Advertisement

Occupational Health Science

, Volume 2, Issue 2, pp 181–201 | Cite as

A Fresh Look at Socio-Demographics in Work-Family Conflict: a Cluster Analysis Approach

  • Kyle J. Page
  • Jacqueline K. Deuling
  • Joseph J. Mazzola
  • Kathleen M. Rospenda
Original Research Article

Abstract

An important gap in work-family literature is the understanding of how socio-demographic variables, such as sex, age, hours worked, age of youngest child, and household income may relate to work-family conflict. Using data from 667 individuals and longitudinal data from 1007 caregivers, separate exploratory cluster analysis by gender provided a three cluster solution for caregiving men, non-caregiving men, and caregiving women and a four cluster solution for non-caregiving women. Differences in work interfering with family were found in caregiving men, caregiving women, and non-caregiving women clusters. Non-caregiving men, non-caregiving women, and caregiving women had differential levels of family interfering with work by cluster. Cohen’s D revealed that age had the largest effect size between clusters for individuals and caregivers. Findings and implications are discussed.

Keywords

Work-family conflict Socio-demographics Cluster analysis 

Balancing work and family is a central concern that affects nearly all individuals. A large portion of the current work-family research focuses on antecedents (e.g., Burke 1988; Michel et al. 2011), sources of conflict (e.g., Greenhaus and Beutell 1985), and specific outcomes (e.g., Wolff et al. 2014). Thus far, socio-demographic variables have been considered, but research routinely finds conflicting results with their specific role, prevalence, and effects on experience of work-family conflict (WFC). Countless individual studies and meta-analyses denote that socio-demographic variables have a complex relationship with WFC as societal and cultural norms have varying effects (e.g., Korabik et al. 2017; Shockley et al. 2017). As such, research varying in breadth and depth has explored a plethora of variable-centered approaches (e.g., Regression, Hierarchical Linear Modeling, and Structural Equation Modeling) that increase in sophistication and model complexity. To date, research rarely takes a person-centered approach (e.g., Cluster Analysis or Latent Profile Analysis) to create a more accurate depiction of an employee’s life. In essence, person-centered approaches differ from variable-centered approaches by creating a profile of each individual through a configuration of several variables. This is done to get a more detailed understanding and to see how a profile of an individual relates to changes in key variables rather than simply observing the relationship between variables. Said differently, differences are considered across conceptually distinct groups whose membership is defined by similarities between group members on a set of predictors of interest (Craig and Smith 2000; Sinclair et al. 2005).

As direct effects, moderating, and mediating relationships between socio-demographic variables and WFC have had mixed and inconclusive results, new techniques should be considered that allow a greater understanding of how a combination of variables interact. In an attempt to further our understanding, this study utilizes cluster analysis to combine socio-demographic variables and builds on prior WFC research for age (e.g. Boyar et al. 2003), hours worked (e.g., Ng and Feldman 2008), age of youngest child (e.g., Byron 2005), gender (e.g., Emslie and Hunt 2009), and income (Michel et al. 2011). With a deeper understanding, better intervention programs, work-family policies, and strategies to maintain work and family roles may be created and/or discovered.

Work-Family Literature & Socio-Demographic Clusters

Work-family conflict, defined as “a form of interrole conflict in which the role pressures from the work and family domains are mutually incompatible in some respect” (Greenhaus and Beutell 1985, p. 77), is distinguished by the direction of conflict, specifically family interfering with work (FIW) or work interfering with family (WIF; Netemeyer et al. 1996), and has exploded in popularity as a research topic. This is in part due to substantial workforce changes such as increases in dual-career couples, single-parent households, and a blurring of gender roles as women have begun to place a higher priority on their work and careers and men take a more active role in family matters (Brough and O’Driscoll 2015; Byron 2005; Lyness and Judiesch 2008). Galinsky et al. (1993) stated that 83% of working mothers and 72% of working fathers experience some conflict between work and family expectancies while Sutton and Noe (2005) noted that American employees are spending less time at home and more time at work than at any point in history. In Finding Time – The Economics of Work-Life Conflict, Boushey (2016) explains that employers are demanding more of employees’ time and that resolving such conflict is vital for individuals and families. Further, the topic of balance between work and family is a concern within organizations (Society of Human Resource Management Work-place Forecast 2016) as a balance between work and family is an important piece of obtaining life satisfaction (Drobnic 2011). If the relationship between socio-demographics and WFC is changing as a result of the evolving workforce, it is vital to develop a thorough understanding of which variables are associated with increased conflict and who experiences these increases.

A multitude of theories have been used throughout work-family research (for a thorough review see Perry-Jenkins and Wadsworth 2013). Some theories, such as expansionist theory (Barnett and Hyde 2001) or work-family facilitation theory (e.g., Greenhaus and Powell 2006), have examined the benefit associated with membership in multiple roles while other theories, such as role stress (e.g., Shockley and Allen 2013), scarcity theory (Marks 1977), and conservation of resources theory (e.g., Halbesleben et al. 2014; Hobfoll 1989), have focused on the detrimental impact of occupying multiple roles. This study follows the latter group of theories drawing particularly on scarcity theory which states that as one takes on more roles a division of time and attention between said roles occurs and leaves less time available for any one specific role (Marks 1977). This benefit versus detriment perspective applies to the socio-demographic variables that are the focus of this study. Adding an additional role, such as becoming an employee or a parent, is likely to reduce the amount of resources available and serve as a detriment. This may be seen in time availability. If more time is spent at work, less time will be available for family. If a new child is added into the mix, not only do individuals have a social role to fill in terms of relationships with others, but these individuals also have to care for a truly dependent person. Next, the role of specific socio-demographics on WFC will be considered.

Age

Researchers have posited that the “question is not whether WFC is experienced throughout the lifespan, but rather when is this conflict greatest, and what factors explain the relationship between age and WFC” (Huffman et al. 2013, p. 4). As such, there has been consideration of differing levels and types of WFC at various life stages (Baltes and Young 2007). Younger workers may be less committed to a career due to focus on personal identity (Erikson 1968), furthering education, or starting a family (Evans and Bartolomé 1984; McDaniels and Gysbers 1992). These employees frequently hold jobs with lower-level responsibilities (Bauman 2001; Beck 1992). With age, employees are likely to work more hours and simultaneously increase work-related and family-related responsibilities. Specifically, the likelihood of personal life changes, such as parenthood, may cause drastic changes in juggling family and work life (Kaufman and Uhlenberg 2000). As resources are considered to be finite (scarcity theory; Marks 1977), devoting more resources to one particular role means that fewer resources will be available for another (Sieber 1974). Therefore, individuals in middle adulthood are likely to have heightened levels of WFC. Near the age of retirement, when the children are grown, mortgage(s) are paid, and time is more abundantly spent on leisure activities, WFC levels may decrease. It is also expected that employees during late adulthood have developed more effective coping mechanisms throughout a lifetime of problem-solving, communication, and experience (Baltes and Young 2007). Thus, age may serve as both a detriment and a benefit for WFC.

The relationship between age and WFC has been well studied but discrepancies are frequent. For example, the relationship for WIF and FIW has been found to be: (1) dependent on family stage (Allen and Finkelstein 2014), (2) curvilinear in an inverted U shape (Huffman et al. 2013), and (3) significant only for workers aged fifty-five years or older (Hill et al. 2014). Additionally, it is possible that age has a weak relationship with distal outcomes (i.e., WFC) and should be used in conjunction with other socio-demographic variables to better predict differences among individuals or groups of individuals.

