Understanding childcare satisfaction and its effect on workplace outcomes: The convenience factor and the mediating role of work-family conflict


Correspondence should be addressed to Stephanie C. Payne, Department of Psychology, Texas A&M University, 4235 TAMU, College Station, TX 77843-4235, USA (e-mail: scp@tamu.edu).


Building on Hobfoll's (1989, 2001) conservation of resources theory, we posit childcare is an essential resource to working parents. In addition to previously demonstrated childcare satisfaction (CCS) dimensions, we propose and demonstrate empirical support for a convenience dimension of CCS. Satisfaction with caregiver convenience refers to a parent's evaluation of the caregiver's location and availability. We hypothesize that time-related dimensions of CCS (caregiver dependability and convenience) relate to employee well-being and withdrawal, because they diminish time-based family interfering with work (FIW). We also propose quality-related CCS dimensions (caregiver attentiveness, communication, and cost) relate to psychological well-being, because they reduce strain-based FIW. Survey data from a sample of 316 university employees (faculty and staff) who were parents of under school-age children (infancy through preschool) revealed time-based FIW as an explanatory mechanism for the relationships between satisfaction with caregiver convenience and both turnover intentions and absenteeism (due to childcare issues). In addition, strain-based FIW mediated the effects of satisfaction with caregiver attentiveness on well-being and satisfaction with caregiver cost on well-being. This study expands previous research on CCS by demonstrating that CCS is related to important work outcomes in part because it reduces time- and strain-based FIW.


According to the US Census Bureau, in 2002, 55% of mothers with infant children and 72% of mothers without infants were part of the labour force. In addition, 62% of married couples with children are dual career (U.S. Bureau of Labor Statistics, 2007). It is therefore not surprising that in 2004 roughly two million children in the United States were cared for in one of 729,040 childcare centres across the country during the bulk of their parents’ working hours (U.S. Census Bureau, 2007). In addition, 704,000 children were enrolled in full-day childcare in 21,800 centres in England in 2005 (U.K. Office for National Statistics, 2007). This large number of working parents, as well as the number children cared for in childcare centres, illuminates the importance of reliable, good quality childcare. Without childcare, most parents are unable to work. Furthermore, if parents are not satisfied with their childcare arrangements, they may not be able to fully engage in their work. Correspondingly, employers should be concerned about adequate childcare and parents’ satisfaction with these arrangements (Goff, Mount, & Jamison, 1990; Kossek, 1990; Kossek & Nichol, 1992).

Work-family researchers have acknowledged the importance of child rearing (e.g., Byron, 2005), childcare (e.g., Milkovich & Gomez, 1976), and childcare satisfaction (CCS; e.g., Buffardi & Erdwins, 1997) on the work-family interface. For example, CCS has been associated with work-family conflict (WFC; Buffardi & Erdwins, 1997), job satisfaction (Goff et al., 1990), turnover intentions (Glass & Estes, 1996), and well-being (Ruben, 2006). In fact, some researchers have speculated that ‘what appears to matter more for role conflict perceptions is not so much the type of arrangement used but the satisfaction with the caregiving arrangements’ (Kossek, Colquitt, & Noe, 2001, p. 42). However, no research has attempted to empirically test why CCS is related to these outcomes. Based on conservation of resources theory (Hobfoll, 1989, 2001), we propose childcare is an essential resource for working parents and satisfaction with certain characteristics of childcare will relate to theoretically relevant outcomes. We further speculate that time- and strain-based family interference with work is important explanatory mechanisms for these relationships. The purpose of our study is to empirically test these propositions in order to gain a deeper understanding of why organizations should be concerned about their employees’ CCS.

Our research responds to calls in the literature to examine the processes linking the work and family domains (e.g., Lambert, 1990). Our study also responds to calls for an examination of both the directionality and form of WFC (Allen, Herst, Bruck, & Sutton, 2000; Ford, Heinen, & Langkamer, 2007), how family variables affect employee behaviours (Eby, Casper, Lockwood, Bordeaux, & Brinley, 2005), the multidimensionality of CCS (Buffardi & Erdwins, 1997; Poms, Botsford, Kaplan, Buffardi, & O’Brien, 2009), as well as the need for additional psychometric evidence for work-family measures (Casper, Eby, Bordeaux, Lockwood, & Lambert, 2007).

Childcare satisfaction

For dual-career parents, finding adequate and quality childcare can be quite challenging (Fuqua & Labensohn, 1986; Skinner, 1980). Even after childcare arrangements have been made, how parents feel about the arrangement can be a source of anxiety and strain. CCS is an evaluative judgment of one's caregiver and the conditions of the childcare arrangement. Bradbard and Endsley (1986) noted all childcare is not created equal and multiple factors are likely to affect CCS including type, quality, cost, and convenience.

Building on Bradbard and Endsley's (1986) work, Buffardi and Erdwins (1997) developed a 20-item measure of CCS that included three factors: caregiver communication, dependability, and attentiveness. Satisfaction with caregiver communication indicates parent's evaluation of the information provided about the child's day as well as how the parent feels about interactions with the caregiver. Satisfaction with caregiver dependability refers to parent's evaluation of the extent to which the caregiver is available on a regular basis to provide service. Satisfaction with caregiver attentiveness reflects a parent's evaluation of the amount of attention that a child receives from his or her caregiver and the number of other children that are also cared for. Buffardi and Erdwins (1997) demonstrated differential relationships between these dimensions and organizational commitment, job satisfaction, inter-role conflict, and maternal separations, further validating their distinction.

