Orientation: Recently, a new work-nonwork interference instrument was developed to measure the interference between work and nonwork roles in the South African context (Koekemoer, Mostert & Rothmann, 2010). However, no information is available on the psychometric properties of this instrument.

Research purpose: The objectives of this study were to investigate the internal validity (construct, discriminant and convergent validity), reliability and external validity (relationship with theoretically relevant variables, including job characteristics, home characteristics, burnout, ill health and life satisfaction) of the instrument.

Motivation for the study: Work-family interaction is a key topic receiving significant research attention. In order to facilitate comparison across work-family studies, the use of psychometrically sound instruments is of great importance.

Research design, approach and method: A cross-sectional survey design was used for the target population of married employees with children working at a tertiary institution in the North West province (n = 366). In addition to the new instrument, job characteristics, home characteristics, burnout, ill health and life satisfaction were measured.

Main findings: The results provided evidence for construct, discriminant and convergent validity, reliability and significant relations with external variables.

Practical/managerial implications: The new instrument can be used by researchers and managers as a test under development to investigate the interference between work and different nonwork roles (i.e. parental role, spousal role, work role, domestic role) and specific relations with antecedents (e.g. job/home characteristics) and well-being (e.g. burnout, ill health and life satisfaction).

Contribution/value-add: This study provides preliminary information on the psychometric properties of a new instrument that measures the interference between work and nonwork.

Key focus of the study and problem statement

In a very recent attempt to address certain issues relating to the measurement of interference between work and personal life, Koekemoer, Mostert and Rothmann (2010) developed a new instrument that measures the interference between work and various nonwork roles (i.e. parental, spousal, religion/spirituality and domestic roles) in both directions (i.e. work→nonwork interference (W-NWI) and nonwork→work interference (NW-WI). Although this study provided a valuable new instrument that addressed several theoretical and measurement limitations, no evidence was provided regarding the psychometric properties of the instrument. This validation study attempts to demonstrate the empirical distinctiveness of the new instrument by examining its internal validity (i.e. construct, discriminant and convergent validity and reliability) and external validity (i.e. the relationship with job and home characteristics and well-being).

Background to the study

One specific aspect of significant importance that has been voiced by previous researchers is the measurement of interaction between work and personal life (Bellavia & Frone, 2005; Frone, 2003; Tetrick & Buffardi, 2006). Two fundamental issues are mentioned. The first issue pertains to the use of instruments that fail to measure the involvement of individuals in multiple social roles and the possible conflict between these multiple roles (currently, interference of work with family is predominantly measured, thereby neglecting valid measurement of work interference with other roles in private life). The second issue concerns the lack of thorough reporting of psychometric properties regarding the instruments being used to measure interaction between work and personal life (Parasuraman & Greenhaus, 2002; Robinson, Shaver & Wrightsman, 1991; Schultheiss, 2006; Voydanoff, 2007) (for an overview of other issues, see Koekemoer et al., 2010).

Understanding the interference between work and different nonwork roles is also problematic for South African researchers, as there is no valid and reliable instrument that measures the interference between work and nonwork roles in the South African context. Currently, the main instrument that is used in South African studies is the internationally developed ‘survey work-home interaction – Nijmegen’ (SWING) of Geurts et al., 2005. Although the SWING has been researched well and is psychometrically sound (see Marais, Mostert, Geurts & Taris, 2009; Mostert & Oldfield, 2009; Pieterse & Mostert, 2005; Rost & Mostert, 2007), it is not without limitations. Several of the SWING items confound with external variables, causes and consequences. The items also specifically measure time-based and strain-based interference. These are limitations raised as concerns by previous researchers (see Bellavia & Frone, 2005; Frone, 2003). Also, the SWING only measures interference between two dimensions (i.e. work and home) and fails to measure interference between work and other dimensions in the nonwork domain. This limitation is of particular importance in South Africa. In an exploratory study, Koekemoer and Mostert (2010) found that South African employees experience interference between work and various other nonwork dimensions in their lives (i.e. work interferes with family, leisure, domestic activities, exercise, community activities and the extended family). It was also reported that involvement in several nonwork roles interfered with the work role.

To address these limitations, Koekemoer et al. (2010) developed an instrument that measures interference of work with four nonwork roles (viz. work-parent interference (WPI), work-spouse interference (WSI), work-religion/spirituality interference (WRI) and work-domestic interference (WDI)) and interference of four nonwork roles with the work role (viz. parent-work interference (PWI), spouse-work interference (SWI), religion/spirituality-work interference (RWI) and domestic-work interference (DWI)). Close attention was paid to the scale development procedures described in the psychometric literature (DeVellis, 2003; Robinson et al., 1991) and they were adhered to and various of the theoretical and measurement limitations were addressed. Rasch analyses and item correlations were used. Items were evaluated and eliminated in order to retain the best items. A total of 30 items were retained. In the final instrument, W-NWI was measured with 15 items: four WPI items, four WSI items, four WRI items and three WDI items. NW-WI was also measured with 15 items: four PWI items, four SWI items, four RWI items and three DWI items. All the items were rated on a four-point Likert scale ranging from 0 (‘never’) to 3 (‘always’).

Trends from the research literature

Psychometric properties of work-family instruments
Although a variety of instruments is found widely across international work-family conflict studies (for summaries of work-family conflict (WFC) and family-work conflict (FWC) studies, see Allen et al., 2000; Byron, 2005; Mesmer-Magnus & Viswesvaran, 2005), the use thereof may be problematic when researchers compare results without being aware of the psychometric properties (i.e. the validity and reliability) of the instruments (DeVellis, 2003; Robinson et al., 1991). Although a variety of indicators of psychometric properties of instruments measuring work-family interaction are found in the literature (i.e. internal validity, construct validity, discriminant validity, convergent validity, reliability and external validity, relationships with external variables), the reporting and use thereof are inconsistent across studies.

One of the main psychometric properties being reported across work-family scale development studies is the construct validity of instruments. However, there is not always consistency regarding the analyses employed during factor analysis (e.g. exploratory (EFA) vs. confirmatory factor analysis (CFA)). Although the use of exploratory factor analysis during the development of new instruments is highly recommended in the literature on psychometrics (DeVellis, 2003; Nunnally & Bernstein, 1994), few work-family studies have made use of exploratory factor analysis (Aryee, 1992; Grzywacz & Marks, 2000; Kopelman, Greenhaus & Connolly, 1983; Premeaux et al., 2007). Instead, the majority of work-family studies used confirmatory factor analysis, which has become a very popular data-analytic technique for the clinical and social sciences (Carlson & Frone, 2003; Curbow, McDonell, Spratt, Griffen & Agnew, 2003; Geurts et al., 2005).

