Introduction
Over the past fifty years, work has grown more precarious in terms of wages and benefits (Fligstein and Shin 2004; Hacker 2006; Kalleberg 2011). But, so too has another crucial dimension of work, especially for low-wage workers: work time and work schedules. Rather than a regular 40 hour work-week once envisioned by Labor, work schedules, especially in the service-sector, are often distinguished by insufficiency in the number of hours, non-standard timing in when those hours are scheduled, variability in the timing and number of hours, a lack of predictability over when shifts will occur, and limited worker control over shift timing (Carré and Tilly 2017; Golden 2015; Lambert 2008; Schneider and Harknett 2019a).
A large and prominent literature has found that these features of work scheduling are associated with key measures of adult worker well-being. Early work examined how the duration of work hours shaped such worker outcomes, with a particular focus on part-time work on the one hand (Bardasi and Francesconi 2004; Kauhanen and Nätti 2015; van Rijswijk et al. 2004) and over-work on the other (Kivimäki et al. 2015; Sparks et al. 1997; Virtanen and Kivimäki 2018). A parallel line of literature has provided a comprehensive examination of how non-standard work hours, such as night shifts, negatively affect worker well-being and family processes (Han 2005; Han, Miller, and Waldfogel 2010; Presser 1999; Presser 2000). Especially for white-collar workers, an influential literature has examined work schedule control and flexibility (Kelly and Moen 2007; Kelly, Moen, and Tranby 2011; Kim et al. 2020). Since the passage of new laws that seek to regularize advance notice of schedules and predictability, there has been an outpouring of work, led by Susan Lambert and colleagues, that examines the role of unpredictability and variability in shaping worker outcomes (Henly and Lambert 2014; Lambert and Henly 2014; Lambert, Henly, and Kim 2019; Schneider and Harknett 2019a; Pilarz, Cuesta, and Drazen 2020; Schneider and Harknett 2021; Lambert and Fugiel 2023; Cho et al. 2024).
While much of this literature has focused on workers’ own financial security, health, well-being, and job-related outcomes, an important strand of this work has also examined the consequences of parental exposure to scheduling challenges for child well-being and parenting outcomes. As with research on worker outcomes, this scholarship has generally examined how specific dimensions of parental work schedule quality are associated with parenting-related and child outcomes, in isolation. For instance, one set of papers examines non-standard timing and parental stress (Hwang and Jung 2020), child behavior and development problems (Gassman-Pines 2011; Joshi and Bogen 2007; Li et al. 2014), and parenting time (Connelly and Kimmel 2011; Craig and Powell 2012; Pilarz and Awkward-Rich 2024; Rapoport and Le Bourdais 2008), but does not examine variability, predictability, or control. Another line of more recent work examines how variability and predictability shape parental well-being and childcare (i.e. Ananat and Gassman-Pines 2021; Harknett, Schneider, and Luhr 2022), but not timing or duration.
We advance this important body of existing research by taking a multi-dimensional approach to examining the consequences of exposure to poor quality work schedules on parenting outcomes. Our approach responds to prominent recent research which points to the need to consider the multi-dimensional aspects of work scheduling comprehensively (Lambert and Fugiel 2023). This work identifies five dimensions of work schedule quality: duration, timing, variability, predictability, and control (Lambert and Fugiel 2023). We operationalize these measures and examine their associations with two key outcomes: parental stress and parents’ time with children. These outcomes are important since they can have long-term effects on both parents and children, shaping, for instance behavioral and developmental outcomes in children (Joshi and Bogen 2007; Neece, Green, and Baker 2012), children’s physical health (Fabricius et al. 2012), and parents’ mental health (Puff and Renk 2014; Quittner, Glueckauf, and Jackson 1990). While these dimensions of schedule quality are often grouped together under the broad phenomenon of “work schedule instability” or “nonstandard scheduling” and while these dimensions are often co-occurrent, these different dimensions may lead to distinct consequences for parents and their children and may matter differently for parental stress and for parenting time. By examining all of these dimensions at once, using a common set of outcomes and common analytical strategy, we provide new insight into how these different dimensions of schedule shape important parenting outcomes.
