The mediating role of work-leisure conflict on job stress and retention of it professionals.

Zhao, Lin ; Rashid, Humayun

INTRODUCTION

During the past two decades, information technology (IT) function has experienced a rapid growth in most organizations due to the intense competition in the field, which has led more demanding and complex roles for knowledge workers (Huarng, 2001; Fox, 2002; Maudgalya et al, 2006). Various studies (Li & Shani, 1991; Ford, Heinen & Langkamer, 2007; Slattery, Selvarajan & Anderson, 2009) have analyzed the consequences of these changes, and in this paper we attempt to take a deeper look at job stressors and determine how they impact the retention likelihood of IT employees. Other scholars (Judge & Colquitt, 2004; Monsen & Boss, 2009) have pointed out that work-life conflict is a significant factor in employee's decision to leave or retain in an organization. In particular, the importance of work-leisure conflict in terms of the impact of various job stressors on retention is another crucial area that this research addresses.

Through a survey of IT professionals in two large centers of a global bank, we collected 575 responses to test our model. Findings followed by discussions and practical implications are also presented.

THEORETICAL FRAMEWORK AND HYPOTHESES

To study the impact of work-leisure conflict on the relationship between job stress and retention, we propose the following theoretical framework (see Figure 1) with five interrelated components.

[FIGURE 1 OMITTED]

Work-life conflict reflects "how work spills over into family time" (Russell, O'Connell & McGinnity, 2009). Work spillover into personal life of employees can manifest itself in several ways such as: amount of time spent at work and away from home (Piotrkowski, 1979); mental preoccupation and absorption at work that impacts life outside work (Kanter, 1977); and, physical challenge of work can fatigue an employee and drain the energy needed for carry out and non-work activities--he or she becomes too tired to effectively engage in personal activities and relationships (Crouter et. al., 1983).

As an important part of nonwork variables, leisure is defined as "a period of time free from paid work or other obligatory activities" (Parker, 1971). Following this definition, leisure is perceived as "opposite" and "neutral" (unrelated) to work practices (Parker, 1971; Parker, 1983). Leisure can potentially compensate for negative experience or insufficient positive outcomes associated with paid work (Pearson, 2008). On the contrary, work-leisure conflict is negatively related to job satisfaction (Ford, Heinen & Langkamer, 2007), organizational commitment (Siegel et al., 2005), retention (Monsen & Boss, 2009) and life quality (Rice, Frone & McFarlin, 1992). Based on these negative spillover effects, we hypothesize that:

H1: Work-leisure conflict is negatively related to retention.

Job stress in general has negative consequences on job outcomes for the employees and organizations, and it typically leads to higher intention to quit and increased employee turnover (Netemeyer, Burton, & Johnston, 1995). Following the literature (Netemeyer, Burton & Johnston, 1995; Gilboa et al., 2008; Monsen & Boss, 2009), we break down job stress into three main components: role ambiguity, role conflict and role overload. and specifically test their relationship to retention as well as how work-leisure conflict may play a mediating role this relationship.

Role ambiguity is the degree to which clear information is lacking. Specifically, IT professionals may deal with unclear expectations from users and changes under uncertain authority (Li & Shani, 1991). According to Ashforth and Saks (1996), role ambiguity is positively correlated with intention to quit. Similarly, Rafferty and Griffin (2006) argued that uncertainty at work in general was associated with intention to quit. When roles are not well defined, typical reaction of employees is negative leading to withdrawal which can eventually lead to employee leaving the organization (Harris & Mossholder, 1996). Based on these earlier findings of the multi-facet negative implications of role ambiguity we proposed the following two hypotheses:

H2: Role ambiguity is positively related to work-leisure conflict.

H3: Role ambiguity is negatively related to retention.

Role conflict is the multiple requirements and expectations from the role that impact role performance (Rizzo, House & Lirtzman, 1970) and clash due to their nature (Handy, 1985; Schi, 1996). It often occurs when conflicting demands are placed upon the individuals by their supervisors, peers, or subordinates, so this type of stress is more dominant in jobs that have vague descriptions and require abstract thinking and decision-making (Menon & Akhilish 1994). Since we are interested in job stress of IT professionals who are involved in such dynamics, we hypothesize:

H4: Role conflict is positively related to work-leisure conflict.