An additional indicator of validity that is seldom reported or tested in work-family studies is discriminant validity. According to Small and Riley (1990), discriminant validity is the extent to which measures of different dimensions are indexing different factors (i.e. when factors do not correlate highly). Discriminant validity is used specifically to indicate clear distinctiveness between constructs in order to further provide proof of empirically distinct dimensions in instruments (Carlson & Frone, 2003). In the literature, few work-family studies tested or provided evidence for discriminant validity (Carlson & Frone, 2003; Carlson, Kacmar & Williams, 2000; Grzywacz & Marks, 2000; Small & Riley, 1990). In the majority of these studies, correlations, factor correlations from CFA or different correlates with outcomes were used as evidence for discriminant validity. Only Netemeyer, Boles and McMurrian (1996) tested for discriminant validity using the chi square (χ²) difference test. Although the use of the chi square (χ²) difference test is not well known among work-family studies, it has been recommended by previous researchers (Anderson & Gerbing, 1988; Deery, Erwin & Iverson, 1999).

Similar to discriminant validity, convergent validity is also not used widely in work-family studies. Of the main work-family studies, only Small and Riley (1990) tested convergent validity using correlations with similar instruments. Although not always found within the work-family field of scale development, the use of correlations with other instruments as evidence for convergent validity is widely found or tested in other scale development studies (Cowin et al., 2008; Taormina, 2004; Trout, Ryan, Vigne & Epstein, 2003). In contrast to the inconsistency in the use and reporting of validity indicators, reliability measures (i.e. Cronbach’s alpha coefficients) are found in the majority of work-family studies.

In terms of the relationship of work-family instruments with external variables (i.e. antecedents and outcomes), the majority of studies report or indicate these relationships. Some studies use correlations to indicate these relationships (Carlson & Frone, 2003; Geurts et al., 2005; Kopelman et al., 1983; Mallard & Lance, 1998; Netemeyer et al., 1996), whilst others use multiple regression analyses (Curbow et al., 2003; Grzywacz & Marks, 2000; Premeaux et al., 2007), or combine correlation analyses and multiple regression analyses (Aryee, 1992; Kirchmeyer, 1992). Some studies do not report the relationship of a work-family instrument with external variables (Small & Riley, 1990; Stephens & Sommer, 1996).

Relations with causes and consequences
In the literature there are various overview studies that summarise the relationship between WFC and FWC and various external variables, including a variety of antecedents and consequences (e.g. Geurts & Demerouti, 2003; Guest, 2002; Voydanoff, 2005a). Some of the more recent studies indicate specifically the relationship of WFC with antecedents such as work demands (i.e. time demands and strain demands), job schedule demands, emotional demands, quantitative demands, work pressures and work support (i.e. managerial support), role autonomy and role overload (Boyar, Carr, Mosley & Carson, 2007; Choi, 2008; Fu & Shaffer, 2001; Greenhaus & Powell, 2003; Lingard & Francis, 2005; Montgomery, Panagopolou & Benos, 2006; Nasurdin & Hsia, 2008; Voydanoff, 2005b). FWC has been found to be related to antecedents such as family demands (e.g. household demands), autonomy, work and family pressure and parental demands (Boyar et al., 2007; Fu & Shaffer, 2001; Greenhaus & Powell, 2003; Keene & Reynolds, 2005; Voydanoff, 2005b). Consequences related to WFC include emotional exhaustion, depersonalisation, decrease in quality of life and quality of family life, impaired career outcomes, depression, decrease in job satisfaction, decrease in organisational commitment and decrease in life stress (Aryee, Srinivas & Tan, 2005; Choi, 2008; Greenhaus, Collins & Shaw, 2002; Lingard & Francis, 2005; McLean, 2002; Montgomery et al., 2006; Nikandrou, Panayotopoulou & Apospori, 2008; Weigel, Weigel, Berger, Cook & DelCampo, 1995).

Potential value added by the study and research objectives
It is very important to use instruments that are psychometrically sound and for which there is well-established evidence of validity and reliability − proper psychometric reporting holds serious implications for the validity of relationships with other variables (DeVellis, 2003; Robinson et al., 1991). This study seeks to adhere to the call to provide sound psychometric evidence for a new instrument by providing preliminary information on its internal validity (construct, discriminant and convergent validity and reliability). In addition, for the investigation and reporting of the external validity of instruments, it is important to indicate relationships between work-family constructs and various external variables (see Carlson & Frone, 2003; Carlson et al., 2000; Geurts et al., 2005). This study will therefore also focus on the external validity of the new instrument and examine its relationship with other theoretically relevant variables.

More specifically, the objectives of this study were to investigate, in relation to the instrument, its:

• construct validity

• discriminant and convergent validity

• reliability

• relationship with theoretically relevant external variables (including job characteristics, home characteristics, burnout, ill health and life satisfaction).

What will follow
In the next section, the following hypotheses will be tested to reach the specific objectives:

Objective 1 (construct validity)
Hypothesis 1a: Work-nonwork interference is a four-dimensional construct, consisting of work-parent interference, work-spouse interference, work-religion/spirituality interference and work-domestic interference.

Hypothesis 1b: Nonwork-work interference is a four-dimensional construct, consisting of parent-work interference, spouse-work interference, religion/spirituality-work interference and domestic-work interference.

Hypothesis 2: A second-order, two-factor model that distinguishes between the different directions is superior to a one-factor model that assumes no directionality.

Objective 2 (discriminant and convergent validity)
Hypothesis 3: The four work-nonwork dimensions and four nonwork-work dimensions are, although highly related, empirically distinct constructs (i.e. there is evidence of discriminant validity).

Hypothesis 4: The work-nonwork dimensions correlate strongly with the negative work-home interference (WHI) scale of the SWING (i.e. evidence of convergent validity).

Hypothesis 5: The nonwork-work dimensions correlate strongly with the negative home-work interference (HWI) scale of the SWING (i.e. evidence of convergent validity).

Objective 3 (reliability)
Hypothesis 6: All the dimensions (i.e. work-parent interference, work-spouse interference, work-religion and/or spirituality interference, work-domestic interference, parent-work interference, spouse-work interference, religion and/or spirituality-work interference and domestic-work interference) are reliable (i.e. Cronbach’s alpha coefficients ≥ 0,70).