In doing so, we advance a small number of papers that have taken a multi-dimensional approach to examining scheduling on adult and parenting outcomes, though rarely specifically parental stress and/or parenting time. Specifically, Walther and Pilarz (2024) make a key contribution to this literature by examining how schedule timing, work hour variability, work schedule predictability, and control are associated with child externalizing behaviors. Supporting the perspective that a multi-dimensional perspective is valuable, they find significant heterogeneity in how these dimensions are related to child behavior. However, their work differs from ours in that they do not directly examine parental stress or parenting time and only generalize to parents in Illinois and New York who received child-care subsidies. In contrast, Luhr, Schneider, and Harknett (2022) examine parental stress using a national sample of service-sector workers, but only for three of the dimensions of work scheduling and only for mothers. Similarly, Pilarz et al. (2020) analyze parental time using data from the Future of Families and Child Well-being Study, but only for three work scheduling dimensions and only for fathers, examining differences by co-residence status. Work by Lambert et al. (2019) and Lambert and Fugiel (2023) examines all five dimensions of work scheduling using national samples, but focuses on adult worker’s own outcomes, not parenting, similarly to Schneider and Harknett (2019a), who examine four dimensions, but only for worker’s own health outcomes.
We take up this multi-dimensional examination of work schedule quality and parental stress and parental time with children in the context of the low-wage US service sector, focusing on hourly workers in food service and retail. One in ten children has at least one parent employed in the sector (Harknett et al. 2022) and the sector contains the single largest concentration of low-wage workers (Osterman and Shulman 2011). While challenging work schedules are found across sectors, levels of work hour insufficiency, non-standard timing, variability, low predictability, and limited control are particularly high in the service sector (Lambert, Fugiel, and Henly 2014). In recognition of the challenges faced by workers in this sector, recent legislation that seeks to regulate work scheduling practices generally focuses specifically on large firms in retail and food service (Wolfe, Jones, and Cooper 2018). These laws contain multiple provisions that govern work hour duration, timing, predictability, and control, making the joint analysis of these dimensions valuable for future policy making.
In addition to estimating the direct associations between these distinct dimensions of work scheduling and parenting outcomes, we build on promising earlier work by Lambert et al. (2019) and Cho et al. (2024), which highlights how schedule control and predictability shape the effects of other work scheduling dimensions, including variability in work hours, non-standard timing, and short notice, on key outcomes for adult workers. Expanding on this seminal work, we examine how these distinct dimensions of work scheduling may interact amongst working parents, specifically focusing on how control over scheduled hours may moderate any negative associations between non-standard timing, variability, and predictability and parenting outcomes.
Further, the association between these dimensions of work schedule quality and parental stress and parenting time may vary by parent gender (Renk et al. 2003; Sayer, Bianchi, and Robinson 2004; Scott and Alwin 1989; Simon 1992), co-residential status (Boll and Schüller 2023; van der Heijden, Poortman, and van der Lippe 2016), and child age (Larson and Richards 1991; Putnick et al. 2010; Small, Eastman, and Cornelius 1988) and we are attentive to such heterogeneity.
Drawing on detailed measures of these scheduling dimensions alongside variables capturing parental stress and parenting time outcomes from a large sample of parents working hourly in the service sector collected by the Shift Project, we find a complex set of associations that, while consistent with theory, highlights the distinct way in which different dimensions of work scheduling are associated with these two key parenting outcomes. While duration, predictability, and control are significantly associated with both outcomes, timing only has a significant effect on parental stress, and variability only has a significant effect on parenting time. We also find evidence, albeit limited, that control over schedule timing moderates the relationship between non-standard timing and parental stress. The significant heterogeneity in impact across and within dimensions underscores the importance of a multidimensional conceptualization of work schedules.