H5: Role conflict is negatively related to retention.

Role overload can manifest itself both qualitatively (difficult work) or quantitatively (too much work) and has been shown to have a relationship to various strain symptoms (physiological, psychological, and behavioral) among employees (Beehr & Newman, 1978; Cooper & Marshall, 1976; Almer & Kaplan, 2002). Ivancevich, Napier and Wetherbe (1983) found that work overload and time pressure were significant factors resulting IT work stress. Based on these negative consequences of role overload, we hypothesize:

H6: Role overload is positively related to work-leisure conflict.

H7: Role overload is negatively related to retention.

The combination of H1 and H2 shows the mediating role of work-leisure conflict in creating a positive effect towards role ambiguity on retention. H3 posits the direct effect of role ambiguity on retention. According to the clarification of conditions and decision points for mediational type inferences provided by Mathieu and Taylor (2006), we present H1, H2 and H3 to clearly emphasize mediating relationship without the confusion of indirect effects. Similarly, the combination of H1 and H4 shows the mediating role of work-leisure conflict on the positive effect of role conflict on retention. H5 posits the direct effect of role conflict on retention. The combination of H1 and H6 shows the mediating role of work-leisure conflict on the positive effect of role overload on retention. H7 posits the direct effect of role overload on retention.

METHOD

RESEARCH SETTING, DATA SOURCES, AND SAMPLING

We surveyed knowledge workers of Chinese and Indian origin, in a global European bank through an online survey. There were 577 complete responses obtained from two sites, for a response rate of 30%. A comparison of the responses from two sites revealed no significant mean differences for study variables. After checking the data consistency and homogeneity, we retained 568 individual responses for further analysis.

MEASUREMENT AND OPERATIONALIZATION

For all study constructs, we directly adapted the scale items from the literature. The Appendix lists the operational items we used for each construct, and Table 1 provides the univariate statistics for the constructs and the intercorrelations among them.

RETENTION.

Job satisfaction and retention have been found to be closely related in several studies (Kotze & Roodt, 2005). Our factor analysis revealed that in certain situations they may in fact be measuring the same things, so we have decided to create a new composition called retention likelihood that incorporates job satisfaction along with intention to quit and desire to remain as a single measure.

WORK-LEISURE CONFLICT.

Previous studies have measured a wide variety of possible effects of work spillover on home life and find that leisure is one of the four most important dimensions (Small & Riley, 1990; Stevens, Kiger & Riley, 2006). Following Small & Riley (1990), we used five items to measure work-leisure conflict.

ROLE AMBIGUITY.

It is defined as "the absence of adequate information which is required in order for persons to accomplish their role in a satisfactory manner" (Senatra, 1980). Since our data are collected from different countries, we adopted the measure from Glazer and Beehr's (2005) study which examined the role stressors on employee attitudes in multi-culture contexts.

ROLE CONFLICT.

It is defined as "the simultaneous occurrence of two (or more) sets of pressures such that compliance with one world make difficult or impossible compliance with the other" (Wolfe & Snoek, 1962). We also adopted the measure from Glazer and Beehr (2005). According to our factor analysis, one item was dropped from the original measure due to cross loading.

ROLE OVERLOAD.

It is defined as "having too much work to do in the time available" (Beehr et al., 1976). We adopted the scale from Peterson et al.'s (1995) cross nation study on role stress of middle managers.

METHOD OF ANALYSIS

The analytical approach involves measurement assessment of the key constructs and testing the hypothesized model. For the subjective measures, a combination of exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) procedures are used to assess the psychometric properties. We explicitly focused on the evidence for the convergent and discriminant validity of the study constructs.