Objective 4 (external validity)
Hypothesis 7: The different work-nonwork dimensions are related in unique ways to job characteristics and well-being.

Hypothesis 8: The different nonwork-work dimensions are related in unique ways to home characteristics and well-being.

Research approach

Research method

Although the participants were mainly white (80.35%), participants from the African (14.75%), Indian (3.00%) and coloured (0.80%) groups were also included in the sample. Males (34.70%) as well as females (65.00%) were included in the study. The majority of participants had postgraduate degrees (47.81%), while other participants had university degrees (12.57%), technical college diplomas (6.00%), technikon diplomas (8.20%) or grade 12 certificates (19.95%). In total, 26.77% of the participants worked as administrative assistants, while 25.68% worked in the administrative offices. The majority of participants worked in academic faculties, including the faculty of health sciences (13.39%), natural sciences (11.46%), education (10.38%), engineering (9.58%), arts (6.83%), economic and management sciences (6.56%) and theology (2.70%). A number of participants worked as lecturers (9.84%), senior lecturers (11.46%), associate professors (6.56%) and professors (7.10%).

Measuring instruments
Items of the job and home characteristics measures were randomly combined with items of the W-NWI and W-NWHI scales (i.e. job-related items combined with W-NWHI items and home-related items combined with NW-WI items) in order to ensure more accurate and valid responses from the participants. In each instance, the questions to the participants were phrased: ‘How often does it happen that...’.

The following measures were utilised in this study:

• job demands

• job resources

• home demands

• home resources

• burnout

• ill health

• life satisfaction.

Job demands: Work pressure was measured with a three-item scale developed by Bakker, Demerouti and Schaufeli (2003a). The items referred to demanding quantitative aspects of the job. An example is: ‘You have to work extra hard in order to meet your deadlines’. Emotional work demands were measured with the five-item scale of Bakker, Demerouti and Schaufeli (2003b). An example is: ‘People at work upset you emotionally with their words’. Cognitive demands were measured by the four-item scale developed by Peeters, Montgomery, Bakker and Schaufeli (2005). An example is: ‘Your work requires you to concentrate continuously’. All these job demands items were rated on a four-point Likert scale ranging from 0 ('never') to 3 ('always'). Reliable Cronbach’s alpha coefficients larger than 0.70 (Nunnally & Bernstein, 1994) were found in these studies (Bakker et al., 2003b; Peeters et al., 2005): work pressure (α = 0.76); emotional work demands (α = 0.74) and cognitive work demands (α = 0.89).

Job resources: Work autonomy was measured with the scale developed by Bakker, Demerouti and Verbreke (2004) (three items, e.g. ‘You have freedom in carrying out your work-related duties’). Work support was measured with the scale developed by Bakker et al. (2004) (three items, e.g. ‘You ask your colleagues for help if necessary.’). Work development possibilities were assessed by three items that were conceptually mirrored from existing scales of home developmental possibilities developed by Demerouti, Bakker and Voydanoff (2010). An example item is: ‘Can you develop yourself sufficiently in your work?’ All these job resources items were rated on a four-point Likert scale ranging from 0 ('never') to 3 ('always'). Cronbach’s alpha coefficients were found that ranged between 0.68 and 0.74 for work autonomy and between 0.81 and 0.85 for work support (Bakker et al., 2004; Bakker, Demerouti & Euwema, 2005).

Home demands: Home pressure was measured with a three-item scale developed by Peeters et al. (2005). The items referred to demanding quantitative aspects of the home. An example is: ‘Do you find that you are busy at home?’ Emotional home demands were also measured with a three-item scale developed by Peeters et al. The scale assessed whether the participants had to deal with emotionally charged situations at home and whether they were confronted by events that touched them personally. A sample item is: ‘How often do emotional issues arise at home?’ All of these home demands items were rated on a four-point Likert scale ranging from 0 ('never') to 3 ('always'). Peeters et al. (2005) reported acceptable Cronbach’s alpha coefficients of 0.80 for home pressure and 0.76 for emotional home demands.

Home resources: The home resources scale was developed by Demerouti et al. (2010) and conceptually mirror existing scales of job resources, as several scholars have successfully used a job-related measure as a model for constructing a symmetrical home-related measure (Frone & Rice, 1987; Frone, Russell & Cooper, 1992; Parasuraman, Purohit, Godshalk & Beutell, 1996). Home autonomy was assessed with four items, including ‘You have control over how you use your free time’. Home support was measured with four items, including ‘If necessary, your partner or family members will help you with a particular task’. Home developmental possibilities were assessed by three items, including ‘You can develop your talents during your free time’. All of these home resources items were rated on a four-point Likert scale ranging from 0 ('never') to 3 ('always'). Acceptable alpha coefficients were found by Demerouti et al. (2010).

Burnout: Emotional exhaustion was measured using eight items (e.g. ‘I feel emotionally drained by my work’) and depersonalisation was measured by five items (e.g. ‘I do not really care what happens to some recipients’) from the MBI-HSS (Maslach & Jackson, 1986). The items were rated on a seven-point scale ranging from 0 ('never') to 6 ('every day'). Cognitive weariness was measured using the five items of Van Horn, Taris, Schaufeli and Schreurs (2004) (e.g. ‘I have trouble concentrating’). The items were rated on a seven-point scale ranging from 0 ('a few times a year') to 6 ('every day'). Van Horn et al. (2004) found the scale to be reliable, with an alpha coefficient of 0.92.

Ill health: Items were adapted from the general health questionnaire (GHQ; Goldberg & Williams, 1988) to measure physical ill health (four items, e.g. ‘Have you recently been getting any headaches?’), anxiety (five items, e.g. ‘Have you recently been getting edgy and bad-tempered?’) and depression (four items, e.g. ‘Have you recently felt that life is entirely hopeless?’). Items were rated on a four-point scale ranging from 0 ('not at all') to 3 ('much more than usual'). Reliable alpha coefficients were found, ranging between 0.83 and 0.81 for physical ill health, between 0.84 and 0.89 for anxiety and between 0.79 and 0.89 for depression (Mostert, 2009; Oldfield & Mostert, 2007).