Taking an economic perspective, which emphasizes the costs associated with work and family demands (Drago & Golden, 2006), Poms et al. (2009) added a fourth factor to Buffardi and Erdwins’ (1997) CCS measure, caregiver cost, which taps parent satisfaction with financial considerations surrounding childcare. They further demonstrated support for a high-order model in which an overall CCS construct comprises four factors.

The one remaining factor identified by Bradbard and Endsley (1986) but not currently included in the CCS measure is convenience1. Satisfaction with caregiver convenience is a parent's evaluation of the location of the caregiver relative to home/work and the days and times childcare is available relative to the days and times the parent needs childcare. Like cost, convenience is also economic in nature because ‘time is money’ and transportation costs money. Previous research supports the importance of the location of childcare to the selection of a caregiver (Bagley, Woods, & Glatter, 2001; Folk & Yi, 1994). The institution of multiple ‘on-site’ childcare centres in which the centre is located within the same building or an adjacent building to the employing organization (Connelly, DeGraff, & Willis, 2004) conveys the perceived value of having childcare near one's place of work. Proximity of the childcare centre to the workplace determines how quickly the parent of a small child could get to the child in case of an emergency (Wolfe, 2007), to tend to special needs (e.g., breast feeding), to meet with the caregiver, or to visit with the child during the workday (e.g., share lunch); thus, convenience likely plays an important role in CCS. Correspondingly, we propose five factors of CCS.

  • Hypothesis 1CCS is a hierarchical multi-dimensional construct, consisting of five latent subfactors: attentiveness, communication, convenience, cost, and dependability.

Directionality and source of WFC

WFC is a form of inter-role conflict where pressures from the work and family domains are incompatible, and therefore participation in one role is made more difficult by participation in another (Kopelman, Greenhaus, & Connolly, 1983). Demands or preoccupations with either role domain (work or family) can cross the boundary of the other, resulting in WFC (Crouter, 1984; Frone, Russell, & Cooper, 1992; Premeau, Adkins, & Mossholder, 2007). Correspondingly, researchers conceptualize WFC as a bi-directional process, with the demands from the work domain spilling over into the home domain (work interfering with family; WIF) and the demands of the family domain spilling over into the work domain (family interfering with work; FIW). Meta-analytic research supports this distinction (Mesmer-Magnus & Viswesvaran, 2009). The source of the conflict related to the need for childcare while at work originates in the family domain (cf. Goff et al., 1990). Therefore, we predicted CCS will relate most strongly to FIW.

Work-family researchers also distinguish between three sources or forms of WFC: time based, strain based, and behaviour based (Greenhaus & Beutell, 1985). The viability of the three forms of conflict has been demonstrated in many studies (e.g., Carlson, Kacmar, & Williams, 2000; Stephens & Sommer, 1996). Theoretically, CCS is expected to relate to time- and strain-based WFC, so we limit our examination to these two sources. To date, researchers have not examined the time- and strain-based components of WFC separately with regard to CCS.

Time-based WFC

Involvement in multiple roles, each with their own demands, creates competition for an individual's time (Greenhaus & Beutell, 1985; Stephens & Sommer, 1996). Time demands from one role can make it physically impossible to participate in or meet the expectations of another role (Greenhaus & Beutell, 1985). In addition, pressures associated with one role can create a psychological preoccupation with that role even when an individual is physically trying to meet the demands of a second role (Greenhaus & Beutell, 1985).

Theoretically, the effects of time-based conflict can be explained by resource drain or a transfer of finite resources such as time, energy, or attention from one domain to another (Staines, 1980). According to the conservation of resources theory (Hobfoll, 1989, 2001), individuals seek to acquire and maintain resources, including objects, personal characteristics, conditions, and energies. Stress occurs when there is a loss, or a threat of loss, of resources. Using conservation of resources theory to explain WFC, Grandey and Cropanzano (1999) proposed that resources (e.g., time, energy) ‘are lost in the process of juggling both work and family roles’ (p. 352). We speculate that managing childcare responsibilities depletes time and energy resources thereby preventing a working parent from responding in the workplace with a full range of these resources.

We conceptualize childcare as a time-related resource that temporarily emancipates parents from childcare responsibilities and allows them to allocate that time to work. In fact, one of the 74 key resources in Hobfoll's (1988) measure, Conservation of Resources Evaluation, is ‘help with childcare’. We extend the conservation of resources theory, by proposing that the characteristics or dimensions of resources (e.g., quality) are particularly important to the perceptions of WFC. Hobfoll focuses particularly on quantity of resources discussing losses and gains and the sufficiency of resources. Many resources are prefaced with the word ‘adequate’ in his measure (e.g., ‘adequate clothing’, ‘adequate food’, ‘adequate income’). However, he also hints at quality for a few resources: ‘good marriage’ and ‘good relationship with my children’.

Borrowing further from conservation of resources theory, we propose that two time-related dimensions of childcare free time for working parents to devote to work. Unreliable and inconvenient care interferes with parents’ ability to work when and where they need to. Therefore, satisfaction with the time-oriented CCS dimensions is expected to negatively relate to time-based FIW.

  • Hypothesis 2Caregiver (1) dependability and (2) convenience are negatively related to time-based FIW.

Time-based FIW as a mediator of CCS – withdrawal relationships

Childcare is critical for employed parents to be able to physically go to work. Illness of a child or difficulties finding accessible and reliable childcare can all interfere with a parent's ability to attend work (Fuqua & Labensohn, 1986). As a result, time-based CCS is expected to relate to an employee's physical presence and time at work. In this study, we operationalize presence at work as the opposite of the withdrawal variables absenteeism (due to childcare issues) and turnover intentions.