FINDINGS

MEASUREMENT ANALYSIS

We estimate a fully disaggregated measurement model with the key observed indicator to ensure that the measures correspond only to their hypothesized constructs and evidenced acceptable reliability and validity. A confirmatory factor analysis of the study constructs using AMOS software yielded the following fit statistics: x2 = 625.03, d.f. = 307, p < 0.01; NFI = 0.94; NNFI (TLI) = 0.96; CFI = 0.97; RMR = 0.05; RMSEA = 0.043 (90% confidence interval = .038 - .048). On statistical, absolute, and relative fit, as well as substantive grounds, the posited measurement model provides a good fit to the data. Table 2 provides further support for the convergent and discriminant validity of the constructs. The estimated loadings for the relationship between individual indicants and their underlying construct are, without exception, large and significant (t-value > 8.0, p < .01). In addition, the reliability estimates are large and significant, ranging from .81 to .92, with an average reliability index of .88, which exceeds the conventional .70 criterion. In terms of discriminant validity, the variance extracted not only exceeds the average variance shared but also exceeds or is close to .50, the threshold value that Fornell and Larcker (1981) recommend. The preceding evidence provides robust support for the convergent and discriminant validity of study constructs.

HYPOTHESIZED MODEL ANALYSIS

Our empirical results are summarized in Table 3. Six out of seven hypotheses are supported.

Role ambiguity is negatively related to work-leisure conflict (P = .06, p < .05), which is negatively related to retention (P = -.43, p < .01). Thus, H1 and H2 are supported. In addition, role ambiguity has significant direct effect on retention (P = -.52, p < .01) after controlling for the effect of work-leisure conflict, so H 3 is supported. Based on Shrout and Bolger (2002), the mediation hypothesis is supported if both the antecedent -> intervening and the intervening -> outcome coefficients are significant. Therefore, the results indicate that work-leisure conflict partially mediates the relationship between role ambiguity and retention.

Similarly, role conflict is positively related to work-leisure conflict (P =.18, p < .01), and also directly related to retention (P =-.24, p < .01). So H4 and H5 are supported. The results reflect that an increasing emphasis on role conflict reduces retention, and work-leisure conflict partially mediates the relationship between role conflict and retention.

Role overload is positively related to work-leisure conflict (P =.26, p < .01), but it doesn't have significant direct effect on retention (P =-.04 p > .10). So H6 is supported, but H7 is not supported. The results suggest that work-leisure conflict fully mediates the relationship between role overload and retention.

DISCUSSION

In certain professions like IT, it is not un-common for professionals to work in high stress environments and put in extra hours to meet project deadlines. No wonder that turnover is typically very high in the IT profession. Our research gives better insights into this phenomenon and provides a window into how the various job related stressors and the work-leisure conflict may impact retention likelihood of employees.

Work-leisure conflict has turned out to be an important mediator between job stress and retention likelihood. Its strong inverse relationship with retention likelihood makes it an important factor in retaining employees even beyond the effects of job stressors. Typically work-leisure is part of the broader work-life balance, but we believe that it may be more useful to understand work-leisure thoroughly since when it comes to stress, people typically look for some type of leisure activities to get their mind away from job stress so they can find a way to relax themselves. If employees are not able to find this 'exit' from job stress then it increases their chances of leaving the company and find some other work where they can minimize the work-leisure conflict.

When employees have ambiguity in their role, their chances of staying with the company go down dramatically. People generally want to resolve ambiguity so they can perform well at work, especially in some complex fields like IT. This inverse relationship is the most strongly related to retention, relative to other two job stressors, which is consistent with the literature reviewed by Gilba et al. (2008). Moreover, it is roughly independent of the work-leisure conflict although such conflict does slightly mediate the effect of ambiguity on retention.

IT professionals have to deal with conflicts frequently at work due to technology change and technology involved organizational changes, and they often intend to quit jobs if the role conflict can't be handled effectively. However, as our findings show, this situation can get worse when employees also experience work-leisure conflict. In other words, when employees face role conflict and also experience work-leisure conflict, their retention likelihood goes down dramatically. When employees have high role conflict, giving them more leisure time may help with retention, but it will be critical to sooner than later since role conflict has a direct effect on retention likelihood as well.