Background
Parental Stress, Parenting Time and their Consequences
There is a large body of research that motivates the examination of parental stress and of parenting time with children as important aspects of the family process and the production of child well-being and ultimately attainment and mobility. The association between higher levels of parental stress and poorer child behavior and development is well-established in the literature (Joshi and Bogen 2007; Neece et al. 2012; Ward and Lee 2020). There is also evidence that parental stress plays an important role in mediating the relationship between childcare instability and behavioral problems in children, especially in low- income families (Pilarz and Hill 2017). Relatedly, Puff and Renk (2014) find that parental stress is a key mediator of the relationship between parents’ financial strain and children’s behavioral outcomes. Parental stress is also particularly salient amongst low-income parents and is associated with less effective parenting strategies and behaviors (Deater-Deckard 2004; Deater-Deckard and Panneton 2017). Other implications of parental stress include reduced sensitivity to children’s needs (Pelchat et al. 2003) and reduced parent involvement (Ponnet et al. 2013), both of which point to decreased quality and quantity of parenting time.
Parenting time, in terms of both quantity and quality, is highly stratified in the United States by socio-economic status (Kalil and Ryan 2020; Schneider, Hastings, and LaBriola 2018; Sherman and Harris 2012). More highly educated and higher-income parents are able to spend more time in developmentally beneficial activities such as reading, play, and academic skills enrichment (Kalil 2015; Kalil, Ryan, and Corey 2012) as well as more time in regular family meals (Musick and Meier 2012) than their more disadvantaged counterparts. Despite lower-educated and lower-income parents’ desire to engage in these developmentally important activities, financial and temporal constraints prevent them from doing so (Kalil and Ryan 2020). Kalil et al. (2012) find evidence of a “developmental gradient” in differences in childcare and play time across maternal education. These inequalities in parental engagement are consequential because parental developmental time is associated with effects on children’s physical and mental health (Fabricius et al. 2012; Han and Fox 2011; Han and Miller 2009). Parent-child interactions in early childhood, including reading and play, are extremely important for children’s literacy (Wood 2002). Fomby and Musick (2018) find persistent positive effects of total mother-child interaction time on children and adolescents’ behavioral and academic outcomes. Regular family meals are associated with improved nutritional health among children and adolescents (Hammons and Fiese 2011), as well as improved psychological well-being and lower incidence of high-risk behaviors among adolescents (Goldfarb, Tarver, and Sen 2014).
Variation in parental stress and in parenting time have been attributed to economic hardship and poverty (Deater-Deckard 2004; Kalil 2015), the magnitude of children’s needs (Pelchat et al. 2003), as well as to marital stress, family conflict, and lack of social support (Carrillo et al. 2017; Deater-Deckard and Panneton 2017; Ponnet et al. 2017). In addition, a number of mechanisms are likely to connect working conditions to both parental stress and parental time with children. First, the time availability model posits that the amount of time spent engaging in childcare depends on the total amount of time that parents have available (Haas 1999; Monna and Gauthier 2008) and time pressure has been identified as a key mechanism linking work-life conflict and poorer mental health (Roxburgh 2004). Thus, while limited time availability at home due to effort at work may functionally decrease time available for parenting, time-based strain may increase parental stress for workers. Second, the family stress model identifies economic insecurity as a significant source of pressure that results in emotional distress, depression, and stress among parents (Masarik and Conger 2017; Neppl, Senia, and Donnellan 2016; Shelleby 2018).
Work Schedules and the Production of Parental Stress and Parenting Time
While multiple aspects of working conditions may shape time availability, time-based strain, and family stress, prior research suggests that work schedules are likely to play an especially important role in these processes, ultimately affecting parental stress and parenting time. Work schedules comprise multiple dimensions that encompass not only objective aspects of schedules themselves but also the extent to which workers determine their own schedules. The primary, or objective, dimensions of scheduling are duration, timing, and variability, and the secondary dimensions of scheduling pertaining to worker power over the scheduling process are predictability, control, and congruence (Lambert and Fugiel 2023; Lambert and Henly 2014).