Consistent with this theorizing, researchers have shown that problems with childcare arrangements relate to absenteeism (Kossek, 1990). Furthermore, Goff et al. (1990) found employees who were more satisfied with the quality of their child's care experienced less WFC that was negatively related to absenteeism. Likewise, Hammer and Nguyen (1995) revealed a significant relationship between CCS and absenteeism due to childcare. Previous research also supports a negative relationship between CCS and turnover intentions (Glass & Estes, 1996).

Building on conservation of resources theory, Buffardi and Erdwins’ (1997) speculation, and theoretical models of the work-family interface (Frone et al., 1992; Higgins, Duxbury, & Irving, 1992), we propose time-based FIW as an explanatory mechanism for the CCS withdrawal relationships. In other words, the reason why CCS is related to work withdrawal is because discontent over time-based CCS leads to time-based FIW, which in turn results in work withdrawal. In other words, dissatisfaction with the (in)convenience of childcare leads to higher levels of time-based FIW that employees seek to reduce by not physically attending work. Consistent with this theorizing, other researchers have proposed that when employees experience FIW they are likely to withdrawal in an effort to reduce the conflict (Boyar, Maertz, Pearson, & Keough, 2003; MacEwan & Barling, 1994). Previous research supports a negative relationship between FIW and absenteeism (Hammer, Bauer, & Grandey, 2003; MacEwan & Barling, 1994), as well as turnover intentions (Allen et al., 2000; Boyar et al., 2003). However, not all researchers have incorporated the time- and strain-based FIW distinction in their research. We propose that time-based FIW is an important mediator of time-based CCS and work withdrawal relationships as depicted in Figure 1.

Figure 1.

Hypothesized relationships.
Note. Family interfering with work (FIW).

  • Hypothesis 3Time-based FIW mediates the relationship between caregiver dependability and (1) absenteeism due to childcare issues and (2) turnover intentions.
  • Hypothesis 4Time-based FIW mediates the relationship between caregiver convenience and (1) absenteeism due to childcare issues and (2) turnover intentions.

Strain-based WFC

Strain-based WFC exists when a strain in one role affects an individual's performance in another role (Greenhaus & Beutell, 1985). Any work or family role can increase WFC. Strains can be either physical or psychological and hinder performance and/or the meeting of goals or expectations in another role (Stephens & Sommer, 1996).

Theoretically, strain-based WFC can be explained by spillover, or the idea that the influence of one domain on the other leads to similarities between the two domains (Zedeck, 1992). Spillover can take the form of affect, values, skills, and/or behaviours (Edwards & Rothbard, 2000). A number of researchers have speculated about and demonstrated empirical support for ‘psychological spillover’, or worry and preoccupation in one domain interfering with full participation or enjoyment in another domain (Carr, Boyar, & Gregory, 2008; Crouter, 1984). Control theorists refer to this as cognitive interference (Carver & Scheier, 1981), which can have adverse effects on attitudes and behaviours. Thus, in addition to physical presence, CCS is also expected to relate to an employees’ psychological presence at work.

The more satisfied employees are with their childcare, the less they will worry about their child during the day. As such, they should have lower levels of strain-based FIW. Specifically, we propose that quality-related dimensions of CCS will negatively relate to strain-based FIW. The demands of work can leave parents dissatisfied with the amount of childcare they can personally provide; as such, it is important for working parents to find attentive and invested caregivers (Anderson-Kulman & Paludi, 1986). We speculate that dissatisfaction with caregiver attentiveness is a source of strain for parents; this proposition is consistent with Buffardi and Erdwins (1997) who identified a relationship between satisfaction with caregiver attentiveness and reduced anxiety about separation from a child. It may lead parents to feel like they are failing in their role as a parent by not providing their own care or finding a suitable substitute (Crouter, 1984).

In addition to caregiver attentiveness, parents are likely to seek a caregiver who communicates with them and considers their input (Fuqua & Labensohn, 1986). Communication between the provider and parents helps parents feel informed and involved despite their physical absence. It also facilitates the parents’ ability to assess how well the caregiver is meeting the child's needs. Open communication likely reduces rumination about the child when the parents are at work, thus alleviating strain-based FIW.

Caregiver cost is also expected to relate negatively to strain-based FIW. High-quality childcare is expensive (Blau, 2001). As such, paying for this service can be a source of strain. When parents are satisfied with the cost of their childcare arrangement, they will likely experience less strain-based FIW, because they are less likely to worry about this expense (Carver & Scheier, 1981) or spend time trying to find a new childcare arrangement that they can afford.

  • Hypothesis 5Caregiver (1) attentiveness, (2) communication, and (3) cost are negatively related to strain-based FIW.

Strain-based FIW as a mediator of CCS – well-being relationships

Building further on conservation of resources theory, we propose three quality-related facets of CCS (attentiveness, communication, and cost) relate to two well-being outcomes (job satisfaction and parent well-being) because they reduce strain-based FIW. Previous research has demonstrated a positive relationship between CCS and job satisfaction (Buffardi & Erdwins, 1997; Erdwins, Buffardi, Casper, & O’Brien, 2001; Poms et al., 2009). Likewise, CCS has also been linked to decreased anxiety about separation from one's child (Buffardi & Erdwins, 1997; Erdwins et al., 2001) and decreased role strain (Van Meter & Agronow, 1982).

To explain these patterns of relationships, we argue that parents who are not satisfied with the quality of their childcare experience lower levels of job satisfaction and well-being because they have higher levels of strain-based FIW. In other words, discontent with caregiver attentiveness, communication, and/or cost results in higher levels of strain-based FIW which in turn leads to lower levels of job satisfaction and parental well-being.