Role overload appear to spillover the most into work-leisure conflict so that all the effect of role overload on retention likelihood is fully mediated through work-leisure conflict. This is an important finding, since when people are overload with work demands, it does not necessarily mean that they will quit the job. In fact, as our findings reveal, changes of retention likelihood under high role overload will go down only if they also face high work-leisure conflict. Another way to look at it is that organizations can increase work demands (role overload) and not have any significant impact on the retention likelihood as long as they can give employees some opportunity to have leisure outside of work and keep their work-leisure conflict level down.

Overall, we have been able to show that the three role stressors behave differently when it comes to their impact on retention likelihood, and that work-leisure conflict is an important factor that mediates these differences in terms of retaining organizational talents.

PRACTICAL IMPLICATIONS

The importance of job stress on retention has been widely known and acknowledged by executives. At the same time, current business climate requires extra demands from employees that often involve various role stressors. One practical implication of our research is that role ambiguity has much greater impact on retention likelihood than either role conflict or role overload, so organizations should give the highest priority to tackle role ambiguity in stress management.

Second practical implication of our findings is the importance of keeping work-leisure conflict as low as possible even when other role stressors are high. Our findings suggest that much of the impact of role stressors on retention goes through work-leisure conflict and if the management team can find ways to relieve this conflict, they may be able to retain their employees even though role stressors are high. Being little sensitive to employees' leisure needs outside of work can go a long way in compensating the negative effects of various job stressors.

LIMITATIONS

Prior research has shown that the relationship between role stressors and intention to quit may vary widely depending on the functional role and level of employees within the organizational hierarchy (Cole & Bruch, 2006; Singh, 1998). In our analysis, we have not tested whether the relationship between role stressors, work-leisure conflict and retention likelihood are robust across staff members (front-line workers) and managers, so the findings may not be generalized.

FUTURE RESEARCH

Since work-leisure conflict has shown to play an important role as mediator between role stressors and retention likelihood, it is important to do an extended study to test a much richer work-life conflict model including other related factors in additional to the work-leisure conflict. Also, testing the current model across different functional and hierarchical levels will provide better insights into the generalizability of our findings and if there are differences then a solid understanding of those differences.

APPENDIX

Operational Items used to Measure Various Study Constructs

Unless otherwise noted, we measured the following items on a seven-point Likert scale where 1 = "strongly disagree" and 7 = "strongly agree." The items marked with (R) were reversed to keep the consistency with other measures.

Retention Likelihood

Job satisfaction (Mak and Sockel, 2001)

JS1: All in all, I am satisfied with my job.

JS2: In general, I do not like my job. (R)

JS3: In general, I like working here.

Intention to quit (Baroudi, 1985)

ITQ1: I frequently think of quitting my job. (R)

ITQ2: I am planning to search for a new job during the next 12 months. (R)

ITQ3: If I have my own way, I will be working for this organization years from now.

Desire to remain (Steers, 1977)

DTR1: All things considered, I have a desire and intent to remain with this organization. Work-Leisure Conflict (Small & Riley, 1990)

WLC1: My job makes it difficult for me to enjoy my free time outside of work.

WLC2: The amount of time I spend working interferes with how much free time I have.

WLC3: Worrying about my job makes it hard for me to enjoy myself outside of work.

WLC4: Because I am often tired after work, I don't see friends as much as I would like.

WLC5: My job doesn't affect whether I enjoy my free time outside of work.

Stress

Role Ambiguity (Glazer & Beehr, 2005)

ROAM1: I feel certain about how much authority I have. (R)

ROAM2: I know that I have divided my time properly. (R)

ROAM3: I know what my responsibilities are. (R)

ROAM4: I know exactly what is expected of me. (R)

ROAM5: I know what the critical factor is in getting promoted. (R)

ROAM6: I know how I should handle my free time on the job. (R)

Role Conflict (Glazer & Beehr, 2005)

ROCO1: I have to do things that should be done differently.

ROCO2: I work under incompatible policies and guidelines.

ROCO3: I receive an assignment without the resources to complete it.

ROCO4: I have to buck (bend) a rule or policy in order to carry out an assignment.

ROCO5: I receive incompatible requests from two or more people.

ROCO6: I have to work under vague directives or orders.