Life satisfaction: The items used were from the satisfaction with life scale (SWLS; Diener, Emmons, Larson & Griffin, 1985) to measure life satisfaction (five items, e.g. ‘I am satisfied with my life’; ‘The conditions of my life are excellent’). Items were rated on a seven-point scale, ranging from 1 ('strongly disagree') to 7 ('strongly agree'). Diener et al. (1985) found the scale to be reliable and valid, with an alpha coefficient of 0.87 and test-retest reliability of 0.82.

The construct validity of these measuring instruments was tested using structural equation modelling. The results supported a six-factor model for job characteristics (work pressure, emotional work demands, cognitive work demands, work autonomy, work support and work developmental possibilities: χ² = 398.98 _(N = 366), IFI = 0.92, TLI = 0.91 and CFI = 0.93, RMSEA = 0.06); a five-factor model for home characteristics (home pressure, emotional home demands, home autonomy, home support and home developmental possibilities: χ² = 161.63 _(N = 366), IFI = 0.97, TLI = 0.96 and CFI = 0.97; RMSEA = 0.04); a three-factor model for burnout (emotional exhaustion, depersonalisation and cognitive weariness: χ² = 422.80 _(N = 366), IFI = 0.92, TLI = 0.90 and CFI = 0.92; RMSEA = 0.08); a three-factor model for health (physical health, anxiety and depression: χ² = 167.49 _(N = 366), IFI = 0.97, TLI = 0.96 and CFI = 0.97; RMSEA = 0.07) and a one-factor model for life satisfaction (χ² = 47.20 _(N = 366), IFI = 0.97, TLI = 0.93 and CFI = 0.97; RMSEA = 0.15).

Research procedure
After permission was granted by the university’s ethics committee, lists of all married employees with children were obtained from the various faculties and departments of the university. With the help of field workers, questionnaires were distributed personally among the selected employees and the participants were given various options for returning the questionnaires to the researchers (e.g. internal post, personal collection, personal delivery). These options were given to the participants in order to ensure a better response rate and also to enhance anonymity and confidentiality. Questionnaire booklets were handed out during the data collection. These included a section that explained the research objectives and the process. In addition, field workers also verbally communicated to the potential participants what the goal and process of the research project were. The participants were informed that they were not obliged to participate in the research. They were also informed that, if they did participate in the research and fill in a questionnaire, they were giving consent for the collected data to be used for research purposes. No personal information was or will be made available to any other organisations or persons. The participants were given two to three weeks to complete the questionnaires. The questionnaires also contained a letter that explained the goal and importance of the study and that gave them assurances regarding the anonymity and confidentiality with which the information will be handled.

Statistical analysis
In order to examine the construct validity of the newly developed instrument, confirmatory factor analysis (CFA) was used with Amos structural modelling software (Arbuckle, 2007). Although exploratory factor analysis (EFA) is usually used for analysing newly developed instruments, using a confirmatory approach is known to be very meaningful, as researchers are required to specify the number of factors according to the literature and substantive theoretical knowledge, which result in the testing of the adequacy of fit of the theoretical factor model (Bollen, 1989). CFA is a theory-driven comprehensive statistical approach for testing theory-based hypotheses that has a number of strengths. It is a very popular data-analytic technique for the clinical and social sciences and therefore the use of CFA is very appropriate for this study (Tomarken & Waller, 2005). With CFA, researchers have the ability to specify latent variable models that provide separate estimates of relations among latent constructs and their manifest indicators (the measurement model). Also, the availability of measures of global fit that can provide a summary evaluation of even complex models is known as a key strength of CFA. In addition, users can comparatively evaluate the fit of alternative models that differ in complexity via nested chi-square tests and other means.

The goodness of fit of the models was evaluated using the following absolute goodness-of-fit indices (Jöreskog & Sörbom, 1993): the likelihood-ratio chi square (χ²), the ratio of the chi square to the degrees of freedom (χ²/df) and the root square of approximation (RMSEA). As c² is sensitive to sample size – that is the probability of rejecting a hypothesised model increases with sample size – the use of relative goodness-of-fit measures is strongly recommended (Bentler, 1990). Therefore, the Comparative Fit Index (CFI), Tucker-Lewis Index (TLI) and Incremental Fit Index (IFI) were used. Values smaller than 0.08 for RMSEA are indicative of an acceptable fit, while values greater than 0.10 should lead to model rejection (Cudeck & Brown, 1993). For CFI, TLI and IFI, as a rule of thumb a value greater than 0.90 is considered a good fit (Hoyle, 1995) while a value smaller than 5.00 is acceptable for the c² /df statistic (Bentler & Bonett, 1980).

Because of the large number of items, it was not possible to conduct CFA analysis on a full disaggregation model. Therefore, due to the complexity, extensiveness and the conceptual directional distinction of interference, two theoretical models were tested:

• A four-factor model for W-NWI interference (consisting of work-parent interference, work-spouse interference, work-religion/spirituality interference and work-domestic interference).

• A four-factor model for NW-WI (consisting of parent-work interference, spouse-work interference, religion/spirituality-work interference and domestic-work interference).

The items identified by Koekemoer et al. (2010) were used as indicators of the latent factors. Because the factors are theoretically so closely related, errors were allowed to correlate. Using alternative models (Lehmann, 2001), the two hypothesised four-factor models were compared with several competing models. Similar models were tested separately for the two directions of interference (i.e. five alternative models for W-NWI and five alternative models for NW-WI). These competing models were similar to the models used in previous scales (Carlson & Frone, 2003; Carlson & Kacmar, 2000; Carlson et al., 2000; Curbow et al., 2003; Geurts et al., 2005; Netemeyer et al., 1996) and models based on theoretical knowledge.

In order to illustrate the directionality of the W-NWI instrument, an additional, second-order factor model was tested with CFA. Firstly, a one-factor model was tested. This model assumes that all the W-NWI and NW-WI dimensions load onto one factor (i.e. directionality is not distinguished). Secondly, a two-factor model was tested. This model assumes that W-NWI and NW-WI are distinct constructs, each measured with the four W-NWI and NW-WI dimensions as indicators (i.e. distinguishing between the direction of interference).