Satisfaction with quality-related dimensions of CCS frees psychological resources that can then be applied to work (Rothbard, 2001). Any parents who have left their child screaming with a provider can empathize with the negative affect and worry that follows their departure. This strain-based FIW can linger with parents during their workday, resulting in distraction or a loss of perceived mastery of the family domain while at work (cf. Carr et al., 2008). Strain-based FIW can be interpreted as a personal failure, which people are less capable or unable to maintain the work boundary (Frone, Russell, & Cooper, 1997), letting family problems ‘spillover’ into the work domain. Furthermore, Premeau et al. (2007) posit that low-grade, persistent worry about balancing multiple roles, and the preoccupation with the work and family domains is taxing on a person, thus causing stress, strain, and dissatisfaction. Consistent with our theorizing, previous research supports negative relationships between strain-based FIW and job satisfaction (Carlson et al., 2000; Ford et al., 2007) as well as various forms of well-being (or lack thereof) including stress (Ford et al., 2007; Kelloway, Gottlieb, & Barham, 1999), negative physical symptoms (Netemeyer, Boles, & McMurrian, 1996), and physical and affective well-being (Lapierre & Allen, 2006).

  • Hypothesis 6Strain-based FIW mediates the relationship between caregiver attentiveness and (1) job satisfaction and (2) well-being.
  • Hypothesis 7Strain-based FIW mediates the relationship between caregiver communication and (1) job satisfaction and (2) well-being.
  • Hypothesis 8Strain-based FIW mediates the relationship between caregiver cost and (1) job satisfaction and (2) well-being.


Participants and procedure

Parents of under school-age children who were employees at a large southern university in the United States were recruited by e-mail to participate in a survey about childcare needs and satisfaction in the Fall 2008 semester. A total of 1,162 employees were identified by the Human Resources (HR) Department as employees who carried at least one child under school age on their insurance. The following demographic data for these employees were made available by HR: 46% women, 60% Caucasian, and 87% married. Following a cover e-mail encouraging participation by administration, an e-mail invitation with a link to an anonymous survey was sent to all of these employees. The e-mail invitation and link were also forwarded to various listservs of sponsoring organizations on campus (e.g., Women's Faculty Network) that likely broadened the potential population.

Four hundred and seven individuals responded to the survey for an approximate 35% response rate. Analyses were limited to university-employed respondents with under school-age children who completed the entire survey. Given that many couples work for the university, a careful screening based on spouse employer, sex of the respondent, and age of the youngest child revealed 11 pairs of responses that may have come from couples. To avoid including dependent data, one of each of these pairs of responses was randomly eliminated. For most analyses (with the exception of absenteeism due to childcare issues; n= 275), the final sample size was 316.

The majority of the respondents were female (75%), Caucasian (75%), and married or in a committed relationship (94%). Of those who were married/committed, most (85%) reported that their spouse/partner was also employed. Respondent ages ranged from 21 to 64 with a mean age of 35 years (SD= 6.25). On average, participants had worked 6.59 years (SD= 5.49) at the university in various capacities: 31% faculty, 57% staff, 10% graduate student, and 2% other (e.g., post doc). Respondents spent $500–$600 per month (modal response) on childcare for their youngest child and 18% used the university-supported childcare centre.


All measures were responded to on 5-point agreement scale (1 = strongly disagree, 5 = strongly agree), unless otherwise indicated. All coefficient alphas exceeded .70 (see Table 1).

Table 1.  Means, standard deviations, and correlations among variables
  1. Note. N= 275–316. FIW, family-interfering-with work; WIF, work-interfering-with family;

  2. a1 = male, 2 = female;

  3. b0 = not employed, 1 = employed;

  4. c1 = < $25k, 9 =$>200k with $25k increments in between; Reliabilities (Coefficient alphas) reported on the diagonal in parentheses. *p < .05; **p < .01.

 1. Sexa1.750.43                   
 2. Partner employmentb0.900.40.24**                  
 3. Faculty position0.260.44−.09.03                 
 4. Staff position0.570.50.07.03−.77**                
 5. Incomec4.**.48**−.28**               
 6. Organizational tenure6.595.49.06.00−.12*.20**.11              
 7. Time-based FIW2.851.06−.03.08.41**−.45**.17**−.15**(.86)             
 8. Strain-based FIW2.360.95−.04.01.13*−.15**−.08−.03.46**(.92)            
 9. Time-based WIF3.041.03−.11−.10.22**−.23**.08−.07.45**.15**(.85)           
10. Strain-based WIF3.001.02−.02−.05.18**−.17**.06−.08.45**.42**.50**(.87)          
11. Overall caregiver satisfaction4.180.87.01−.07.02−.01.07.15**−.18**−.22**−.20**−.25**         
12. Caregiver communication4.330.68−.03−.04−.*−.13*−.13*−.14*−.19**.70**(.86)        
13. Caregiver dependability4.370.62−.02−.07−.−.16**−.10−.20**−.21**.66**.74**(.75)       
14. Caregiver attentiveness4.160.83.03−.09−.*−.16**−.14*−.16**−.24**.69**.75**.65**(.82)      
15. Caregiver cost3.800.86.04−.*−.15*−.15*−.19**−.24**.49**.55**.58**.49**(.79)     
16. Caregiver convenience4.021.05−.10−.12.01−.02.12.11−.16**−.19**−.18**−.20**.50**.41**.54**.51**.45**(.96)    
17. Job satisfaction4.150.74−.02.07.06−.01.10−.07−.14*−.18**−.32**−.33**.14*.11.09.20**.14*.11(.89)   
18. Well-being3.490.81−.02−.01−.18**.23**.01.04−.44**−.51**−.42**−.57**.15**.10.10.15*.14*.14*.44**(.88)  
19. Turnover intention2.271.08.05−.07−.01.01−.07.07.18**.25**.30**.39**−.15**−.06−.06−.16**−.12−.12*−.68**−.45**(.73) 
20. Absenteeism3.682.13.18**.19**−.**.16*.11.21**−.15*−.13*−.22**−.13*−.17**−.22**−.07−.20**.12

Work-family conflict

Each of the four components of WFC was assessed with three items from Carlson et al.'s (2000) measure of WFC. For completeness, WIF was assessed and included in Table 1.