Hours Worked

One of the causes of WFC is insufficient time for both work and family domains (Greenhaus and Beutell 1985). Most Americans are believed to suffer from a “time bind” where the hours spent at work have a negative impact on their non-work lives (Hochschild 1997). According to the rational view (Gutek et al. 1991) the level of WFC a person experiences depends on the amount of time that person spends in each domain. Previous studies have generally found a positive relationship between hours worked and WIF/FIW (e.g., Gutek et al. 1991; Major et al. 2002); however, Barnett (1998) noted that these linear findings have been weak, inconsistent, and may be non-linear. Similarly, a meta-analysis by Byron (2005) found that individuals who spend more time at work have higher levels of WIF and FIW.

One reason that the hours worked and WFC relationship may be weak is that having an identity as an employee in addition to any non-work identities creates a more complex sense of self. These multiple roles and identities, as described by self-complexity theory (Linville 1987), may serve as a benefit to easing WFC. By having multiple identities to draw on during times of stress, the effects of stressors on the mind and body may lessen. Being a workaholic or seemingly addicted to work, however, may be a detriment to perceptions of WFC (Gillet et al. 2017). Thus, it may be that hours worked has a weak, distal relationship with WFC that may best be detected when combining multiple socio-demographic variables. In line with this, Barnett (1998) states that long hours should be construed as a risk factor when combined with other variables that lead to negative outcomes and that there is no a priori reason to expect long hours to have the same effects across workers and conditions. Barnett (1998) further posits that one of the largest inconsistencies seems to be the relationship between hours worked and WFC. This may be seen as WIF has been found to have a linear (Adkins and Premeaux 2012) and curvilinear (Ganster and Bates 2003; Ng and Feldman 2008) relationship with hours worked and FIW has been found to have no relationship (Byron 2005) and a curvilinear relationship (Adkins and Premeaux 2012) with hours worked. These inconsistencies show that there is much to learn about the relationship between hours worked and WFC.

Children

Children amplify emotional and time constraints (Galinsky 1999; Taylor and Dao 2003). Research has found that having children is a “critical catalyst” for differences in the balance between work and family life, that the greatest gender differences were found in central stages of family life (constructed by age of respondent and age of youngest child), and that when life stage is not considered, gender differences are concealed (Martinengo et al. 2010). In a meta-analysis, Byron (2005) found that age of the youngest child significantly affected WFC. The same study found the number of individuals that have children (i.e., parents) moderates the relationship between gender and WFC or rather as the number of parents in a study increases, mothers experience more WFC than fathers but when there are fewer parents in a study, men tend to experience higher levels of WFC. Byron (2005) also noted that single parents experience more WFC than parents who are married, whereas married and single individuals without children have similar levels of WFC. However, self-complexity theory (Linville 1987) would suggest that having the role or identity of “parent” may help buffer WFC effects.

From the aforementioned examples, it appears that the relationship between WFC and age of youngest child may exacerbate or mitigate other socio-demographic variables, such as age or gender. While some studies are beginning to combine variables to better explore how socio-demographic variables may interact and affect WFC levels (i.e., Allen and Finkelstein 2014), the intricacies of the relationship are largely understudied.

Income

Income is complex when considering WFC. Following scarcity theory, one may argue that higher income may indicate more time spent at work and time away from non-work roles and responsibilities and thus serve as a detriment to WFC. Alternatively, family income may be beneficial as more money allows for more resources to manage WFC. Meta-analyses (Byron 2005; Michel et al. 2011) have found that employees with higher incomes have more WFC. In studies focused on dual earners, the use of childcare facilities has been associated with increases in women's conflict; further, women's use of coping strategies is more associated with WFC and work-family enrichment than men’s (Matias and Fontaine 2015). Thus, the simple use of WFC resources may increase women’s conflict creating a more complex mosaic for income’s role in WFC. By taking a person-centered approach, we aim to identify the more nuanced relationship between income and WFC.

Gender

Gender has been the focus of multiple investigations due to the belief that women tend to experience greater family demands while men tend to experience greater work demands (e.g., Powell and Greenhaus 2010). Research supports a weak relationship between gender and WFC (Byron 2005). Emslie and Hunt (2009) found that men’s and women’s accounts of WFC differed in two important ways: (1) men focused on the way paid work dominated their lives while women discussed the issues of balancing home and work life and (2) that men reported past problems while women focused on recent or current problems.

There has been a narrowing of the gender gap in housework (Bianchi et al. 2000) and childcare (Bianchi et al. 2006); however, gender differences are still found. Despite narrowing, American mothers still spend nearly twice as much time as fathers per day in household activities and caring for household members (U.S. Bureau of Labor Statistics 2014a) while men work almost an hour more per day or half an hour more when comparing full time workers (U.S. Bureau of Labor Statistics 2014b). Gender role theory (e.g., Rajadhyaksha et al. 2015) suggests that the primary domain for men is work while for women it is family. This is compounded when work and family interactions are imbedded in cultural, institutional, and economic context (Ollier-Malaterre and Foucreault 2017). Further, it has been argued that women have more permeable and flexible role segmentation which may stem from gender differences in mental models about boundary management between work and family which are also influenced by gender role socialization (Andrews and Bailyn 1993; Crosby 1991; Innstrand et al. 2009; Rothbard 2001). Despite evidence of similarity of WFC perceptions, it has been found that men and women have slight yet significant differences in multiple comparisons (for more detail see Shockley et al. 2017). With this in mind, our analysis was split by gender to allow for greater understanding with less confounding information. In summary, it is unclear how age, hours worked, income, and age of youngest child combine to affect WFC. Grounded and diverse research has been conducted, but there are still inconsistencies of how socio-demographic variables affect level of WFC.

Socio-Demographic Clusters

So far, research has generally adhered to two overarching paradigms of variable-centered approaches: (1) the direct relationships of antecedent or outcomes to WFC or (2) mediating/moderating relationships that may explain further variance. Variable-centered approaches are appropriate if the goal is to understand how individuals are different or to capture relationships among a limited number of variables within a group but not if the goal is to understand the configurations of variables and how they operate within individuals (Block 1971). A vast amount of knowledge has been discovered by employing variable-centered approaches but these approaches ignore that participants may come from different subpopulations in which observed variables may differ while person-centered (e.g., cluster analysis) approaches help categorize individuals more accurately into distinct profiles (Morin et al. 2011). Byron (2005) noted that demographics are poor predictors of WFC when not combined with other more domain specific variables. It is plausible that the use of person-centered approaches will provide information that is previously undiscovered.

Person-centered approaches, specifically cluster analysis, will allow exploration of multiple variables that may be combined. These combinations may then be used to understand individuals’ experience of WFC and how groups with varying combinations differ. The utilization of cluster analysis to explore how a variety of socio-demographic variables (i.e., age, hours worked, household income, and age of youngest child) interact may help researchers conceptualize the role of socio-demographics and may enable companies to produce better and more customized work-life balance programs.