CFA analyses were also used to prove the discriminant validity of the various dimensions. Following the example of previous researchers (Anderson & Gerbing, 1988; Bagozzi & Philips, 1982; Mallard & Lance, 1998), discriminant validity was tested by making use of the chi square (χ²) difference test. By calculating the difference between a model that allows the correlation between the constructs (with multiple indicators) to be constrained to unity (i.e. perfectly correlated) and another model that allows the correlations to be free (unconstrained model or target model), discriminant validity can be tested. If the two models do not differ significantly on a chi-square difference test, the researcher fails to conclude that the constructs differ. However, a significantly lower chi-square value for the model in which the trait correlations are not constrained to unity would indicate that the traits are not perfectly correlated and that discriminant validity is achieved (Bagozzi & Philips, 1982; Bagozzi, Yi & Phillips, 1991). According to Anderson and Gerbing (1988), discriminant validity can be assessed for two estimated constructs by constraining the estimated correlation parameter between them to 1.0 and then performing a chi-square difference test on the values obtained for the constrained and unconstrained models. This procedure was carried out for one pair of constructs at a time, thereby calculating the χ² difference test for all the various correlations between different dimensions.

Convergent validity was assessed by investigating the correlation coefficients between the various dimensions of the W-NWI scale and the SWING instrument (Geurts et al., 2005) using the SPSS program (SPSS Inc., 2009). The SWING was selected because it is a widely used, well-researched and psychometrically sound instrument that measures the interference in both directions (work→home and home→work). The SWING has also been validated in different South African samples (Marais et al., 2009; Mostert & Oldfield, 2009; Pieterse & Mostert, 2005; Rost & Mostert, 2007). Although Polit and Beck (2006) recommend that correlations greater than 0.70 may be regarded as evidence for strong correlations (or similarities) between the measures in social research, other researchers have indicated that correlation coefficients should only meet or exceed 0.35 in order to be cited as evidence for convergent validity (Hammill, Brown & Bryant, 1989). Cronbach’s alpha coefficients were used to assess the reliability of the newly developed scales. Descriptive statistics (means, standard deviations) were used to describe the data. To determine the relationship between the scales and various external variables, Pearson product-momentum correlation coefficients were used. The level of statistical significance was set at p < 0.05. Effect sizes (Cohen, 1988; Steyn, 2002) were used in addition to statistical significance to determine the practical significance of relationships. Effect sizes indicate whether the obtained results are important, while statistical significance may often show results that are of little practical relevance (Steyn, 2002). A cut-off point of 0.30 (medium effect) (Cohen, 1988) was set for the practical significance of the correlation coefficients.

Construct validity of the work-nonwork interference instrument

Firstly, a one-factor model (M2) that assumes that all 15 items measuring the interference from work to nonwork roles load onto the same underlying latent variable (i.e. W-NWI) was compared to the theoretical four-factor model (M1). This alternative model assumes that items cannot be distinguished on the basis of the specific role being influenced by the work domain (see Tetrick & Buffardi (2006) for a discussion of the various factor structures in work-family research). With all 15 items loading onto one factor, the fit of M2 was significantly worse when compared to M1 (Δ χ² (19) = 566.73, p ≤ 0.001).

Next, a two-factor model (M3) was compared to the theoretical model (M1). This model distinguishes between two latent variables (i.e. work-home interference and work-religion/spirituality interference) and is based on previous work-family literature and positive psychology literature, where no distinction is made between home-related roles (such as parent, spouse or domestic roles) and religion/spirituality is seen as an additional well-being dimension (Snyder & Lopez, 2002). When compared to the hypothesised model (M1), the fit of this alternative model was also significantly worse (Δ χ² (18) = 307.25, p ≤ 0,001).

Two three-factor models (M4 and M5) were also compared to the theoretical model (M1). Model 4 distinguishes between the interference of domestic-related activities and family relationship interference (therefore consisting of three latent variables, i.e. work-family interference, work-domestic interference and work-religion/spirituality interference). Model 5, however, differs from Model 4 in the sense that the assumption is made that the parental items and domestic items load together to form a caretaker latent variable, in addition to the spouse latent variable and the religion/spirituality latent variable. Model 5 therefore assumes that parental activities are closely related to domestic activities and are different from the relationship with one’s spouse. The results indicated that the fit of both of these models (M4 and M5) was also significantly worse than that of the theoretical model (M1) (for M4 = Δ χ² (16) = 209.82, p ≤ 0.001 and M5 = Δ χ² (16) = 179.27, p ≤ 0.001). Finally, all of the fit indices suggested that the theoretical four-factor model (M1) for W-NWI was a better fit for the data, with all the values (i.e. IFI, TLI, CFI) far exceeding the conventional standard of 0.90 (Hoyle, 1995).

Similar procedures were employed for the NW-WI scale, with a four-factor theoretical model of NW-WI being compared to alternative models similar to the W-NWI models (Hypothesis 1b). Firstly, a four-factor hypothesised model was tested for the NW-WI scale. This model also assumed that the interference between nonwork roles and the work domain would differ depending on the specific roles and that it would result in four expected interference dimensions (i.e. parent-work interference, spouse-work interference, religion/spirituality-work interference and domestic-work interference). The results showed acceptable fit for this hypothesised four-factor model (see Table 1 – M1 – four-factor theoretical model). All the fit indices suggested that this theoretical four-factor model (M1) for NW-WI fitted the data well, with all the values (IFI, TLI, CFI) far exceeding the conventional standard of 0,90 (Hoyle, 1995). Consequently, this model was used as the baseline model in subsequent analyses regarding NW-WI. Next, various alternative factor models were systematically compared to this four-factor NW-WI theoretical model (M1) (the results are also summarised in Table 1).

Firstly, a one-factor model (M2) was compared to the theoretical four-factor model (M1). For this alternative model (where all 15 items loaded onto one factor), the fit of M2 was significantly worse compared to M1 (Δ χ² (6) = 164.42, p ≤ 0.001). Next, a two-factor model (M3) was compared to the theoretical model (M1). Upon comparison with the hypothesised model (M1), the fit of this alternative model (M3) was again significantly worse (Δ χ² (5) = 105.63, p ≤ 0.001). In addition, two three-factor models (M4 and M5) were compared to the theoretical model (M1). The results indicated that the fit of both of these models (M4 and M5) was also significantly worse than that of the theoretical model (M1) (for M4 = Δ χ² (3) = 91.55, p ≤ 0.001 and M5 = Δ χ² (3) = 21.04, p ≤ 0.001). Although M5 showed a fit closest to M1, the fit indices suggested that the theoretical four-factor model (M1) for NW-WI was still a better fit of the data, with all the values (i.e. IFI, TLI, CFI) far exceeding the conventional standard of 0.90 (Hoyle, 1995).

These results provide support for Hypotheses 1a and 1b.

