. Journal of the Academy of Marketing Science, 40, 8–34. Baldridge, D. C., Eddleston, K. A., & Veiga, J. F. (2006). Saying no to being uprooted: The impact of family and gender on willingness to relocate. Journal of Occupational and Organizational Psychology, 79, 131–149. Barber, A. E., Daly, C. L., Giannantonio, C. M., & Phillips, J. M. (1994). Job search activities: An examination of changes over time. Personnel Psychology, 47, 739–766. Bielby, W. T., & Bielby, D. D. (1992). I will follow him: Family ties, gender-role beliefs, and reluctance to relocate for a better job. American Journal of Sociology, 97, 1241–1267. Brett, J. M., & Reilly, A. H. (1988). On the road again: Predicting the job transfer decision. Journal of Applied Psychology, 73, 614–620. Brett, J. M., & Stroh, L. K. (1995). Willingness to relocate internationally. Human Resource Management, 34, 405–424. Challiol, H., & Mignonac, K. (2005). Relocation decision-making and couple relationships: A quantitative and qualitative study of dual-earner couples. Journal of Organizational Behavior, 26, 247–274. Deaux, K., & LaFrance, M. (1998). Gender. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (pp. 788–827). Boston: Oxford University Press. De Hauw, S., & Greenhaus, J. H. (2014). Building a sustainable career: The role of work–home balance in career decision making. Academy of Management Proceedings, 2014(1), 13367. Eby, L. T. (2001). The boundaryless career experiences of mobile spouses in dual-earner marriages. Group & Organization Management, 26, 343–368. Eby, L. T., & Allen, T. D. (1998). Perceptions of relocation services in relocation decision making: An exploratory field study. Group & Organization Management, 23, 447–469. Eby, L. T., Casper, W. J., Lockwood, A., Bordeaux, C., & Brinley, A. (2005). Work and family research in IO/OB: Content analysis and review of the literature (1980–2002). Journal of Vocational Behavior, 66, 124–197. Eby, L. T., & Russell, J. E. A. (2000). Predictors of employee willingness to relocate for the firm. Journal of Vocational Behavior, 57, 42–61. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39–50. Gould, S., & Penley, L. E. (1985). A study of the correlates of the willingness to relocate. Academy of Management Journal, 28, 472–478. Greenhaus, J. H., & Powell, G. N. (2012). The family-relatedness of work decisions: A framework and agenda for theory and research. Journal of Vocational Behavior, 80, 246–255. Haddock, S. A., Zimmerman, T. S., Lyness, K. P., & Ziemba, S. J. (2006). Practices of dual earner couples successfully balancing work and family. Journal of Family and Economic Issues, 27, 207–234. Landau, J. C., Shamir, B., & Arthur, M. B. (1992). Predictors of willingness to relocate for managerial and professional employees. Journal of Organizational Behavior, 13, 667–680. Lau, R. S., & Cheung, G. W. (2012). Estimating and comparing specific mediation effects in complex latent variable models. Organizational Research Methods, 15, 3–16. Lee, T. W., Mitchell, T. R., Sablynski, C. J., Burton, J. P., & Holtom, B. C. (2004). The effects of job embeddedness on organizational citizenship, job performance, volitional absences, and voluntary turnover. Academy of Management Journal, 47, 711–722. March, J. G. (1994). A primer on decision making: How decisions happen. New York: Free Press. Marks, S. R. (2006). Understanding diversity of families in the 21st century and its impact on the work–family area of study. In M. Pitt-Catsouphes, E. E. Kossek, & S. Sweet (Eds.), The work and family handbook: Multi-disciplinary perspectives, methods, and approaches (pp. 41–65). Mahwah, NJ: Erlbaum. Masterson, C. R., & Hoobler, J. M. (2015). Care and career: A family identity-based typology of dual-earner couples. Journal of Organizational Behavior, 36, 75–93. Mincer, J. (1978). Family migration decisions. Journal of Political Economy, 86, 749–773. Muthén, L. K., & Muthén, B. O. (2002). How to use a Monte Carlo study to decide on sample size and determine power. Structural Equation Modeling, 9, 599–620. Powell, G. N., & Greenhaus, J. H. (2010). Sex, gender, and decisions at the family → work interface. Journal of Management, 36, 1011–1039. Powell, G. N., & Greenhaus, J. H. (2012). When family considerations influence work decisions: Decision-making processes. Journal of Vocational Behavior, 81, 322–329. Pronin, E., Lin, D. Y., & Ross, L. (2002). The bias blind spot: Perceptions of bias in self versus others. Personality and Social Psychology Bulletin, 28, 369–381. Rodgers, J. R., & Rodgers, J. L. (2000). The effect of geographic mobility on male labor-force participants in the United States. Journal of Labor Research, 21, 117–132. Rudman, L. A., & Phelan, J. E. (2008). Backlash effects for disconfirming gender stereotypes in organizations. Research in Organizational Behavior, 28, 61–79. Rusbult, C. E., Farrell, D., Rogers, G., & Mainous, A. G., III (1988). Impact of exchange variables on exit, voice, loyalty, and neglect: An integrative model of responses to declining job satisfaction. Academy of Management Journal, 31, 599–627. Sekaran, U. (1982). An investigation of the career salience of men and women in dual-career families. Journal of Vocational Behavior, 20, 111–119. Sillars, A. L., & Scott, M. D. (1983). Interpersonal perception between intimates: An integrative review. Human Communication Research, 10, 153–176. Spector, P. E., & Brannick, M. T. (2011). Methodological urban legends: The misuse of statistical control variables. Organizational Research Methods, 14, 287–305. Stroh, L. K. (1999). Does relocation still benefit corporations and employees?: An overview of the literature. Human Resource Management Review, 9, 279–308. Tharenou, P. (2008). Disruptive decisions to leave home: Gender and family differences in expatriation choices. Organizational Behavior and Human Decision Processes, 105, 183–200. Valcour, M., Bailyn, L., & Quijada, M. A. (2007). Customized careers. In H. Gunz, & M. Peiperl (Eds.), Handbook of career studies (pp. 188–214). Newbury Park, CA: Sage. Van der Velde, M. E. G., Bossink, C. J. H., & Jansen, P. G. W. (2005). Gender differences in the determinants of the willingness to accept an international assignment. Journal of Vocational Behavior, 66, 81–103.