Person-centered approaches are beginning to be used more regularly in WFC literature. For example, latent profile analysis was used to study typology of work-family balance and three types of individuals were found through the combination of work and family conflict and enrichment (Rantanen et al. 2013). Specifically, cluster analysis has been used rarely in the work-family literature. Cullen et al. (2009) utilized cluster analysis of child care demands (age and number of children), hours spent caregiving per week, and hours worked per week, in a sample of dual-earner couples. Three profiles were created and differences in WIF but not FIW were found. Mitchelson (2009) used cluster analysis to group non-, adaptive-, and mal-adaptive perfectionists and found that perfectionism predicts WFC beyond the Big Five, trait affectivity, and achievement. Additionally, cluster analysis was used to create boundary management styles to test differences in WFC (Kossek et al. 2012). In related, but not specifically WFC literature, cluster analysis was used to create job ecologies that responded to the job demands-job control model and the analogous home demands-home control model to examine life-course fit (Moen et al. 2008). These examples show that although person-centered approaches are not the forefront of WFC research, there is support for their use and effectiveness. Therefore, this study will use cluster analysis to investigate differences of men and women (separately) through the combination of age, hours worked, age of youngest child, and household income, in an effort to better understand how the constellation of socio-demographics relate to WFC.
  • Research Question 1: Once separated by gender, are there multiple clusters with respect to age, hours worked, age of youngest child, and household income?

  • Research Question 2: Which variable(s) are key to the socio-demographic cluster solution?

  • Research Question 3: How do these clusters relate to WFC?

Current Studies

This study utilizes archival data from three different samples with one sample consisting of longitudinal data (3 time points). The first study (referred to as Study 1) combines two samples (total N = 667) from a more heterogeneous sample of the general population of working adults (20 or more hours a week) while the second study (referred to as Study 2) considers how the created clusters act in a longitudinal time frame (NWave1 = 1007, NWave2 = 713, NWave3 = 689) among employees with more homogeneous and intensive caregiving responsibilities (i.e., caring for children or aging and disabled family members). Sample 1 was previously used in Mitchelson (2009), Sample 2 was previously used in Deuling and Burns (2017), and Sample 3 was previously used in Rospenda et al. (2013), Wolff et al. (2014); Wolff et al. (2013).

Study 1

Method

Participants and Procedure

Sample 1

Participants were working adults with family obligations recruited by undergraduate students enrolled in psychology courses in a Midwestern American university. Undergraduates were given extra credit to recruit two working adults who worked at least an average of 20 h per week and had a significant other and/or lived with dependent children. Measures were presented online in random order. The average age was approximately 38 years old (SD = 11.5) with the ethnic distribution of the sample being 47% White/European American, 28% Black/African-American, 10% Asian, 5% Arab/Middle Eastern, 3.5% Hispanic, and the remainder either identifying as multiracial or “other” while 62% were female and 38% were male.

Sample 2

Participants were working adults (20 or more hours a week) with family obligations of living with a significant other and/or having dependent children. The average age was approximately 40 years old (SD = 9.4) with the ethnic distribution of the sample being 81% White/European American, 6% Hispanic, 4% African-American, 2% Asian American, 1% Native American, 2% multiracial, and the remainder listed as “other” or unidentified while roughly 60% were female and roughly 40% were male. Adult workers were recruited via a non-profit online database of research participants. At time 1, participants completed the socio-demographic measure and one week later, participants completed the work-family conflict scale (93% response rate). Participants received a $5 gift certificate for an online retail merchant for completion of each time wave of the survey.

Measures

Work-Family Conflict

An 18-item multidimensional scale (Carlson et al. 2000) where 9-items reflected WIF (Sample 1, α = .90; Sample 2, α = .84) and 9-items reflected FIW (Sample 1, α = .92; Sample 2, α = .88) was used. This measure uses a 5-point Likert-response scale with anchors of Strongly Disagree (1) to Strongly Agree (5).

Socio-Demographics

Socio-demographic information was collected for age, gender, number of children, age of youngest child, marital status, and hours worked per week (see Tables 1 and 2).
Table 1

Study 1 (men): cluster income, age of youngest child, marital status, and living situation

 

Young Childfree Man (N = 39)

Middle-age Man (N = 114)

Established Man (N = 91)

Age mean

27.1

36.2

52.9

Hours worked mean

39.2

45.7

39.1

Number of children

0.3

1.8

1.1

Income

 Under $15,000

2.6

4.4

 $15,000 - $30,000

12.8

13.2

1.1

 $30,001 - $45,000

23.1

13.2

17.6

 $45,001 - $75,000

28.2

39.5

33.0

 $75,001 - $100,000

15.4

21.9

31.9

 $100,001 - $125,000

5.1

5.3

7.7

 $125,001 - $150,000

5.1

1.8

3.3

 Above $150,000

7.7

0.9

5.5

Age of youngest child

 No children

94.9

3.5

51.6

 Above 18 years

2.6

25.3

 13–18 years

2.6

20.2

23.1

 6–12 years

33.3

 Below 6 years

43.0

Marital Status

 Single

25.6

5.3

4.4

 Exclusively dating

5.1

4.4

2.2

 Married, engaged, living together

69.2

86.8

87.9

 Separated, divorced, widowed

3.5

5.5

Living Situation

 Alone

15.8

6.2

6.7

 Friend/roommate

5.3

1.8

1.1

 Relative/parent

15.8

7.1

2.2

 Spouse/significant other

63.2

85.0

90.0

WIF

2.73

2.67

2.45

FIW

2.57b

2.52b

2.16a

Income, age of youngest child, marital status, and living situation are in percentages. WIF, work interfering with family; FIW, family interfering with work. Different letters indicate significantly different means by cluster. Total N = 244. Different letters indicate significantly different means by cluster

Table 2

Study 1 (women): cluster income, age of youngest child, marital status, and living situation

 

Young Childfree Woman (N = 58)

Low Income Woman (N = 68)

Middle-age Woman (N = 114)

Established Woman (N = 132)

Age mean

26.8

28.0

39.9

49.1

Hours worked mean

38.6

31.1

42.5

34.1

Number of children

0.1

1.7

1.7

1.2

Income

 Under $15,000

6.9

14.7

5.3

 $15,000 - $30,000

13.8

42.6

1.8

16.7

 $30,001 - $45,000

19.0

22.1

7.9

11.4

 $45,001 - $75,000

43.1

14.7

56.1

23.5

 $75,001 - $100,000

12.1

4.4

20.2

21.2

 $100,001 - $125,000

5.2

1.5

7.9

9.8

 $125,001 - $150,000

2.6

6.8

 Above $150,000

3.5

5.3

Age of youngest child

 No children

98.3

3.5

40.9

 Above 18 years

1.7

0.9

26.5

 13–18 years

2.9

26.3

27.3

 6–12 years

25.0

41.2

4.5

 Below 6 years

72.1

28.1

0.8

Marital Status

 Single

19.0

19.1

3.5

2.3

 Exclusively dating

6.9

4.4

4.4

0.8

 Married, engaged, living together

70.7

63.2

78.8

82.4

 Separated, divorced, widowed

3.4

13.2

13.3

14.5

Living Situation

 Alone

10.3

14.7

4.4

8.3

 Friend/roommate

1.5

0.9

2.3

 Relative/parent

17.2

22.1

11.4

7.6

 Spouse/significant other

72.4

61.8

83.3

81.8

WIF

2.69

2.78

2.79a

2.48b

FIW

2.45a

2.52a

2.42a

2.10b

Income, age of youngest child, marital status, and living situation are in percentages. WIF, work interfering with family; FIW, family interfering with work. Different letters indicate significantly different means by cluster. Total N = 372. Different letters indicate significantly different means by cluster

Analysis

To test whether there is a group structure for the variables of age of participant, hours worked, age of the youngest child, and household income (Research Question 1), Ward’s method of hierarchical clustering with squared Euclidean distance based on standardized versions of the clustering variables was used for each gender separately (Ward 1963). Traditional methods propose plotting the agglomeration coefficient on the y-axis and the number of clusters on the x-axis, which would resemble a scree plot in exploratory factor analysis (Ketchen Jr. and Shook 1996) to find a particular number of clusters. Additionally, length of the vertical lines on the dendrogram of each cluster analysis was utilized. To address Research Question 2, which sought to find which variables were key to the cluster solution, ANOVAs were run for each of the clustering variables using the clusters as the independent variable and clustering variables as the dependent variable. Lastly, to address Research Question 3, which sought to find how the created clusters relate to WIF/FIW, ANOVAs were run for each gender separately with WIF/FIW being the dependent variable and cluster membership being the independent variable.