Consequently, the two second-order factor models were tested (i.e. the one-factor model and the two-factor model, Hypothesis 2). The results of the CFA analyses showed that the one-factor model indicated poor fit to the data (χ² = 568.02 _(N = 366), IFI = 0.66, TLI = 0.52 and CFI = 0.66; RMSEA = 0.27), whereas the two-factor model showed a good fit to the data (χ² = 136.58 _(N = 366), IFI = 0.93, TLI = 0.90 and CFI = 0.93; RMSEA = 0.13). Upon comparison of the one-factor and two-factor models, it seemed that the two-factor model fitted the data significantly better than the one-factor model (Δ χ² (1) = 431.44, p ≤ 0.001), which provides support for Hypothesis 2.

Discriminant validity

Allen, T.D., Herst, D.E.L., Bruck, C.S., & Sutton, M. (2000). Consequences associated with work-to-family conflict: A review and agenda for future research. Journal of Occupational Health Psychology, 5(2), 278–308.

Anderson, J.C., & Gerbing, D.W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 3, 411–423.

Arbuckle, J.L. (2007). Amos 17.0.0. Amos Development Company. Copyright 1983–2008. NET Framework Version 1.1.

Aryee, S. (1992). Antecedents and outcome of work-family conflict among married professional women: Evidence from Singapore. Human Relations, 45, 813–837.

Aryee, S., Srinivas, E.S., & Tan, H.H. (2005). Rhythms of life: Antecedents and outcomes of work-family balance in employed parents. Journal of Applied Psychology, 90, 132–146.

Bagozzi, R.P., & Phillips, L.W. (1982). Representing and testing organizational theories: A holistic construal. Administrative Science Quarterly, 27, 449–489.

Bagozzi, R.P., Yi, P.Y., & Phillips, L.W. (1991). Assessing construct validity in organizational research. Administrative Science Quarterly, 36, 421–458.

Bailyn, L., & Harrington, M. (2004). Redesigning work for work-family integration. Community, Work & Family, 7(2), 197–208.

Bakker, A.B., Demerouti, E., & Euwema, M.C. (2005). Job resources buffer the impact of job demands on burnout. Journal of Occupational Health Psychology, 10(2), 170–180.

Bakker, A.B., Demerouti, E., & Schaufeli, W.B. (2003a). The socially induced burnout model. In A.M. Columbus (Ed.), Leading edge research in cognitive psychology, (pp.46-62). New York: Nova Science Publishers.

Bakker, A.B., Demerouti, E., & Schaufeli, W.B. (2003b). Dual processes at work in a call centre: An application of the Job Demands-Resources Model. European Journal of Work and Organizational Psychology, 12, 393–417.

Bakker, A.B., Demerouti, E., & Verbeke, W. (2004). Using the job demands-resources model to predict burnout and performance. Human Resource Management, 43, 83–104.

Bellavia, G.M., & Frone, M.R. (2005). Work-family conflict. In E.K. Barling, K. Kelloway & M.R. Frone (Eds.), Handbook of work stress (pp. 113–148). Thousand Oaks: Sage.

Bentler, P.M., & Bonett, D. (1980). Significance tests and goodnesses of fit in the analysis of covariance structures. Psychological Bulletin, 88, 588–606.

Bentler, P.M. (1990). Comparative fit indexes in structural equation models. Psychological Bulletin, 107, 238–246.

Bollen, K.A. (1989). Structural equations with latent variables. New York: Wiley.

Boyar, S.L., Carr, J.C., Mosley, D.C., & Carson, C.M. (2007). The development and validation of perceived work and family demand scales. Educational and Psychological Measurement, 67, 100–115.

Bryman, A., & Bell, E. (2003). Business research methods. (2nd edn.). New York: Oxford University Press.

Bulger, C.A., Matthews, R.A., & Hoffman, M.E. (2007). Work and personal life boundary management: Boundary strength, work/personal life balance, and the segmentation-integration continuum. Journal of Occupational Health Psychology, 12(4), 365–375.

Byron, K. (2005). A meta-analytic review of work-family conflict and its antecedents. Journal of Vocational Behavior, 67, 169–198.

Carlson, D.S., & Frone, M.R. (2003). Relation of behavioral and psychological involvement to a new four-factor conceptualization of work-family interference. Journal of Business and Psychology, 17(4), 515–535.

Carlson, D.S., & Kacmar, K.M. (2000). Work-family conflict in the organization: Do life role values make a difference? Journal of Management, 26, 1031–1054.

Carlson, D.S., Kacmar, K.M., & Williams, L.J. (2000). Construction and initial validation of a multidimensional measure of work-family conflict. Journal of Vocational Behavior, 56, 249–276.

Choi, J. (2008). Work and family demands and life stress among Chinese employees: The mediating effect of work-family conflict. The International Journal of Human Resource Management, 19(5), 878–895.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences. (rev. edn.). Orlando: Academic Press.

Cowin, L.S., Hengstberger, C., Eagar, S.C., Gregory, L., Andrew, S., & Rolley, J. (2008). Competency measurements: Testing convergent validity for two measures. Journal of Advanced Nursing, 64(3), 272–277.

Cudeck, R., & Browne, M.W. (1993). Alternative ways of assessing model fit. In K.A. Bollen & J.S. Long (Eds.), Testing structural equation models (pp. 1–9). Newbury Park: Sage.

Curbow, B., McDonell, K., Spratt, K., Griffen, J., & Agnew, J. (2003). Development of the work-family interface scale. Early Childhood Research Quarterly, 18, 310–330.

Day, A.L., & Chamberlain, T.C. (2006). Committing to your work, spouse and children: Implications for work-family conflict. Journal of Vocational Behavior, 68, 116–130.

Deery, S., Erwin, P., & Iverson, R. (1999). Industrial relations climate, attendance behaviour, and the role of trade unions. British Journal of Industrial Relations 37(4), 533–558.

Demerouti, E., Bakker, A.B., & Voydanoff, P. (2010). Does home life interfere with or facilitate job performance? European Journal of Work and Organizational Psychology, 19(2), 128–149.

DeVellis, R.F. (2003). Scale development, theory and application. (2nd edn.). Thousand Oaks: Sage.

Diener, E.D., Emmons, R.A., Larson, R.J., & Griffin, S. (1985). The satisfaction with life scale. Journal of Personality Assessment, 49, 71–75.