The extent to which work scheduling practices operate through channels such as time availability, time strain, and family stress to affect parental stress and parenting time likely depends on the specific dimension of work scheduling. Duration of work hours, for example, may functionally decrease time available for parenting, but may bolster economic security, thus reducing parental stress.
A very large and influential literature examines the association between these dimensions of work schedule quality and parenting outcomes, but generally does so by focusing on specific dimensions of work scheduling, with any one study rarely examining each of the dimensions simultaneously. We suspect this is because most existing data sets lack detailed measurement of all of these dimensions, because new aspects of work scheduling quality, such as predictability and variability, have emerged more recently, and because Lambert and Fugiel’s (2023) conceptual architecture is relatively new. We draw on this body of high-quality prior research, that generally treats these dimensions separately, to motivate how each dimension of work scheduling might affect parental stress and parental time with children. Where this literature review draws on studies that use many different sources of data, collected at different times, model differently measured parenting outcomes, and deploy quite varied modeling strategies, our simultaneous application of multiple dimensions of work schedules aims to provide a harmonized and comprehensive assessment of how each dimensions shapes the important outcomes of parental stress and parenting time.
Dimensions of Schedules: Duration
Duration, or the amount of time spent at work, could be negatively associated with time with children from a time-availability perspective. Meanwhile, insufficient hours may increase parental stress due to economic anxiety. Prior research finds that the amount of time that parents spend working outside the home is not associated with parental stress, independent of other factors like non-standard or irregular work schedules (Nomaguchi and Johnson 2016; Sanik 1993; Rönkä et al. 2017). However, number of hours could be related to parental stress in the context of underemployment. Involuntary part-time work due to insufficient hours, or underemployment, is a common issue in the service sector and is linked to poorer mental health and diminished financial well-being (Bell and Blanchflower 2019; Bell and Blanchflower 2021; Kim and Golden 2022). Working parents who do not get enough hours may have higher levels of parental stress due to increased economic insecurity.
On the other hand, parents who work more hours are likely to spend less time with their children since they have fewer hours available to do so. Prior literature focusing on maternal employment has found that mothers working full-time report spending considerably less time with their children compared to unemployed mothers (Kendig and Bianchi 2008). Work hours have been strongly linked to feelings of time strain and reduced time with children amongst mothers and fathers (Milkie et al. 2004; Pilarz and Awkward-Rich 2024; Yeung et al. 2001).
Dimensions of Schedules: Timing
Out of the dimensions of scheduling identified by Lambert and Fugiel (2023), timing has been explored the most in relation to parental stress and parenting time. Much of this literature is focused specifically on parental exposure to non-standard work schedules – regularly scheduled shifts that occur outside of 8–4 or 9–5 hours, such as night, evening, or weekend shifts, particularly regularly scheduled shifts that occur outside of 8–4 or 9–5 hours. Parental depressive symptoms and reduced parent-child interaction have been found to partially mediate the relationship between parental non-standard work hours and child well-being (Li et al. 2014). Prior work on non-standard work schedules and parental stress that focuses on mothers has linked nighttime, weekend, and rotating shifts to higher levels of parental stress (Joshi and Bogen 2007) and has identified non-standard work hours as a moderator of the relationship between work-family conflict and parental stress (Hwang and Jung 2020).