Results

In an effort to be thorough, analyses were conducted with gender not separated between clusters. As expected these analyses resulted in clusters heavily aligned with gender. Specifically, for Study 1, four clusters were present with three of the cluster heavily sorted by gender and the last being split approximately 60% (women) to 40% (men).

When split by gender, a 3-cluster solution for men and a 4-cluster solution for women were supported. The clusters and their accompanying socio-demographic distributions for men and women are provided in Tables 1 and 2, respectively. Each cluster was given a descriptive name to illustrate demographics of employees that largely populate each cluster. Clusters that were similar for men and women were given similar names. In a further effort to be thorough, analysis was run separately for Sample 1 and 2. Although not identical with the clusters created from the combination of Sample 1 and 2, the results were similar. In Sample 1, 3 clusters were supported for men and four clusters were supported for women while in Sample 2, 3 clusters were supported for both men and women.

For Research question 2, a series of one-way ANOVAs were run by cluster to test which variables were key to the cluster solutions (with Tukey post-hoc). For study 1, hours worked, age of participant, age of youngest child, and household income were significant for both men and women (p < .01). Additional analyses (Cohen’s D) were used to assess the individual contribution for each socio-demographic variable between created clusters. Cohen’s D was considered to reveal which demographic information was most useful in the creation of our clusters. This was done by comparing all cluster means and providing the minimum and maximum Cohen’s D ranges for each demographic. For men, Cohen’s D ranges were from −2.06 to 4.03 for age and −.53 to .63 for hours worked. For women, Cohen’s D ranges were from −3.93 to 1.37 for age and −.81 to 1.10 for hours worked.

For Research question 3, none of the clusters differed on level of WIF (F(2, 231) = 2.3, p = .10) for men. There were, however, significant differences in level of FIW (F(2,231) = 4.88, p < .01). The Established Man (M = 2.16) had lower FIW than both the Young Childfree Man (M = 2.57; d = .33, p < .05) and the Middle-age Man (M = 2.52; d = .28, p < .05). For women clusters (F(3,345) = 2.81, p < .05) the Established Woman (M = 2.48; d = .33, p < .05) significantly varied on level of WIF compared to the Middle-age Woman (M = 2.79). For FIW (F(3,345) = 5.13, p < .01), the trend continued with the Established Woman (M = 2.10) significantly lower than the Young Childfree Woman (M = 2.45; d = −.44, p < .05), Low Income Woman (M = 2.52; d = −.52, p = .005), and the Middle-age Woman (M = 2.42; d = −.40, p < .05).

Study 2 - Participants and Procedure

Data were collected in a three-wave survey of employed adults (at least 18 years old and employed for 20 h/week) by purchasing numbers randomly selected from block groups in Chicagoland. The survey was a combined technique of random digital dial (RDD) telephone recruiting and mailed self-report survey. In households with more than one eligible respondent, the Troldahl-Carter-Bryant method (Lavrakas 1987) was used to select who would participate.

Participants received a $30 American Express gift card incentive to complete each wave of the questionnaire. Respondents who completed all three waves of questionnaires had no significant difference in gender or race but were more likely to be older (β = .03, p < .001) with the average age being 43 years old (SD = 9.9). Of the participants who completed all three phases 43.7% of the participants were white, 34.8% were African American, 15.7% were Latino/a, and 4.9% were “other” race/ethnicity, while .9% was missing. Approximately 46% were male while roughly 54% were female. This sample consisted of caregivers who were responsible for children under age 18 (75.7%), children over age 18 (15.4%), a spouse/partner (24.1%), or parents (21.5%), and/or siblings, aunts or uncles, and grandparents (less than 6% each).

Measures

Work-Family Conflict

WFC was measured with a 22-item measure developed by Kelloway et al. (1999) where 11-items measured WIF and FIW respectively. Items are rated on a Likert-type scale (1 = never, 4 = almost always). All waves showed acceptable alpha levels (WIF, α = .89 to .90; FIW, α = .90 to.91).

Socio-Demographics

Socio-demographic information was collected for age, gender, number of children, age of youngest child, marital status, and hours worked in Wave 1 (see Tables 3 and 4).
Table 3

Study 2 (men): cluster income, mean age of youngest child, and marital status

 

Low Income Man (N = 31)

Middle-age Man (N = 160)

Established Man (N = 119)

Age mean

32.5

37.2

48.1

Hours worked mean

31.6

48.5

40.4

Age of youngest child

6.1

3.5

9.9

Number of children

1.9

2.1

1.8

Income

 $10,000 or below

19.4

0.6

 $10,001 - $20,000

54.8

 $20,001 - $30,000

12.9

6.3

6.7

 $30,001 - $50,000

9.7

15.6

16.8

 $50,001 - $70,000

3.2

12.5

26.9

 $70,001 - $90,000

16.3

18.5

 Above $90,000

48.8

31.1

Marital status

 Never married

12.9

.6

0.9

 Committed relationship

22.6

3.1

6.0

 Married

61.3

93.1

84.6

 Separated, divorced, widowed

3.2

3.1

8.5

 WIF Wave 1

2.11b

2.54a

2.34b

 WIF Wave 2

2.09a

2.47b

2.38

 WIF Wave 3

1.91b

2.07b

2.26a

 FIW Wave 1

1.85a

2.21b

2.04

 FIW Wave 2

1.83

1.96

1.94

 FIW Wave 3

1.82

2.04

1.86

Income and marital status are in percentages. WIF, work interfering with family; FIW, family interfering with work. Different letters indicate significantly different means by cluster. Total N = 310. It should be noted that as attrition occurs, N diminishes for each cluster. Different letters indicate significantly different means by cluster

Table 4

Study 2 (women): cluster income, mean age of youngest child, and marital status

 

Low Income Woman (N = 156)

Middle-age Woman (N = 85)

Established Woman (N = 86)

Age mean

36.1

37.1

48.6

Hours worked mean

33.3

36.0

44.4

Age of youngest child

7.5

2.9

12.1

Number of children

2.0

1.8

1.6

Income

 $10,000 or below

17.3

1.2

 $10,001 - $20,000

18.6

2.3

 $20,001 - $30,000

23.7

4.7

 $30,001 - $50,000

26.3

7.1

31.4

 $50,001 - $70,000

10.3

11.8

22.1

 $70,001 - $90,000

1.9

25.9

10.5

 Above $90,000

1.9

55.3

27.9

Marital status

 Never married

25.5

4.7

11.8

 Committed relationship

26.8

5.9

8.2

 Married

24.2

80.0

42.4

 Separated, divorced, widowed

23.5

9.4

37.6

 WIF Wave 1

2.16a

2.56b

2.38b

 WIF Wave 2

2.26a

2.52b

2.30

 WIF Wave 3

1.99a

2.35b

2.08a

 FIW Wave 1

1.96a

2.60b

2.20c

 FIW Wave 2

1.94a

2.36b

1.97a

 FIW Wave 3

1.90a

2.36b

1.97a

Income and marital status are in percentages. WIF, work interfering with family; FIW, family interfering with work. Different letters indicate significantly different means by cluster. Total N = 327. It should be noted that as attrition occurs, N diminishes for each cluster. Different letters indicate significantly different means by cluster