Role Overload (Peterson et al., 1995)

ROOV1: There is a need to reduce some parts of my role.

ROOV2: I feel overburdened in my role.

ROOV3: I have been given too much responsibility.

ROOV4: My workload is too heavy.

ROOV5: The amount of work I have to do interferes with the quality I want to maintain.

AUTHORS' NOTE

Both authors contributed equally.

REFERENCES

Almer, E.D. & S.E. Kaplan (2002). The effects of flexible work arrangements on stressors, burnout, and behavioral job outcomes in public accounting. Behavioral Research in Accounting, 14, 1-34.

Ashforth, B. E. & A. M. Saks (1996). Socialization tactics: Longitudinal effects on newcomer adjustment. Academy of Management Journal, 39, 149-178.

Beehr, T. & J. Newman (1978). Job stress, employee health, and organizational effectiveness: A facet analysis model, and literature review. Personnel Psychology, 31, 665-699.

Blankertz, L. E. & S. E. Robinson (1996). Who is psychosocial rehabilitation worker? Psychiatric Rehabilitation Journal, 19(4), 3-13.

Boxall, P., K. Macky & E. Rasmussen (2003) Labour turnover and retention in New Zealand: The causes and consequences of leaving and staying with employers. Asia Pacific Journal of Human Resources, 41(2), 195-214.

Cole, M. S. & H. Bruch (2006). Organizational identity strength, identification, and commitment and their relationships to turnover intention: Does organizational hierarchy matter? Journal of Organizational Behaviour, 27, 585-605.

Cooper, C. L. & J. Marshall (1976). Occupational sources of stress: a review of the literature relating to coronary heart disease and mental ill health. Journal of occupational psychology, 49(1), 11-28.

Ford, M. T., B. A. Heinen & K. L. Langkamer (2007). Work and family satisfaction and conflict: A meta-analysis of cross-domain relations. Journal of Applied Psychology, 92(1), 57-80.

Fornell, C. & D. F. Larcker (1981) Evaluating Structural Equation Models with Unoberservable Variables and Measurement Error. Journal of Marketing Research, 18, 39-50.

Fox, R. (2002). IT Professional Pressure Cooker. Communications of the ACM, 45(10), 9-10.

Gilboa, S., A. Shirom , Y. Fried & C. L. Cooper (2008). A meta-analysis of work demand stressors and job performance: Examining main and moderating effects. Personnel Psychology, 61, 227-271.

Glazer, S. & T. A. Beehr (2005). Consistency of implications of three role stressors across four countries. Journal of Organizational Behavior, 26, 467-487.

Handy, C. (1985). Understanding Organizations. Harmondsworth: Penguin.

Harris, S. G. & K. W. Mossholder (1996). The Affective Implications of Perceived Congruence with Culture Dimensions During Organizational Transformation. Journal of Management, 22(4), 527-47.

Huarng, A. S. (2001). Burnout syndrome among information system professionals. Information Systems Management, 18(2), 15-20.

Hytter, A. (2007) Retention Strategies in France and Sweden. Irish Journal of Management, 28(1), 59-79.

Ivancevich, J., H. Napier, & J. Wetherbe (1983). Occupational Stress, Attitudes and Health Problems in the Information Systems Professional. Communications of the ACM, 26(10), 800-806.

Judge, T. J. & J. A. Colquitt (2004). Organizational justice and stress: The mediating role of work-family conflict. Journal of Applied Psychology, 89(3), 395-404.

Kanter, R. (1977). Men and women of the corporation. New York: Basic Books.

Kotze, K. & G. Roodt (2005). Factors that Affect the Retention of Managerial and Specialist Staff: An Exploratory Study of an Employee Commitment Model. Journal of Human Resource Management, 3(2), 48-55.

Li, E. & A. Shani (1991). Stress Dynamics of Information Systems Managers, A Contingency Model. Journal of Management Information Systems, 7(4), 107-130.

Mathieu, J. E. & S. R. Taylor (2006). Clarifying conditions and decision points for mediational type inferences in organizational behavior. Journal of Organizational Behavior, 27, 1031-1056.