Childcare satisfaction2

We assessed CCS with the latest CCS Survey developed initially by Buffardi and Erdwins (1997) and subsequently revised by Poms et al. (2009). Eighteen items assessed five dimensions of CCS: attentiveness, communication, convenience, cost, and dependability (see Table 2). Caregiver convenience items were developed by Buffardi and Erdwins (1997). Respondents were asked the initial stem ‘Please take a moment to consider the child care arrangement for your youngest child. How satisfied do you feel with each of the aspects described below?’ In addition, respondents were asked to assess their overall CCS with the following item: ‘Considering everything, how satisfied are you with your childcare situation?’ Reponses were made on a 5-point satisfaction scale (1 = extremely dissatisfied, 5 = extremely satisfied).

Table 2.  Childcare satisfaction items
  1. *Note. We inadvertently left out one of Poms et al.'s cost items: ‘The flexibility your caregiver allows for delayed payment’.

Caregiver attentiveness
 The amount of direct attention/supervision given to your child.
 The number of other children cared for at the same time.
 Physical facilities or space in which your child stays (i.e., cleanliness, safety, facility size, and appropriate toys).
 The influence and interaction of other children with your child.
Caregiver communication
 Your caregiver's disciplinary style.
 The way your caregiver deals with meals/snacks.
 The relationship between you and your child's caregiver.
 Your caregiver's responsiveness to your suggestions about your child.
Caregiver convenience
 The location of the caregiver.
 The amount of time it adds to your commute to drop off your child at the caregiver's.
 Your ability to get your child quickly should an emergency arise.
Caregiver cost*
 Your caregiver's policy about charging for late pickup.
 Your caregiver's policy about charging when your child does not attend.
 The fees charged.
Caregiver dependability
 Your caregiver's dependability.
 The relationship between your child and your caregiver(s).
 The attitude of your caregiver toward flexible drop-off and pickup times.
 Your caregiver's policy about taking care of your sick child.

Job satisfaction

Three items adapted from Cammann, Fichman, Jenkins, and Klesh (1983) were used to assess overall job satisfaction. A sample item read: ‘All in all I am satisfied with my job’.


We adapted seven items from Goldberg's (1972) General Health Questionnaire to assess psychological well-being over the last month. We modified the items so they could be responded to on a 5-point agreement scale.

Childcare-related absenteeism

Respondents were asked the following two questions: ‘In the spring 2008 semester, approximately how many days of work were you absent because your child was sick and he/she could not be cared for by your regular provider because of his/her illness?’ and ‘In the spring 2008 semester, approximately how many days of work were you absent because of inadequate (lack of) childcare?’ Response options were N/A, never, less than 1 day, 1 day, 2 days, 3 days, 4 or more days. The latter five responses were summed across the two items for a total absence score due to childcare issues.

Turnover intentions

Two items adapted from Cammann et al. (1983) were used to assess turnover intentions. An example item read: ‘I will probably look for a new job in the next year’.

Control variables

We examined demographic variables that serve as proxies for family and work responsibilities (sex, age, marital status, income, partner employment status, number of children under school-age, position, and organizational tenure) as potential covariates.


Table 1 depicts the means, standard deviations, reliabilities, and correlations among the variables examined in the study. Consistent with Becker's (2005) recommendations, we controlled for demographic variables that were significantly related to the dependent variables in the corresponding analyses: sex, income, partner employment status, faculty status, staff status, and organizational tenure. Faculty and staff status variables are dummy codes for corresponding status (e.g., faculty = 1, non faculty = 0).

Hypothesis 1 proposed CCS as a hierarchical multi-dimensional construct, consisting of five latent subfactors. To test this hypothesis, we conducted a hierarchical confirmatory factor analysis in which the five CCS dimensions loaded onto one latent CCS factor. We compared the fit of those models to other theoretically feasible models including a one-factor model, a four-factor model in which dependability and convenience load on the same factor, a four-factor model in which the cost and convenience items all load on the same factor, and a five-factor model. A five-factor model for CCS was supported, as the chi-squared difference test showed a significantly better fit for the five factor as compared to the one- and four-factor models (χ2(10) = 632.34, p < .01; χ2(4) = 226.32, p < .01; and χ2(4) = 355.44, p < .01, respectively). As depicted in Table 3, the hierarchical model also fit the data equally well, supporting Hypothesis 1.

Table 3.  Childcare satisfaction confirmatory factor analyses results
  1. Note. N= 331. df = df for χ2; p = significance ofχ2. RMSEA, root mean square error of approximation; NNFI, non-normed fit index; CFI, comparative fit index; SRMR, standardized root mean square residual.

One factor963.97119.00.15.680.720.09
Four factor (cost + convenience)557.95113.00.11.820.850.12
Four factor (dependability + convenience)687.07113.00.13.770.810.12
Five factor331.63109.00.08.910.930.06
Hierarchical: five to one363.14114.00.08.900.920.07

Hypotheses concerning time-based FIW

All remaining hypotheses were tested with hierarchical regressions. In Hypothesis 2, we proposed that caregiver (1) dependability and (2) convenience are negatively related to time-based FIW. After controlling for organizational tenure, faculty status, staff status, and income, caregiver convenience (β=−.15, p < .01) accounted for a significant amount of variance in time-based FIW but caregiver dependability did not (β=−.08, ns), providing support for Hypothesis 2b, but not 2a.