Results

As in Study 1, analysis was run with gender not separated between clusters to be thorough. As expected this analyses resulted in clusters heavily aligned with gender. Specifically, three clusters were created with one being primarily men, one being primarily women, and the final being a mixed gender cluster. Thus, we split the samples by gender. When analysis was run separately by gender, a 3-cluster solution was supported for men and women caregivers separately (Research Question 1). The clusters and their accompanying socio-demographics distributions for men and women are provided in Tables 3 and 4, respectively. Each cluster was given a descriptive name to illustrate which individuals had been sorted into each cluster. For Research Question 2, age of participant, hours worked, age of youngest child, and household income were all significant for men and women caregivers (p < .001). Similarly, Cohen’s D was conducted by comparing the highest mean cluster to the other clusters. For men, Cohen’s D ranges were from −1.90 to 2.70 for age, −.90 to 1.51 for hours worked, −1.50 to .61 for age of youngest child, and −.23 to .09 for number of children. For women, Cohen’s D ranges were from −1.65 to 1.96 for age, −.24 to .92 for hours worked, .61 to 1.50 for age of youngest child, and −.37 to −.05 for number of children.

While we had no a priori theory regarding change in WIF/FIW over time, all waves of our longitudinal data were used to consider the utility of created clusters over time. Specifically, we investigated whether demographic cluster differences are stable over time as it is important to see if time will render the initial cluster solution useless. As such, the data shows that even after two years, the original demographic clusters detect differences in WIF/FIW.

For men caregivers, there was a significant difference in WIF in Wave 1 (F (2,290) = 7.19, p = .001) in Wave 2 (F(2,198) = 3.43 p < .05) and Wave 3 (F(2, 190) = 5.31, p < .01). Tukey post hoc comparisons reveal the Middle-age Man cluster had the highest WIF across waves; specifically for Wave 1 (M = 2.54), which differed from the Established Man (M = 2.34; d = .33, p < .05) and the Low Income Man (M = 2.11; d = .35, p < .005). Similarly, for Wave 2 there was a difference between the Middle-age Man (M = 2.47) and the Low Income Man (M = 2.09; d = .51, p < .05). Wave 3 was similar to Wave 1 with the Middle-age Man (M = 2.26) differing from the Established Man (M = 2.07; d = .40, p < .05) and the Low Income Man (M = 1.91; d = .69, p < .05). There were no significant differences in level of FIW for the men caregiver clusters at Wave 2 (F(2,199) = .40, p > .05) or at Wave 3 (F(2,192) = 2.55, p < .10) but for Wave 1, there was a significant difference in FIW (F(2,298) = 3.77, p < .05) with differences found between the Middle-age Man (M = 2.21) and the Low Income Man (M = 1.85; d = .49, p < .05).

For women caregivers, there was a significant difference in WIF for Wave 1(F(2,309) = 10.21, p < .001) where the Low Income Woman had a lower level than the Middle-age Woman (M = 2.56; d = .63, p < .001) and the Established Woman (M = 2.38; d = .31, p = .051). For Wave 2 (F (2,220) = 3.56, p < .05) there was a significant difference in WIF when comparing the Middle-age Woman (M = 2.52) and the Low Income Woman (M = 2.26; d = .76, p < .05). For Wave 3, there was a significant difference in WIF (F (2,184) = 6.60, p < .005) for the Middle-age Woman (M = 2.35) compared to both the Established Woman (M = 2.07; d = .49, p < .05) and the Low Income Woman (M = 1.99; d = .51, p = .001).

For FIW in Wave 1, there was a significant difference (F(2,312) = 19.29, p < .01) between clusters where the Middle-age Woman (M = 2.60) differed from the Established Woman (M = 2.20; d = .54, p < .01) and the Low Income Woman (M = 1.96; d = .84, p < .001), and the Established Woman (M = 2.20) differed from the Low Income Woman (M = 1.96; d = .32, p < .05). For Wave 2, there was a significant difference in FIW (F (2,217) = 10.49, p < .001) with the Middle-age Woman (M = 2.36) having higher ratings than the Established Woman (M = 1.97; d = .70, p = .001) and the Low Income Woman (M = 1.94; d = .76, p < .001). Finally, this trend continues in Wave 3 with a significant difference in FIW (F (2, 188) = 10.43, p < .01) with the Middle-age Woman (M = 2.36) having higher ratings compared to the Established Woman (M = 1.97; d = .64, p < .005) and the Low Income Woman (M = 1.90; d = .78, p < .001).

Discussion

Work-family research has identified inconsistencies in findings between socio-demographic variables and the experience of WFC (Byron 2005; Michel et al. 2011). Using cluster analysis, a person-centered approach, we found that combinations of socio-demographic variables created meaningful clusters and differences in levels of WFC. Although each socio-demographic variable was key to the clustering solution, age played a larger role than other variables for both men and women. Amongst caregivers, age was followed by the role of hours worked per week for men and age of youngest child for women in effect size by cluster. Across studies and in general, the Middle-age Woman cluster had the highest levels of WFC. This has been an assumption in most theory and research on WFC but has been a finding that is difficult to capture. More specific findings are discussed next.

In our more heterogeneous sample (Study 1), the Middle-age Woman cluster had higher WIF than the Established Woman. The Middle-age Woman on average was 10 years younger than the Established Women, worked approximately 8 more hours per week, and had younger children at home. Thus, the combination of increased workload and younger children may be related to an increased level of WIF. Further, the Established Woman had the lowest level of FIW of all the clusters. This cluster was roughly 49 years old, 41% had no children while others had children that are at least 13 years old or older, and 82% were married.

Similarly, the Established Man had lower FIW than both the Young Childfree Man and the Middle-age Man. The Established Man cluster was older. Half of these individuals had no children (or possibly no children that still lived at home) and half had children that were at least 13 years old. Further, 88% were married and 90% lived with a spouse or significant other. Taken together, as these established men and women are older, it is possible that they will hold more authoritative positions in organizations due to tenure. Although this is not directly tested in the data, it is possible that work-family policies are created and enacted by more established employees in influential decision-making groups. As this is beyond our data, future research should consider this possibility. This could be done by assessing level of employment and tenure within a position or by asking specific questions about whether an employee has direct or indirect influence over the work-family policies created. It is less than advantageous for companies to have individuals who are removed from the peak of WFC in their own lives to be creating work-family policies. One caveat is that this finding is not replicated within the more homogeneous sample of caregiver clusters (Study 2). Future research should consider the boundary conditions of this effect as it did not generalize to a sample with more intense family obligations. Considering a sample with more intense work obligations may also reveal interesting implications for experienced WFC amongst socio-demographic clusters.