Maudgalya, T. , S. Wallace, N. Daraiseh & S. Salem (2006). Workplace stress factors and 'burnout' among information technology professionals: a systematic review. Theoretical Issues in Ergonomics Science, 7(3), 285-297.

Menon, N., & K. Akhilish (1994). Functionally dependent stress among managers. Journal of Managerial Psychology, 9, 13-22.

Monsen, E. & R. W. Boss (2009). The Impact of Strategic Entrepreneurship inside the Organization: Examining Job Stress and Employee Retention. Entrepreneurship Theory & Practice, 33(1), 71-104.

Netemeyer, R., S. Burton & M. Johnston (1995). A Nested Comparison of Four Models of the Consequences of Role Perception Variables. Organizational Behavior & Human Decision Processes, 61, 77-93.

Parker, S. (1971). The future of work and leisure. London: MacGibbon & Kee.

Parker, S. (1983). Leisure and Work. London: Allen and Unwin.

Peterson, M. F., P. B. Smith, A. Akande, S. Ayestaran, S. Bochner, V. Callan, N. G. Cho, J. C. Jesuino, M. D'Amorim, P. Francois, K. Hofmann, P. L. Koopman, K. Leung, T. K. Lim, S. Mortazavi, J. Munene, M. Radford, A. Ropo, G. Savage, B. Setiadi, T. N. Sinha, R. Sorenson & C. Viedge (1995). Role Conflict, Ambiguity, and Overload: A 21-Nation Study. The Academy of Management Journal, 38(2), 429-452.

Piotrkowski, C. S. (1979). Work and the family system. New York: Free Press.

Rafferty, A. E. & M. A. Griffin, (2006). Perceptions of organizational change: A stress and coping perspective. Journal of Applied Psychology, 91, 1154-1162.

Rice, R. W., M. R. Frone & D. B. McFarlin (1992). Work--nonwork conflict and the perceived quality of life. Journal of Organizational Behavior, 13(2), 155-168.

Rizzo, J. R., R. J. House & S. I. Lirtzman (1970). Role conflict and ambiguity in complex organizations. Administrative Science Quarterly, 15(2), 150-163.

Russell, H., P. J. O'Connell & F. McGinnity (2009). The impact of flexible working arrangements on work-life conflict and work pressure in Ireland. Gender Work and Organization, 6(1), 73-97.

Shrout, P. E. & N. Bolger (2002) Mediation in Experimental and Nonexperimental Studies: New Procedures and Recommendations. Psychological Method, 7(4), 422-445.

Singh, J. (1998). Striking a balance in boundary-spanning positions: an investigation of some unconventional influences of role stressors and job characteristics on job outcomes of salespeople. Journal of Marketing, 62, 69-86.

Slattery, J., T. Selvarajan & J. Anderson (2008). Influences of New Employee Development Practices on Temporary Employee Role Stressors and Work-Related Attitudes of Temporary Employees. International Journal of Human Resource Management, 19(12), 2268-2293.

Small, S. A. & D. Riley (1990). Toward a Multidimensional Assessment of Work Spillover into Family Life. Journal of Marriage and Family, 52(1 ), 51-61.

Stevens, D. P., G. Kiger & P. J. Riley (2006). His, hers, or ours? Work-to-family spillover, crossover, and family cohesion. The Social Science Journal, 43(3), 425-436.

Ward, E. A. (1988) Relation of job satisfaction and job knowledge and their effect on intention to turnover. Psychological Reports, 63(2), 611-615.

Wolfe, D. M. & J. D. Snoek (1962). A study of tensions and adjustments under role conflict. Journal of Social Issues, 18, 102-121.

Lin Zhao, Purdue University Calumet

Humayun Rashid, University of Southern California

Table 1 Summary of Statistics and Intercorrelations for the Study
Constructs

                       1       2       3       4     5   Mean   s.d.

1. Retention           1                                 4.80   1.48
2. Work-leisure     -.58 *     1                         4.24   1.76
   conflict
3. Role ambiguity   -.58 *   .31 *     1                 2.71   1.27
4. Role conflict    -.61 *   .59 *   .42 *     1         3.94   1.63
5. Role overload    -.49 *   .63 *   .28 *   .63 *   1   3.78   1.57

* Correlation is significant at the .01 level (two-tailed).