The next two hypotheses proposed time-based FIW as a mediator of the relationships between CCS (dependability, H3; and convenience, H4) and withdrawal (absenteeism due to childcare issues, H3a and H4a; and turnover intentions, H3b and H4b). Hypothesis 3a and 3b were not pursued because caregiver dependability did not relate significantly to time-based FIW.

Hypotheses 4a and 4b were tested using Baron and Kenny's (1986) recommended procedure. The first step, regressing the mediator on the independent variable, was supported when testing Hypothesis 2b, which demonstrated that caregiver convenience related to time-based FIW. The second step involved regressing the dependent variables [absenteeism (β= .26, p < .05) and turnover intentions (β= .17, p < .05)] on the mediator (time-based FIW conflict), controlling for relevant covariates (sex and partner employment status for absenteeism). These results were significant. The third step involves independently regressing the dependent variables (absenteeism and turnover intentions) on the independent variable (caregiver convenience). Caregiver convenience related significantly to absenteeism (β=−0.18, p < .05) and turnover intentions (β=−0.12, p < .05). When time-based FIW was added to the model, caregiver convenience became nonsignificant with both absenteeism (β=−0.12, p > .05) and turnover intentions (β=−0.09 p > .05). The Sobel (1982) test indicated the decreases for absenteeism (z=−2.28, p < .05) and turnover intentions were significant (z=−2.00, p < .05). Thus, Hypothesis 4a and 4b were both supported.

Hypotheses concerning strain-based FIW

Our next set of hypotheses concerned strain-based FIW. First, we proposed caregiver (1) attentiveness, (2) communication, and (3) cost would be negatively related to strain-based FIW. Controlling for staff status because it related significantly to strain-based FIW, we confirmed that caregiver cost (β=−0.15, p < .05) and attentiveness (β=−0.12, p < .05) were both significantly related to strain-based FIW, but caregiver communication was not (β=−0.11, ns), providing support for Hypotheses 5a and 5c, but not Hypotheses 5b.

Hypotheses 6, 7, and 8 proposed strain-based FIW conflict as a mediator of the CCS well-being relationships. Hypothesis 7 was not supported given the lack of a significant relationship between caregiver communication and strain-based FIW as demonstrated when testing Hypothesis 5b.

Testing Hypothesis 6a, job satisfaction was significantly related to strain-based FIW (β=−0.18, p < .05) and caregiver attentiveness (β= 0.20, p < .05). When strain-based FIW was added to this equation predicting job satisfaction, caregiver attentiveness remained significant (β= 0.18, p < .05). Thus, Hypothesis 6a was not supported. The Sobel test was conducted to assess if partial mediation was occurring; this test indicated that the decrease was not significant (z= 1.63, p= 0.10) thus indicating that there is not partial mediation.

Testing Hypothesis 6b, well-being was also significantly related to both strain-based FIW (β=−0.48, p < .05) and caregiver attentiveness (β= 0.13, p < .05). In adding strain-based FIW to the equation predicting well-being, caregiver attentiveness was no longer significant (β= 0.08, ns). The Sobel test indicated that this decrease is significant (z= 2.06, p < .05), supporting Hypothesis 6b.

Testing Hypothesis 8a, job satisfaction was also significantly related to caregiver cost (β= 0.14, p < .05). When strain-based FIW was added to the equation predicting job satisfaction, caregiver cost remained significant (β= 0.12, p < .05). The Sobel test also indicated that the change was not significant (z=−1.90, ns). Thus, Hypothesis 8a was not supported.

Testing Hypothesis 8b, well-being was significantly related to caregiver cost (β= 0.15, p < .05). When strain-based FIW was added to the equation predicting well-being, caregiver cost was no longer significant (β= .09, ns). The Sobel test indicated that the change was significant (z=−2.37, p < .05), supporting Hypothesis 8b.

Exploratory analyses

Although work-family researchers have differentiated between time- and strain-based FIW, when one experiences time-based conflict, they are also likely to experience strain-based conflict as supported by their strong positive relationship with one another (r= .46). In order to examine discriminant validity for the different dimensions of CCS and types of FIW, additional mediation relationships were examined. Mediated relationships emerged for caregiver convenience and cost only, providing some evidence for discriminant validity3.


The purpose of this study was to determine the importance and utility of a convenience dimension of CCS and the extent to which time- and strain-based FIW, theoretically linked to time- and quality-related dimensions of CCS, explain relationships between CCS and four work-related outcomes. Building on Hobfoll's (1989, 2001) conservation of resources theory, we conceptualize childcare as a crucial resource for working parents that they must strive to maintain. Extending early work on the multidimensionality of CCS (Bradbard & Endsley, 1986; Buffardi & Erdwins, 1997), we propose that CCS has both time- (caregiver dependability and convenience) and quality-related (caregiver attentiveness, communication, and cost) dimensions. Consistent with our theorizing, time- and strain-based FIW help to explain why CCS dimensions relate to important work outcomes.

Based on Hobfoll's (1989) conservation of resource theory, we proposed time-related CCS is positively related to an employees’ physical presence at work and thus negatively related to absenteeism due to childcare issues and turnover intentions because time-related CCS reduces time-based FIW. We demonstrated that caregiver convenience was significantly related to turnover intentions through both time- and strain-based FIW and with absenteeism through time-based FIW. Caregiver convenience also related to employee well-being through both sources of conflict. Thus, the effects of time-related CCS appear to extend beyond time-related strain and outcomes. In other words, a lack of satisfaction with childcare convenience is associated with multiple sources of FIW and outcomes.