For parenting time, the results are mixed. Some studies find that non-standard schedules, especially evening and nighttime hours, are linked to reduced parenting time with children (Craig and Powell 2012; Han and Fox 2011; Han and Miller 2009; Rapoport and Le Bourdais 2008). Moreover, studies examining time spent in specific parenting activities find that non-standard schedules are associated with less parenting time devoted to activities supporting child development (Gassman-Pines 2011; Rapoport and Le Bourdais 2008; Wight, Raley, and Bianchi 2008). Evening hours, for example, decrease parenting time spent in leisure and social activities with children (Rapoport and Le Bourdais 2008). However, several studies, largely focusing on mothers, find that non-standard work schedules do not decrease total parenting time (Barnett and Gareis 2007; Connelly and Kimmel 2011; Wight et al. 2008) and are not associated with less time in developmental activities with children (Pilarz and Awkward-Rich 2024).
Dimensions of Schedules: Variability
Variability in both amount and timing of work hours could lead to parental stress as it relates to work-family conflict and economic insecurity. Among parents working in the service sector, work-hour volatility is significantly associated with greater difficulty arranging childcare and increased work-life conflict (Ananat and Gassman-Pines 2021; Carrillo et al. 2017; Cho et al. 2024; Luhr et al. 2022). Variation in work hours may also increase parental stress due to higher levels of financial strain. Large proportions of households facing high levels of work-hour volatility are under the poverty line (Cai 2024). Moreover, service-sector workers who experience more volatile work hours are more likely to report material hardship (Schneider and Harknett 2021). Work-hour volatility may also make establishing regular parent-child routines more difficult and so affect time. In terms of variability in timing of work hours, Pilarz and Awkward-Rich (2024) find that mothers who work irregular schedules spend more time with their children. However, this study is unable to differentiate between irregular schedules that are “inflexible” – i.e. not allowing mothers to negotiate family issues – and those that are unstable, and in fact, the authors call for subsequent work to examine how different aspects of work schedules, including variability and worker control, are associated with parenting time.
Dimensions of Schedules: Predictability
Unexpected work schedule changes are associated with work-family conflict (Henly and Lambert 2014; Walther and Pilarz 2024). Ananat and Gassman-Pines (2021) find that daily occurrence of last-minute changes to work schedules, including cancellations and changes in start and end time, were associated with indicators of stress like negative mood and poor sleep quality among working parents. Last-minute schedule changes also often force parents to rely on informal care arrangements (Harknett et al. 2022), which could in turn increase parental stress. Unanticipated changes to parents’ work hours also disrupt childcare routines, affecting important activities like meals (Agrawal et al. 2018) and bedtime (Lee et al. 2019) and thus reducing parenting time with children.
On the other hand, increased schedule predictability has a positive impact on working parents. Ananat, Gassman-Pines, and Fitz-Henley (2022) examined daily reports of schedule instability and well-being experienced by parents working in low-wage service sector jobs before and after the passage of the Fair Work Ordinance in Emeryville, California. The authors found that the FWO, which required employers to give advance notice of schedule changes and provide compensation for last-minute schedule changes, resulted in lower schedule instability and improved well-being of working parents.
Dimensions of Schedules: Control
Lack of control over work schedules prevents working parents from aligning their work hours with their familial obligations, resulting in higher levels of stress and, to a lesser extent, reduced time with children. Schedule flexibility and employee control over work hours have been found to improve well-being among working parents (Jang 2008), while lack of schedule control creates work-family conflict (Henly and Lambert 2014). Nomaguchi and Johnson (2016) find that workplace inflexibility is a key predictor of stress for working fathers, while Pilarz (2021) finds that the association between mothers’ inflexible work schedules – defined in that study as schedules that mothers perceive as limiting their ability to address family needs – and children’s behavioral problems is partially mediated by mothers’ parental stress. For parenting time, the association with flexible work hours is weaker (Baxter 2011; Genadek and Hill 2017). However, Haley-Lock and Posey-Maddox (2016) find that schedule inflexibility and lack of worker control decrease the degree to which mothers working in low-wage jobs are able to engage with their children’s schooling.