Table 2 Factor Loadings and Measurement Properties of Various
Constructs Used

                                                   Composite
Construct/Item          Loading (a)   t-Value   Reliability (c)

Retention likelihood                                  0.89
  JS1                       0.77       -- (b)
  JS2                       0.78       18.37
  JS3                       0.74       19.90
  ITQ1                      0.80       18.19
  ITQ2                      0.78       17.76
  DTQR1                     0.67       15.65
Work-leisure conflict                                 0.88
  WLC1                      0.86       12.44
  WLC2                      0.83       12.21
  WLC3                      0.81       12.10
  WLC4                      0.82       12.28
  WLC5                      0.51       -- (b)
Role ambiguity                                        0.81
  ROAM1                     0.55       -- (b)
  ROAM2                     0.48       10.61
  ROAM3                     0.86       13.41
  ROAM4                     0.88       13.26
  ROAM5                     0.59       10.97
  ROAM6                     0.40        8.18
Role conflict                                         0.88
  ROCO2                     0.70       17.67
  ROCO3                     0.77       19.67
  ROCO4                     0.72       18.07
  ROCO5                     0.85       23.15
  ROCO6                     0.81       -- (b)
Role overload                                         0.92
  ROOV1                     0.80       -- (b)
  ROOV2                     0.88       28.89
  ROOV3                     0.85       21.93
  ROOV4                     0.93       22.52
  ROOV5                     0.70       17.54
Goodness-of-Fit Statistics
  [chi square]            625.03
  d.f.                    307
  p-value                   0.00
  NFI                       0.94
  NNFI (TLI)                0.96
  CFI                       0.97
  GFI                       0.93
  AGFI                      0.91
  RMR                       0.05
  RMSEA                     0.043
(90% Confidence Interval)   0.038-0.048

                          Variance      Variance    Variance
Construct/Item          Extracted (d)    Sharede   Shared (f)

Retention likelihood         0.57         0.07        0.12
  JS1
  JS2
  JS3
  ITQ1
  ITQ2
  DTQR1
Work-leisure conflict        0.60         0.04        0.06
  WLC1
  WLC2
  WLC3
  WLC4
  WLC5
Role ambiguity               0.43         0.02        0.05
  ROAM1
  ROAM2
  ROAM3
  ROAM4
  ROAM5
  ROAM6
Role conflict                0.60         0.08        0.14
  ROCO2
  ROCO3
  ROCO4
  ROCO5
  ROCO6
Role overload                0.70         0.07        0.14
  ROOV1
  ROOV2
  ROOV3
  ROOV4
  ROOV5
Goodness-of-Fit Statistics
  [chi square]
  d.f.
  p-value
  NFI
  NNFI (TLI)
  CFI
  GFI
  AGFI
  RMR
  RMSEA
(90% Confidence Interval)

(a) The estimates are standardized coefficients (all p < 0.01) and
t-values from maximum likelihood solution using AMOS 16.0.

(b) The corresponding coefficient was fixed to set the metric of the
latent construct.

(c) Estimated composite reliability in line with Fornell and Larcker
(1981).

(d) Estimated variance extracted by the corresponding latent construct
from its hypothesized indicators in line with Fornell and Larcker
(1981).

(e) Average of the variance shared between the corresponding latent
construct and all other constructs of study.

(f) The maximum variance shared between the corresponding latent
construct and all other constructs of study.

Table 3 Estimated Coefficients for the Hypothesized Model

                             Dependent Variables

                        Work-leisure conflict  Retention

Independent Variables
Work-leisure conflict            --            -.43 (.06)
Role ambiguity                .06 (.03)        -.52 (.04)
Role conflict                 .18 (.03)        -.24 (.04)
Role overload                 .26 (.03)        -.04 (.04)

Note: The results reported are unstandardized coefficents followed by
standard error in parentheses. Coefficients significant at p = .05 are
in bold.