We also proposed CCS is positively related to employees’ psychological presence at work and overall well-being because it diminishes strain-based FIW. We found caregiver attentiveness and cost related to well-being through both forms of conflict and caregiver cost related to absenteeism through time-based conflict. Thus, quality-related dimensions of CCS are primarily associated with employee well-being but not exclusively through strain-based conflict. Interestingly, satisfaction with caregiver communication did not emerge as a critical factor in the relationships examined.

Although we replicated a positive relationship between CCS (caregiver attentiveness, cost, and overall) and job satisfaction (Buffardi & Erdwins, 1997), a lack of FIW did not explain this relationship. Additional research is needed to identify possible explanatory mechanisms for this relationship, perhaps work-family enrichment. Work-family enrichment theory suggests affect spills over from one domain to the next and fosters relationships between work and family variables (Greenhaus & Powell, 2006). Alternatively, because work allows employees access to better quality childcare by earning a salary or receiving job benefits such as priority to access or a subsidy for childcare costs, a lack of WIF may explain this relationship and suggest an alternative causal ordering in which job satisfaction leads to CCS. However, this is more likely to be the case for employees who take advantage of employer supported childcare and only 18% of our sample did so.

Working parents have very limited excess resources (e.g., time), thus whenever they can ‘save time’ or avoid expending extra time that they do not have to (e.g., commuting to a childcare centre further away), they are likely to want to take advantage of it. Convenient childcare is when and where parents need it the most. The whole notion of on-site childcare is built on the merits of convenience. Thus, parents who are satisfied with the convenience of their caregiver are able to conserve time and effort resources that parents with inconvenient childcare cannot. Convenient childcare may also allay parents’ concerns or worry associated with being separated from children, because they know they can get to their children quickly if they become distressed (e.g., asthma attack) or if there is an emergency.

Directionality and source of WFC

In response to calls for an examination of both the directionality and form of WFC (Allen et al., 2000; Ford et al., 2007), we examined how CCS related to time- and strain-based FIW, as well as WIF. We expected CCS to be more strongly related to FIW than WIF because childcare is a family-related resource; however, stronger correlations emerged with WIF (cf. Poms et al., 2009). Thus, our results may provide some insight into how parents assess the source of a childcare-related conflict, which has implications for the perceived directionality of WFC and correlations with corresponding constructs (FIW and WIF). Greenhaus and Beutell (1985) proposed ‘directionality of work-family conflict is perceived only after a response to the conflict is made’ (p. 84). Interestingly, one of the items in Frone et al.'s (1992) WIF scale reads: ‘How often does your job or career interfere with your responsibilities at home, such as yard work, cooking, … or childcare?’ Thus, work can be perceived as the source of a problem regarding childcare needs, because it prevents parents from being able to care for their children. Furthermore, for employees who utilize an employer-supported childcare centre, childcare is a work-related resource. Childcare could be construed as a resource that facilitates participation in multiple roles and thus bridges the work-family interface. In short, directionality of WFC when work and childcare overlap is determined by what the individual perceives as the source that may vary from person to person and likely relates to individual differences like role salience (Greenhaus & Beutell, 1985).

Our results also provide some insight and implications for the perceived source or form of WFC. Although we hypothesized time-based FIW would mediate relationships between time-oriented CCS dimensions and physical presence at work, our exploratory analyses revealed that strain-based FIW mediated the same significant relationships with turnover intentions (see Table 4). Two exceptions, both with absenteeism due to childcare issues, emerged where only time-based FIW mediated these relationships. Likewise, we hypothesized strain-based FIW would mediate relationships between quality-related CCS dimensions and psychological well-being; yet, time-based FIW also mediated the relationships that were significant for time-based FIW. Thus, it seems that both sources of conflict explain why CCS is related to these outcomes. It may be that time-based conflict quickly leads to strain-based conflict and it is difficult to tease apart these mechanisms when reflecting on long time periods. Shorter time intervals or reporting of specific conflicts may reveal differential relationships between the sources of conflict. It is also important to note that absenteeism could impact satisfaction with childcare. Understandably, many providers will not care for sick children. Thus, parents have to miss work when their children are ill. When this happens it may result in reduced CCS.

Table 4.  Significant mediations
 1.Satisfaction with caregiver convenience -> time-based FIW -> absence (H4a)
 2.Satisfaction with caregiver convenience -> time-based FIW -> turnover intentions (H4b)
 3.Satisfaction with caregiver convenience -> time-based FIW -> well-being
 4.Satisfaction with caregiver convenience -> strain-based FIW -> turnover intentions
 5.Satisfaction with caregiver convenience -> strain-based FIW -> well-being
 6.Satisfaction with caregiver attentiveness -> time-based FIW -> well-being
 7.Satisfaction with caregiver attentiveness -> strain-based FIW -> well-being (H6b)
 8.Satisfaction with caregiver cost -> time-based FIW -> absence
 9.Satisfaction with caregiver cost -> time-based FIW -> well-being
10.Satisfaction with caregiver cost -> strain-based FIW -> well-being (H8b)

Consistent with predictions and previous research (Hammer & Nguyen, 1995), CCS was negatively related to absenteeism related to childcare issues. Eby et al. (2005) noted that ‘research has consistently failed to find a direct relationship between the use of on-site childcare and absenteeism’ (p. 152). Again, this illustrates the importance of the characteristics of the resource and the affect employees have for that resource (i.e., satisfaction) in predicting important work outcomes. Like other researchers (Kossek & Nichol, 1992; Mize & Freeman, 1989), we limited our absenteeism criteria to absences due to childcare issues. Future researchers may want to collect more specific absenteeism criteria as well. Further exploratory analyses revealed that this relationship was largely driven by the absenteeism item about a sick child (rather than unreliable/lack of care). We inadvertently truncated the absenteeism response scale with the highest response option being ‘four or more absences’. Given the large number of respondents who chose this response option (28 for lack of childcare and 101 for illness), we clearly underestimated the number of absences that occurred. An open-ended continuous measure would have captured more variance.