In our more homogeneous sample of caregivers, the findings reflect the higher levels of WFC experienced by the Middle-age Woman cluster. Specifically, the Middle-age Woman cluster had higher WIF and FIW than the Low Income Woman for all time points. These groups are similar in age and hours worked, but the Middle-age Woman had younger children at home (2.9 years versus Low Income Woman at 7.5 years), earned a higher income (81% had an income of $70,000 or more versus Low Income Woman where 86% made $50,000 or less), and were likely married (80% were married versus Low Income Woman where an almost equal distribution of never married, in a committed relationship, married, or had been separated, divorced, or widowed were found). Prior variable-centered research found that age of the youngest child significantly affected WFC (Byron 2005), but single parents experienced more WFC than parents who were married. For income, research has found that those with higher incomes have more WFC (Byron 2005; Michel et al. 2011). For women with caregiving responsibilities, our person-centered approach found that in combination, being married, having younger children, and having a higher income (or a job that has greater responsibilities which results in higher income) related to higher WFC. Future research should consider how organizations may best assist these overly burdened employees as nearly 66 million employees in the United States alone attend to a family member (Feinberg et al. 2011) and that number is expected to rise as family roles shift and advances in medicine increase life expectancy (DePasquale et al. 2017; Fox and Brenner 2012). Specifically targeting the Middle-age Woman with appropriate work-family policies and supports may help these women stay engaged in the workforce.

For caregiving men, the Low Income Man cluster had the lowest WIF and FIW (Wave 1 only). These groups had a similar number of children but the Low Income Man worked fewer hours (31.6 versus Middle-age Man which had 48.5 and Established Man which had 40.4), had a lower salary (87% had an income of $30,000 or less versus Middle-age Man and Established Man who had 77% of individuals making $50,001 or more), and had fewer married men (61.3% versus Middle-age Man of 93.1% married and Established Man of 84.6% married). It is important to note this cluster represented only 10% of the caregiving men in the sample. However, this finding supports the notion that working fewer hours relates to lower WIF, thus supporting scarcity theory (Marks 1977) and the view that spending less time at work allows for more time with family. Further, spending less time at work may be the reason this cluster earned a lower income. As this group had the lowest level of WIF, organizations may not need to invest in work family policies and supports for this type of caregiver. However, these findings amongst caregiving men align with caregiving women findings in that having fewer resources (i.e., lower income) related to lower WIF. Future research on WFC may want to put considerable thought on the demographic composition of their sample as these low income men and women with caregiving responsibilities seem to have lower WFC overall. In the future, if companies and researchers target more homogeneous samples, work-family policies and interventions may be more tailored and effective in easing the burden of WFC.

A benefit of this study is that data from three separate samples were used. Two samples represent a general population and the third used a time-lag design over two years amongst caregivers, which is a homogeneous population with more extreme role requirement in the family domain. This lends support toward external validity as similar results were found for the different samples used. It is important to note that while the number of clusters is relevant to the similarity of results between samples and future replications it is the substantive content of the clusters that should be focused on.

As with all research, there are limitations to this paper. For both studies, age and hours worked were measured in the same manner, but Study 1 included age of youngest child in categories (i.e., below 6, 6 to 12, 13 to 18, above 18, and no children) while Study 2 had the literal age of youngest child. Sample size may also be a limitation as the smallest cluster contained 31 men (Low Income Man; Study 2), while the largest cluster contained 160 men (Middle-age Man; Study 2). Cluster analysis is a technique that is data driven and the composition of different samples may yield different results. This is less problematic in this research as both studies relied on relatively large sample sizes. Further, WFC was measured with different measures (Carlson et al. 2000; Kelloway et al. 1999) which may be both a strength and weakness. It is possible that inconsistencies of results across the two studies may be due in part to different scales being used but it is also a strength that such similar findings were found despite this. As the labor market continues to change, this may result in different numbers of socio-demographic clusters. As such, future research should replicate our cluster approach using more recent samples. As similar numbers of clusters were found in our samples of a heterogeneous general and homogeneous caregiving population, this concern is slightly mitigated.

Future research may focus on how different types of homogeneous groups (e.g., single parents or dual-career couples) vary compared to general samples. It may be that when a specific population is accounted for, in this case caregivers, the amount of unique clusters created increases. When considering a more homogeneous sample, it is likely that differences found between clusters will be more meaningful. This may allow for more targeted, effective, and tailored work policies and supports to be created for a given population or company. It is also possible that when not using a person-centered approach, researchers or consultants may erroneously conclude that a work-family intervention doesn’t work when in fact, within a specific cluster or set of clusters, it may have.

Although the number of clusters created was similar, the information gained from different clusters was unique. Future research is needed as this study is exploratory in nature. There are many other socio-demographic variables that may also affect the work-life balance relationship, including marital status, living situation, race, ethnicity, and sexual orientation. It may also be advantageous to include a greater amount of clustering variables in future research to determine if new profiles emerge. Other than including a wider variety of socio-demographic variables, future research should consider utilizing variables with theoretically and empirically stronger relationships to WFC. By employing new techniques, previous research findings may be supported and a deeper understanding of the relationships between WFC variables, interventions, and outcomes may increase.

Conclusion

Using cluster analysis, demographically distinct groups were created. WFC differences were found between specific clusters but not all clusters. This ultimately supports the use of a combination of socio-demographic variables to separate and classify individuals in a statistically grounded and practically meaningful way. Cluster analysis may allow researchers and practitioners to create different profiles of employees affected by WFC. These profiles illustrate who experiences the greatest and least levels of WFC allowing for more targeted intervention programs to be created. If more targeted research on clusters is conducted, it is possible that a better understanding of the dynamics of WFC in various stages of the life and family cycles will be achieved. This may allow businesses to create better incentive packages to aid, attract, and retain future employees.