Overall, based on this prior literature, we expect that in the sectoral context of work hour insufficiency, usual weekly hours will be negatively associated with parental stress, but will also decrease the amount of parental time with children. We expect that non-standard timing will increase parental stress, but overall prior literature leads to ambiguous expectations for non-standard timing and time with children. We expect that both schedule variability and schedule unpredictability will both increase parental stress and reduce parental time with children. Finally, schedule control should reduce stress and perhaps allow for greater time with children.
Moderating Relationships between Work Scheduling Dimensions
Recent research on short notice, non-standard shift timing, and variable shift timing often assumes that such practices are undesirable for workers (Schneider and Harknett 2019a) and there is, in fact, confirmatory evidence that workers exposed to such practices report wanting more stable and predictable schedules (Schneider and Harknett 2019b). However, it is also possible that such aspects of work scheduling could be the product of desirable flexibility exercised by workers who opt-in to non-standard schedules, vary their hours to fit with non-work needs, and are able to alter their schedules with little notice (Fugiel and Lambert 2019; Henly, Shaefer, and Waxman 2006). Prior research has generally not distinguished between these two possibilities empirically, instead relying on evidence from the field and qualitative work (i.e. Lambert 2008; Carrillo et al. 2017; Henly et al. 2006) that these practices in fact generally represent employer driven instability (i.e. Schneider and Harknett 2019a; Choper, Schneider, and Harknett 2022).
While primarily focused on discerning the distinct dimensions of work scheduling, Lambert and Fugiel (2023) also make the point that interactions between these dimensions may capture such moderating processes. Bolino, Kelemen, and Matthews (2021) put forward a more elaborated version of this framework with specific reference to non-standard schedules, arguing that “schedule factors,” such as schedule control/input, may moderate the effects of non-standard timing on worker outcomes. However, in their review of over 150 related articles, they do not report on any empirical work testing such interactions (Bolino et al. 2021).
In recent research, Lambert et al. (2019), test for such moderation, examining if limited schedule control combines with volatility, variability, non-standard timing, and short notice to shape financial insecurity, job insecurity, and distrust in institutions. While generally signed in the expected direction but not significant, there are significant heightened associations between irregular and non-standard schedules and job insecurity and institutional distrust when workers also have little schedule control. Relatedly, Cho et al. (2024) test the effect of schedule predictability and worker control over hours on the relationship between work hour variability and work-family conflict, finding significant moderating effects of both work scheduling dimensions. Notably, worker control over hours is most impactful when work hour variability is high, while the moderating effect of schedule predictability lessens as variability increases. This conceptual and empirical work motivates further analysis of how the dimension of control may moderate any negative effects of non-standard timing, variability, or lack of advance notice.
Variation by Parent Gender, Co-Residence, and Child Age
The associations between these dimensions of work schedule quality and parental stress and parenting time may also vary by parent gender, co-residential status, and child age. The relationships between work scheduling and parental stress may be moderated by parent gender due to the differing gender role experiences and parental role demands of mothers and fathers (Scott and Alwin 1989; Simon 1992). However, given the convergence in mothers’ and fathers’ childcare time and the diminished predictive power of parent gender for time spent in childcare found in prior literature (Renk et al. 2003; Sayer et al. 2004), the impact of work schedules on parenting time with children is less likely to be moderated by parent gender.
Prior work examining parent-child co-residence finds that non-resident parents, especially fathers, are less involved in childcare overall (Boll and Schüller 2023). Consequently, non-resident parents are less subject to parenting-related time strain compared to their co-resident counterparts, who likely face time pressure due to higher childcare demands (van der Heijden et al. 2016). We therefore expect schedule quality to have more of an effect on both parental stress and time with children for co-resident parents than for non-resident parents.
In terms of child age, work schedule quality might have a greater effect on parental stress amongst parents of early adolescents, who have been found in previous research to experience higher levels of overall parental stress compared to parents of younger children or older adolescents (Putnick et al. 2010; Small et al. 1988). On the other hand, parental work schedules might have a greater impact on parenting time for parents of younger children, due to lack of time available for parenting, since older children spend less time with their families overall (Larson and Richards 1991).