Doumas, D.M., Margolin, G., & John, R.S. (2008). Spillover patterns in single-earner couples: Work, self-care, and the marital relationship. Journal of Family Economic Issues, 29, 55–73.

Frone, M.R. (2003). Work-family balance. In J.C. Quick & L.E. Tetrick (Eds.), The handbook of occupational health psychology (pp. 143–162). Washington: American Psychological Association.

Frone, M.R., & Rice, R.W. (1987). Work-family conflict: The effect of job and family involvement. Journal of Occupational Behavior, 11, 79–96.

Frone, M.R., Russell, M., & Cooper, M.L. (1992). Antecedents and outcomes of work–family conflict: Testing a model of the work–family interface. Journal of Applied Psychology, 77, 65–78.

Fu, C.K., & Shaffer, M.A. (2001). The tug of work and family. Direct and indirect domain-specific determinants of work-family conflict. Personnel review, 30(5), 502–522.

Geurts, S.A.E., & Demerouti, E. (2003). Work/non-work interface:

A review of theories and findings. In M.J. Schabracq, J.A.M. Winnubst & C.L. Cooper (Eds.), The handbook of work and health psychology (pp. 280–312). Chichester: Wiley.

Geurts, S.A.E., Taris, T.W., Kompier, M.A.W., Dikkers, J.S.E., Van Hooff, L.M., & Kinnunen, U.M. (2005). Work-home interaction from a work psychological perspective: Development and validation of a new questionnaire, the SWING. Work and Stress, 19(4), 319–339.

Glaser, K., Evandrou, M., & Tomassini, C. (2006). Multiple role occupancy and social participation among midlife wives and husbands in the United Kingdom. International Journal of Aging and Development, 63, 27–47.

Goldberg, D., & Williams, P. (1988). A user’s guide to the General Health Questionnaire. Windsor: NFER-Nelson.

Greenhaus, J.H., Collins, K.M., & Shaw, J.D. (2003). The relation between work-family balance and quality of life. Journal of Vocational Behavior, 63, 510–531.

Greenhaus, J.H., & Powell, G.N. (2003). When work and family collide: Deciding between competing role demands. Organizational Behavior and Human Decision Processes, 90, 291–303.

Grzywacz, J.G., & Marks, N.F. (2000). Reconceptualizing the work-family interface: An ecological perspective on the correlates of positive and negative spillover between work and family. Journal of Occupational Health Psychology, 5(1), 111–126.

Guest, D.E. (2002). Perspectives on the study of work-life balance. Social Science Information, 41(2), 255–279.

Hammill, D.D., Brown, L., & Bryant, R.R. (1989). A consumers’ guide to test in print. Austin, TX: PRO-ED.

Heraty, N., Morley, M.J., & Cleveland, J.N. (2008). Complexities and challenges in the work-family interface. Journal of Managerial Psychology, 23(3), 209–214.

Hoyle, R.H. (1995). The structural equation modeling approach: Basic concepts and fundamental issues. In R.H. Hoyle (Ed.), Structural equation modeling: Concepts, issues, and applications (pp. 1–15). Thousand Oaks: Sage.

Jöreskog, K.G., & Sörbom, D. (1993). LISREL 8: User’s Reference Guide. Chicago: Scientific Software International.

Keene, J.R., & Reynolds, J.R. (2005). The job costs of family demands, gender differences in negative family-to-work spillover. Journal of Family Issues, 26(3), 275–299.

Kirchmeyer, C. (1992). Perceptions of non-work to work spillover: Challenging the common view of conflict-ridden domain relationships. Basic and Applied Social Psychology, 13(2), 231–249.

Kline, P. (1999). The handbook of psychological testing. (2nd edn.). London: Routledge.

Koekemoer, E., & Mostert, K. (2010). An exploratory study of the interaction between work and personal life: Experiences of South African employees. SA Journal of Industrial Psychology/SA Tydskrif vir Bedryfsielkunde, 36(1), Art. #801, 15 pages. DOI: 10.4102/sajip.v36i1.801

Koekemoer, E., Mostert, K., & Rothman, S. Jr, (2010). Interference between work and nonwork roles: The development of a new South African instrument. SA Journal of Industrial Psychology/SA Tydskrif vir Bedryfsielkunde, 36(1), Art. #907, 14 pages. DOI:10.4102/sajip.v36i1.907

Kopelman, R.E., Greenhaus, J.H., & Connolly, T.F. (1983). A model of work, family, and interrole conflict: A construct validation study. Organizational Behavior and Human Performance, 32, 198–215.

Lee, J.A., & Phillips, S.J. (2006). Work and family: Can you have it all? The Psychologist-Manager Journal, 9, 41–57.

Lehmann, D. (2001). Structural equations modeling. Journal of Consumer Psychology 10, 90–92.

Lewis, S., & Cooper, C.L. (1999). The work-family research agenda in changing contexts. Journal of Occupational Health Psychology, 4(4), 382–393.

Lewis, S., & Cooper, C. (2005). Work-life integration: Case studies of organisational change. Chichester: Wiley.

Lingard, H., & Francis, V. (2005). Does work-family conflict mediate the relationship between job schedule demands and burnout in male construction professionals and managers? Construction Management and Economics, 23, 733–745.

Mallard, A.G.C., & Lance, C.E. (1998). Development and evaluation of a parent-employee interrole conflict scale. Social Indicator Research, 45, 343–370.

Marais, C., Mostert, K., Geurts, S.A.E., & Taris, T.W. (2009). Measuring work-home interaction in indigenous languages: The psychometric properties of a translated version of the Survey Work-Home Interaction-Nijmegen (SWING) instrument. South African Journal of Psychology, 39(2), 202–219.

Maslach, C., & Jackson, S.E. (1986). Maslach burnout inventory manual. Palo Alto: Consulting Psychologists Press.

Matthews, S., & Power, C. (2002). Socio-economic gradients in psychological distress: A focus on women, social roles and work-home characteristics. Social Science & Medicine, 54, 799–810.

McLean, V. (2002). Stress, depression and role conflict in working mothers. South African Psychiatry Review, 5(2), 13–16.

Mesmer-Magnus, J.R., & Viswesvaran, C. (2005). Convergence between measures of work-to-family and family-to-work conflict: A meta-analytic examination. Journal of Vocational Behaviour, 67, 215–232.

Montgomery, A.J., Panagopolou, E., & Benos, A. (2006). Work-family interference as a mediator between job demands and job burnout among doctors. Stress & Health, 22, 203–212.