References

  1. Adkins, C. L., & Premeaux, S. F. (2012). Spending time: The impact of hours worked on work–family conflict. Journal of Vocational Behavior, 80(2), 380–389.CrossRefGoogle Scholar
  2. Allen, T. D., & Finkelstein, L. M. (2014). Work–family conflict among members of full-time dual-earner couples: An examination of family life stage, gender, and age. Journal of Occupational Health Psychology, 19(3), 376–384.CrossRefPubMedGoogle Scholar
  3. Andrews, A., & Bailyn, L. (1993). Segmentation and synergy: Two models linking work and family. In J. C. Hood (Ed.), Men, work, and family (pp. 262–275). Newbury Park: Sage.Google Scholar
  4. Baltes, B., & Young, L. M. (2007). Aging and work/family issues. In G. Adams & K. Shultz (Eds.), Aging and work in the 21st century (pp. 251–256). Mahwah: Lawrence Erlbaum.Google Scholar
  5. Barnett, R. C. (1998). Toward a review and reconceptualization of the work/family literature. Genetic, Social, and General Psychology Monographs, 124(2), 125.Google Scholar
  6. Barnett, R. C., & Hyde, J. S. (2001). Women, men, work, and family: An expansionist theory. American Psychologist, 56(10), 781.CrossRefPubMedGoogle Scholar
  7. Bauman, Z. (2001). The individualized society. Cambridge: Polity.Google Scholar
  8. Beck, U. (1992). Risk society: Towards a new modernity. London: Sage.Google Scholar
  9. Bianchi, S. M., Milkie, M. A., Sayer, L. C., & Robin-son, J. P. (2000). Is anyone doing the housework? U.S. trends and gender differentials in domestic labor. Social Forces, 79, 191–228.CrossRefGoogle Scholar
  10. Bianchi, S. M., Robinson, J. P., & Milkie, M. A. (2006). Changing rhythms of American family life. New York: Russell Sage Foundation.Google Scholar
  11. Block, J. (1971). Lives through time. Berkeley: Bancroft Books.Google Scholar
  12. Boushey, H. (2016). Finding time – The economics of work-life conflict. Cambridge: Harvard University Press.CrossRefGoogle Scholar
  13. Boyar, S. L., Maertz Jr., C. P., Pearson, A. W., & Keough, S. (2003). Work-family conflict: A model of linkages between work and family domain variables and turnover intentions. Journal of Managerial Issues, 175–190.Google Scholar
  14. Brough, P., & O’Driscoll, M. P. (2015). Integrating work and personal life. In R. J. Burke, K. M. Page, & C. L. Cooper (Eds.), Flourishing in life, work, and careers: Individual wellbeing and career experiences (pp. 377–394). Cheltenham: Edward Elgar Publishing.CrossRefGoogle Scholar
  15. Burke, R. J. (1988). Some antecedents of work-family conflict. Journal of Social Behavior & personality, 3(4), 287.Google Scholar
  16. Byron, K. (2005). A meta-analytic review of work–family conflict and its antecedents. Journal of Vocational Behavior, 67(2), 169–198.CrossRefGoogle Scholar
  17. 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(2), 249–276.CrossRefGoogle Scholar
  18. Craig, S. B., & Smith, J. A. (2000). Integrity and personality: A person-oriented investigation. Paper presented at the annual conference of the Society for Industrial Psychology, New Orleans.Google Scholar
  19. Crosby, F. (1991). Juggling: The unexpected advantages of balancing career and home for women and their families. New York: Free Press.Google Scholar
  20. Cullen, J. C., Hammer, L. B., Neal, M. B., & Sinclair, R. R. (2009). Development of a typology of dual-earner couples caring for children and aging parents. Journal of Family Issues, 30(4), 458–483.CrossRefGoogle Scholar
  21. DePasquale, N., Polenick, C. A., Davis, K. D., Moen, P., Hammer, L. B., & Almeida, D. M. (2017). The psychosocial implications of managing work and family caregiving roles: Gender differences among information technology professionals. Journal of Family Issues, 38(11), 1495–1519.Google Scholar
  22. Deuling, J. K., & Burns, L. (2017). Perfectionism and work-family conflict: Self-esteem and self-efficacy as mediator. Personality and Individual Differences, 116, 326–330.CrossRefGoogle Scholar
  23. Drobnic, S. (2011). Introduction: Job quality and work-life balance. In S. Drobnic & A. M. Guillen (Eds.), Work-life balance in Europe. The role of job quality (pp. 1–16). New York: Pelgrave and Macmillan.CrossRefGoogle Scholar
  24. Emslie, C., & Hunt, K. (2009). ‘Live to work’ or ‘work to live’? A qualitative study of gender and work–life balance among men and women in mid-life. Gender, Work and Organization, 16(1), 151–172.CrossRefGoogle Scholar
  25. Erikson, E. H. (1968). Identity: Youth and crisis (No. 7). WW Norton & Company.Google Scholar
  26. Evans, P., & Bartolomé, F. (1984). The changing pictures of the relationship between career and family. Journal of Organizational Behavior, 5(1), 9–21.CrossRefGoogle Scholar
  27. Feinberg, L., Reinhard, S. C., Houser, A., & Choula, R. (2011). Valuing the invaluable: 2011 update, the growing contributions and costs of family caregiving. Washington, DC: AARP Public Policy Institute, 32.Google Scholar
  28. Fox, S., & Brenner, J. (2012). Family caregivers online. Washington, DC: Pew Internet & American Life Project.Google Scholar
  29. Galinsky, E. (1999). Ask the children: What America's children really think about working parents. William Morrow and Company, Inc., 1350 Avenue of the Americas, New York, NY 10019 (US, $25; Canada, $37). Web site: http://www.familiesandwork.org.
  30. Galinsky, E., Bond, J. T., & Friedman, D. E. (1993). National study of the changing workforce. New York: Families and Work Institute.Google Scholar
  31. Ganster, D. C., & Bates, C. (2003). Do long work hours decrease general wellbeing and increase work-family conflict? In Annual Meeting of the Academy of Management, Seattle.Google Scholar
  32. Gillet, N., Morin, A. J., Cougot, B., & Gagné, M. (2017). Workaholism profiles: Associations with determinants, correlates, and outcomes. Journal of Occupational and Organizational Psychology, 90(4), 559–586.Google Scholar
  33. Greenhaus, J. H., & Beutell, N. J. (1985). Sources of conflict between work and family roles. Academy of Management Review, 10, 76–88.CrossRefGoogle Scholar
  34. Greenhaus, J. H., & Powell, G. N. (2006). When work and family are allies: A theory of work-family enrichment. Academy of Management Review, 31(1), 72–92.CrossRefGoogle Scholar
  35. Gutek, B. A., Searle, S., & Klepa, L. (1991). Rational versus gender role explanations for work–family conflict. Journal of Applied Psychology, 76, 560–568.CrossRefGoogle Scholar
  36. Halbesleben, J. R., Neveu, J. P., Paustian-Underdahl, S. C., & Westman, M. (2014). Getting to the “COR” understanding the role of resources in conservation of resources theory. Journal of Management, 40(5), 1334–1364.CrossRefGoogle Scholar
  37. Hill, E. J., Erickson, J. J., Fellows, K. J., Martinengo, G., & Allen, S. M. (2014). Work and family over the life course: Do older workers differ? Journal of Family and Economic Issues, 35(1), 1–13.CrossRefGoogle Scholar
  38. Hobfoll, S. E. (1989). Conservation of resources: A new attempt at conceptualizing stress. American Psychologist, 44(3), 513.CrossRefPubMedGoogle Scholar
  39. Hochschild, A. (1997). The time bind. Working USA, 1(2), 21–29.CrossRefGoogle Scholar
  40. Huffman, A., Culbertson, S. S., Henning, J. B., & Goh, A. (2013). Work-family conflict across the lifespan. Journal of Managerial Psychology, 28(7–8), 761–780.CrossRefGoogle Scholar
  41. Innstrand, S. T., Langballe, E. M., Falkum, E., Espnes, G. A., & Aasland, O. G. (2009). Gender-specific perceptions of four dimensions of the work/family interaction. Journal of Career Assessment, 17, 402–416.CrossRefGoogle Scholar
  42. Kaufman, G., & Uhlenberg, P. (2000). The influence of parenthood on the work effort of married men and women. Social Forces, 78, 931–949.CrossRefGoogle Scholar
  43. Kelloway, E. K., Gottlieb, B. H., & Barham, L. (1999). The source, nature, and direction of work and family conflict: A longitudinal investigation. Journal of Occupational Health Psychology, 4(4), 337–346.CrossRefPubMedGoogle Scholar