Theoretical and practical implications

We extend conservation of resources theory by proposing the importance of the individual's evaluative judgment of the quality of the resource. We apply this to childcare; however, any resource has both quantity and quality dimensions. For example, Hobfoll (1988) lists ‘time with loved ones’ and ‘medical insurance’ as important resources. Certainly more is better; however, higher quality is also better. When studying specific resources, researchers may want to consider measuring both quantity and quality.

We also speculated that CCS relates to strain-based FIW, because when parents are not satisfied with their childcare arrangements they are likely to ruminate about their child and this negative emotional spillover causes cognitive interference. We operationalized this interference with assessments of FIW conflict and found mixed support. Future research is needed to confirm that a lack of CCS leads to rumination, negative affect, and/or cognitive interference.

In terms of practical implications of our research, childcare providers can use this information to support the need for organizations to partner with them to establish employer-sponsored/supported and even subsidized care, as well as on-site childcare centres. This partnership could be especially important given the influence of the convenience factor. An on-site childcare arrangement would make transportation to the childcare centre and getting to a child much easier and less stressful for parents. In addition, organizations benefit when parents have a convenient childcare arrangement because of the decreases in employee absenteeism and turnover intentions. Furthermore, resources saved from commuting could potentially be devoted to work activities.

Our results can also be used to justify to organizational management the need to allocate resources towards high-quality and convenient childcare solutions for their employees given relationships with important work-related outcomes. Such solutions are not exclusively on-site childcare arrangements as the pattern of results that emerged was not limited to on-site childcare users. In fact, only 18% of our sample utilized the university-supported centre. This is consistent with Kossek et al.'s (2001) observation that CCS is more important to role conflict perceptions (and outcomes) than the specific provider. Additional solutions include providing information about nearby childcare providers, establishing partnerships with these providers to give priority and/or discounts to their employees, and childcare subsidies.

A comprehensive CCS measure articulates important dimensions of childcare that would be of interest to multiple stakeholders. This includes parents selecting a childcare provider and making comparisons between multiple providers. The dimensions are likely to be of interest to childcare providers interested in evaluating their services, public agencies concerned with training childcare providers, and providing childcare referral services to parents. CCS dimensions are also likely to be of interest to employers seeking to partner with a childcare vendor or as a quality-control measure to ensure minimum standards are met.

Limitations and directions for future research

Whereas it would have been preferable to have gathered some of our data from alternative sources (e.g., absenteeism), university administrators sought frank responses and a high response rate. Therefore, they stipulated the survey had to be anonymous that eliminates the possibility of linking survey responses to other sources of data. Furthermore, we were particularly interested in absenteeism due to childcare issues and organizational records of absenteeism are unlikely to comprehensively log the reasons for absences and unlikely to exist for faculty (31% of our sample). We further contend that self-reports are the most appropriate way to measure individual perceptions (e.g., well-being) and subjective evaluations (WFC, job satisfaction, CCS). Consistent with Podsakoff, MacKenzie, Lee, and Podsakoff's (2003) suggestions, we included measures with different response scales (agreement and satisfaction), which should help to reduce common method bias.

One additional limitation was that the measurement of absenteeism for the previous semester was inconsistent with our theoretical model. Absenteeism data were for the spring semester thus occurring prior to reports of CCS. However, given university administration's desire for an anonymous survey, we could not link survey responses to future absenteeism data. Furthermore, the validity of self-reported recall measure depends on participants accurately recalling the number of days missed. Our sample consisted of 57% staff, who are required to log leave time online, which may have facilitated accurate recall.

Additional research is needed to determine the generalizability of our findings. One interesting avenue to pursue might be elder care satisfaction and the extent to which similar relationships emerge for this construct. Future research is also needed on the conditions that exacerbate or ameliorate these relationships. Some potential moderators include the type of caregiver (Kossek, Pichler, Meece, & Barratt, 2008), demographic variables, as well as individual differences like work and family salience (Greenhaus & Powell, 2003) and dispositional tendencies.

In summary, our research extends previous research examining the relationship between CCS and WFC using the conservation of resources theory. Results supported caregiver convenience as a fifth factor of CCS as well as time-based FIW as an explanatory mechanism for the effect of satisfaction with caregiver convenience on turnover intentions and absenteeism due to childcare issues. Strain-based FIW was shown to be an explanatory mechanism for the effect of satisfaction with caregiver attentiveness and cost on employee well-being. These results support WFC as a linking mechanism between work and family. Additional research is needed to determine the generalizability of these findings and to determine the causal pattern of these relationships.


  • 1

    Buffardi and Erdwins (1997) identified a factor with two items, one regarding cost and one about location inconvenience (temporal cost), but because of low alphas in two samples, these items were removed from the CCS measure.

  • 2

    At the end of the survey, participants were given the opportunity to provide comments with the following vague prompt: ‘Please feel free to share any additional comments that you have about childcare or this survey’. Thirteen of 44 (30%) comments explicitly about CCS dimensions concerned convenience [i.e., when (day, time, and duration) and where (location)].

  • 3

    Additional mediation tests examined all possible independent variable (five dimensions of CCS), mediator (time- and strain-based FIW), and dependent variable relationships. There were no significant mediated relationships for caregiver communication, caregiver dependability, and no new mediated relationships for caregiver attentiveness. Mediated relationships emerged for caregiver convenience and cost (see Table 4).


The authors would like to thank Sherry Yennello and Jackie Caruso for facilitating the administration of this survey and Ann Huffman for her comments on an earlier version of this paper.