Mostert, K. (2009). The balance between work and home: The relationship between work and home demands and ill-health of employed females. South African Journal of Industrial Psychology, 35, 1–8.

Mostert, K., & Oldfield, G.R (2009). Work-home interaction of employees in the mining environment. South African Journal of Economic and Management Sciences, 12(1), 81–99.

Nasurdin, A.M., & Hsia, K.L. (2008). The influence of support at work and home on work-family conflict: Does gender make a difference? Research Practice in Human Resource Management, 16, 18–38.

Netemeyer, R.G., Boles, J.S., & McMurrian, R. (1996). Development and validation of work-family conflict and family-work conflict. Journal of Applied Psychology, 81(4), 400–410.

Nikandrou, I., Panayotopoulou, L., & Apospori, E. (2008). The impact of individual and organisational characteristics on work-family conflict and career outcomes. Journal of Managerial Psychology, 23(5), 576–598.

Nordenmark, M. (2002). Multiple social roles – a resource or a burden: Is it possible for men and women to combine paid work with family life in a satisfactory way? Gender, Work and Organizations, 9(2), 125–145.

Nunnally, J.C., & Bernstein, I.H. (1994). Psychometric theory (3rd edn.). New York: McGraw-Hill.

Oldfield, G.R., & Mostert, K. (2007). Job characteristics, ill-health and negative work-home interference in the mining industry. South African Journal of Industrial Psychology, 33(2), 68–75.

Olson-Buchanan, J.B., & Boswell, W.R. (2006). Blurring boundaries: Correlates of integration and segmentation between work and nonwork. Journal of Vocational Behavior, 68, 432–445.

Parasuraman, S., & Greenhaus, J.H. (2002). Toward reducing some critical gaps in work-family research. Human Resource Management Review, 12, 299–312.

Parasuraman, S., Purohit, Y.S., Godshalk, V.M., & Beutell, N.J. (1996). Work and family variables, entrepreneurial career success, and psychological well-being. Journal of Vocational Behavior, 48, 275–300.

Peeters, M.C.W., Montgomery, A.J., Bakker, A.B., & Schaufeli, W.B. (2005). Balancing work and home: How job and home demands are related to burnout. International Journal of Stress Management, 12, 43–61.

Perrone, K.M., Webb, L.K., & Jackson, Z.V. (2007). Relationships between parental attachment, work and family roles, and life satisfaction. The Career Development Quarterly, 55, 237–248.

Pieterse, M., & Mostert, K. (2005). Measuring the work-home interface: Validation of the Survey Work-Home Interaction – Nijmegen (SWING) Instrument. Management Dynamics, 14(2), 2–15.

Polit, D.F., & Beck, C.T. (2006). Essentials of nursing research: Methods, appraisals and utilization. Philadelphia: Lippincott, Williams, & Wilkins.

Premeaux, S.F., Adkins, C.L., & Mossholder, K.W. (2007). Balancing work and family: A field study of multi-dimensional, multi-role work-family conflict. Journal of Organizational Behaviour, 28, 705–727.

Robinson, J.P., Shaver, P.R., & Wrightsman, L.S. (1991). Criteria for scale selection and evaluation. In J.P. Robinson, P.R. Shaver & L.S. Wrightsman (Eds.), Measures of personality and social psychological attitudes (pp. 1–16). New York: Academic Press.

Rost, I., & Mostert, K. (2007). The interaction between work and home of employees in the earthmoving equipment industry: Measurement and prevalence. South African Journal of Industrial Psychology, 33(2), 54–61.

Roxburgh, S. (2002). ‘I wish we had more time to spend together…’. The distribution and predictors of perceived family time pressures among married men and women in the paid labor force. Journal of Family Issues, 27(4), 529–553.

Schultheiss, D.E.P. (2006). The interface of work and family life. Professional Psychology: Research and Practice, 4, 334–341.

Small, S.A., & Riley, D. (1990). Toward a multidimensional assessment of work spillover into family life. Journal of Marriage and Family, 52, 51–61.

Snyder, C.R., & Lopez, S.J. (2002). Handbook of positive psychology. New York: Oxford University Press.

SPSS Inc. (2009). SPSS 12.0 for Windows. Chicago: SPSS Incorporated.

Stephens, G.K., & Sommer, S.M. (1996). The measurement of work to family conflict. Educational and Psychological Measurement, 56(3), 475–486.

Steyn, H.S. (2002). Practically significant relationships between two variables. South African Journal of Industrial Psychology, 28(3), 10–15.

Struwig, F.W., & Stead, G.B. (2001). Planning, designing and reporting research. Cape Town: Pearson Education.

Taormina, R.J. (2004). Convergent validation of two measures of organizational socialization. International Journal of Human Resource Management, 15, 76–94.

Tetrick, L.E., & Buffardi, L.C. (2006). Measurement issues in research on the work-home interface. In F. Jones., R.J. Burke & M. Westman (Eds.), Work-life balance, a psychological perspective (pp. 90–114). Hove, New York: Psychology Press, Taylor & Francis Group.

Tomarken, A.J., & Waller, N.G. (2005). Structural equation modeling: Strengths, limitations, and misconceptions. Annual Review Clinical Psychology 1, 31–65.

Trout, A.L., Ryan, J.B., Vigne, S.P., & Epstein, M.H. (2003). Behavioral and emotional rating scale: Two studies of convergent validity. Journal of Child and Family Studies, 12(4), 399–410.

Van Horn, J.E., Taris, T.W., Schaufeli, W.B., & Schreurs, P.J.G. (2004). The structure of occupational well-being: A study among Dutch teachers. Journal of Occupational and Organizational Psychology, 77, 365–375.

Voydanoff, P. (2005a). The differential salience of family and community demands and resources for family-to-work conflict and facilitation. Journal of Family and Economic Issues, 26(3), 395–417.

Voydanoff, P. (2005b). Work demands and work-to-family and family-to-work conflict: Direct and indirect relationships. Journal of Family Issues, 26(6), 707–726.

Voydanoff, P. (2007). Work, family, and community, exploring interconnections. Mahwah, New Jersey: Lawrence Erlbaum Associates.

Weigel, D.J., Weigel, R.R., Berger, P.S., Cook, A.S., & DelCampo, R. (1995). Work-family conflict and the quality of family life: Specifying linking mechanisms. Family and Consumer Sciences Research Journal, 24(5), 5–28.