  44. Ketchen Jr., D. J., & Shook, C. L. (1996). The application of cluster analysis in strategic management research: An analysis and critique. Strategic Management Journal, 17, 441–458.CrossRefGoogle Scholar
  45. Korabik, K., Aycan, Z., & Ayman, R. (Eds.). (2017). The work-family interface in global context. Taylor & Francis.Google Scholar
  46. Kossek, E. E., Ruderman, M. N., Braddy, P. W., & Hannum, K. M. (2012). Work–nonwork boundary management profiles: A person-centered approach. Journal of Vocational Behavior, 81(1), 112–128.CrossRefGoogle Scholar
  47. Lavrakas, P. J. (1987). Telephone survey methods: Sampling, selection, and supervision. Sage Publications, Inc.Google Scholar
  48. Linville, P. W. (1987). Self-complexity as a cognitive buffer against stress-related illness and depression. Journal of Personality and Social Psychology, 52(4), 663–676.CrossRefPubMedGoogle Scholar
  49. Lyness, K. S., & Judiesch, M. K. (2008). Can a manager have a life and a career? International and multisource perspectives on work-life balance and career advancement potential. Journal of Applied Psychology, 93(4), 789.CrossRefPubMedGoogle Scholar
  50. Major, V. S., Klein, K. J., & Ehrhart, M. G. (2002). Work time, work interference with family, and psychological distress. Journal of Applied Psychology, 87, 427–436.CrossRefPubMedGoogle Scholar
  51. Matias, M., & Fontaine, A. M. (2015). Coping with work and family: How do dual-earners interact? Scandinavian Journal of Psychology, 56(2), 212–222.CrossRefPubMedGoogle Scholar
  52. Marks, S. R. (1977). Multiple roles and role strain: Some notes on human energy, time and commitment. American Sociological Review, 42, 921–936.CrossRefGoogle Scholar
  53. Martinengo, G., Jacob, J. I., & Hill, E. J. (2010). Gender and the work-family interface: Exploring differences across the family life course. Journal of Family Issues, 31(10), 1363–1390.CrossRefGoogle Scholar
  54. McDaniels, C., & Gysbers, N. C. (1992). Counseling for career development: Theories, resources, and practice. Jossey-Bass Publishers.Google Scholar
  55. Michel, J. S., Kotrba, L. M., Mitchelson, J. K., Clark, M. A., & Baltes, B. B. (2011). Antecedents of work–family conflict: A meta-analytic review. Journal of Organizational Behavior, 32(5), 689–725.CrossRefGoogle Scholar
  56. Mitchelson, J. K. (2009). Seeking the perfect balance: Perfectionism and work–family conflict. Journal of Occupational and Organizational Psychology, 82(2), 349–367.CrossRefGoogle Scholar
  57. Moen, P., Kelly, E., & Huang, Q. (2008). Work, family and life-course fit: Does control over work time matter? Journal of Vocational Behavior, 73(3), 414–425.CrossRefPubMedPubMedCentralGoogle Scholar
  58. Morin, A. J., Morizot, J., Boudrias, J. S., & Madore, I. (2011). A multifoci person-centered perspective on workplace affective commitment: A latent profile/factor mixture analysis. Organizational Research Methods, 14(1), 58–90.CrossRefGoogle Scholar
  59. Netemeyer, R. G., Boles, J. S., & McMurrian, R. (1996). Development and validation of work–family conflict and family–work conflict scales. Journal of Applied Psychology, 81(4), 400.CrossRefGoogle Scholar
  60. Ng, T. W., & Feldman, D. C. (2008). Long work hours: A social identity perspective on meta-analysis data. Journal of Organizational Behavior, 29(7), 853–880.CrossRefGoogle Scholar
  61. Ollier-Malaterre, A., & Foucreault, A. (2017). Cross-national work-life research: Cultural and structural impacts for individuals and organizations. Journal of Management, 43, 111–136.CrossRefGoogle Scholar
  62. Perry-Jenkins, M., & Wadsworth, S. M. (2013). Work and family through time and space: Revisiting old themes and charting new directions. In Handbook of Marriage and the Family (pp. 549–572). Springer, Boston, MA.Google Scholar
  63. Powell, G. N., & Greenhaus, J. H. (2010). Sex, gender, and decisions at the family→ work interface. Journal of Management, 36(4), 1011–1039.CrossRefGoogle Scholar
  64. Rajadhyaksha, U., Korabik, K., & Aycan, Z. (2015). Gender, gender-role ideology, and the work–family interface: A cross-cultural analysis. In Gender and the Work-Family Experience (pp. 99–117). Springer International Publishing.Google Scholar
  65. Rantanen, J., Kinnunen, U., Mauno, S., & Tement, S. (2013). Patterns of conflict and enrichment in work-family balance: A three-dimensional typology. Work and Stress, 27(2), 141–163.CrossRefGoogle Scholar
  66. Rospenda, K. M., Richman, J. A., Wolff, J. M., & Burke, L. A. (2013). Bullying victimization among college students: Negative consequences for alcohol use. Journal of Addictive Diseases, 32(4), 325–342.CrossRefPubMedGoogle Scholar
  67. Rothbard, N. P. (2001). Enriching or depleting? The dynamics of engagement in work and family roles. Administrative Science Quarterly, 46, 655–684.CrossRefGoogle Scholar
  68. Shockley, K. M., & Allen, T. D. (2013). Episodic work–family conflict, cardiovascular indicators, and social support: An experience sampling approach. Journal of Occupational Health Psychology, 18(3), 262–275.CrossRefPubMedGoogle Scholar
  69. Shockley, K. M., Shen, W., DeNunzio, M. M., Arvan, M. L., & Knudsen, E. A. (2017). Disentangling the relationship between gender and work–family conflict: An integration of theoretical perspectives using meta-analytic methods. The Journal of Applied Psychology, 102(12), 1601–1635.CrossRefPubMedGoogle Scholar
  70. SHRM Workplace Forecast. (2016). Future insights: Top trends affecting the workplace and the HR profession according to SHRM special expertise panels. Alexandria: Published by the Society for Human Resource Management.Google Scholar
  71. Sieber, S. D. (1974). Toward a theory of role accumulation. American Sociological Review, 39, 567–578.CrossRefGoogle Scholar
  72. Sinclair, R. R., Sommers, J. A., Cullen, J. C., & Wright, C. (2005). Performance differences among four organizational commitment profiles. Journal of Applied Psychology, 90, 1280–1287.CrossRefPubMedGoogle Scholar
  73. Sutton, K. L., & Noe, R. A. (2005). Family-friendly pro-grams and work–life integration: More myth than magic? In E. E. Kossek & S. J. Lambert (Eds.), Work and life integration: Organizational, cultural, and individual perspectives (pp. 151–169). London: Erlbaum.Google Scholar
  74. Taylor, B., & Dao, C. (2003). What kids really want that money can't buy. New York: Warner Books.Google Scholar
  75. U.S. Bureau of Labor Statistics. (2014a). Table A-7: Time spent in primary activities by married mothers and fathers with own household children under 18 by employment status of self and spouse and age of youngest child, average for the combined years 2007–11 [American Time Use Survey]. Retrieved from http://www.bls.gov/tus/tables/a7_0711.htm.
  76. U.S. Bureau of Labor Statistics. (2014b). American Time Use Survey- 2013 Results. Retrieved from http://www.bls.gov/news.release/pdf/atus.pdf.
  77. Ward, J. H. (1963). Hierarchical grouping to optimize an object function. Journal of the American Statistical Association, 66, 846–850.Google Scholar
  78. Wolff, J. M., Rospenda, K. M., Richman, J. A., Liu, L., & Milner, L. A. (2013). Work-family conflict and alcohol use: Examination of a moderated mediation model. Journal of Addictive Diseases, 32(1), 85–98.CrossRefPubMedPubMedCentralGoogle Scholar
  79. Wolff, J. M., Rospenda, K. M., & Richman, J. A. (2014). Age differences in the longitudinal relationship between work-family conflict and alcohol use. Journal of Addiction, 2014, 10.CrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG, part of Springer Nature 2018

Authors and Affiliations

  • Kyle J. Page
    • 1
  • Jacqueline K. Deuling
    • 1
  • Joseph J. Mazzola
    • 1
  • Kathleen M. Rospenda
    • 2
  1. 1.Department of I/O PsychologyRoosevelt UniversityChicagoUSA
  2. 2.Department of PsychiatryUniversity of Illinois-ChicagoChicagoUSA

Personalised recommendations