Our Contribution
As detailed above, a large literature has examined the individual dimensions of work scheduling. A first tranche of this literature examined the effects of parental work duration (Connelly and Kimmel 2011; Joshi and Bogen 2007; Kendig and Bianchi 2008; Milkie et al. 2004) and non-standard timing (Connelly and Kimmel 2011; Joshi and Bogen 2007; Rapoport and Le Bourdais 2008) on parental stress and time with children. A parallel line of inquiry examined the effects of schedule control on these outcomes, generally focusing on white-collar workers (Kelly and Moen 2007; Kelly et al. 2011). More recently, there has been a vibrant literature, sparked with innovative work by Susan Lambert, investigating the consequences of schedule variability and predictability, mostly focused on workers’ own outcomes (Henly and Lambert 2014; Lambert and Henly 2014; Lambert et al. 2019; Schneider and Harknett 2019a; Pilarz et al. 2020; Schneider and Harknett 2021; Lambert and Fugiel 2023; Cho et al. 2024), but with important contributions examining parenting-related outcomes (Ananat and Gassman-Pines 2021; Gassman-Pines 2011; Harknett et al. 2022; Henly and Lambert 2005). However, across these decades of inquiry, most studies examine these dimensions of work scheduling individually rather than engaging in a comprehensive analysis of all five dimensions within a single study.
There are, however, several important studies that come closest to taking on this multi-dimensional analysis. To our knowledge, three recent studies examine four or five of these dimensions in an integrated analysis, but focus on adult workers’ own outcomes. Lambert et al. (2019) draw on nationally representative data from the General Social Survey and examine how all five dimensions of work scheduling are associated with financial insecurity, job insecurity, and institutional distrust. Lambert and Fugiel (2023) draw on data from the RAND American Working Conditions Survey to examine a comprehensive set of dimensions of work scheduling and work-life conflict, sleep problems, and satisfaction with working conditions, finding a great deal of variability in how the different dimensions of scheduling are associated with these outcomes. Schneider and Harknett (2019a) draw on data from a national sample of low-wage hourly service workers and show that non-standard timing, variability in number and timing of hours, low predictability, and limited control are negatively associated with workers’ own psychological well-being, sleep quality, and happiness. These studies demonstrate the importance of examining multiple dimensions of work schedules simultaneously, but they include the measures in separate models and only look at adult outcomes.
Three other recent studies take a multi-dimensional approach to examining scheduling, but focus specifically on parents. Walther and Pilarz (2024) make the key contribution in this respect by examining the effect of lack of schedule control, alongside measures of timing, variability, and predictability, on child behavioral problems, finding that variable shifts and unexpected schedule changes, but not variable work hours, limited notice, and low control were associated with child behavior problems. While this study stands out for using multiple dimensions of work schedules in a parenting-related context, their work differs from ours in that they do not directly examine parental stress or parenting time and only generalize to parents select counties in Illinois and New York who received child-care subsidies. Further, while Walther and Pilarz (2024) examine multiple dimensions of scheduling and an adjacent outcome (child behavior), their relatively small sample (N=376) did not permit them to examine heterogeneity by such characteristics as child age, leading them to call for future work on such interactions. We take up this question, examining moderation by parental gender, co-residential status, and child age. Luhr et al. (2022) do directly examine parental stress using a national sample of service-sector workers, but only for three of the dimensions of work scheduling and only for mothers. Pilarz et al. (2020) also incorporate three work scheduling dimensions into their analysis of fathers’ involvement using the FFCWS, but only incorporate mothers’ involvement as it interacts with fathers’. These papers also each estimate separate models for each measure of scheduling, making it difficult to disentangle the unique contributions of each dimension. We then advance prior research by examining how a full set of the key dimensions of work scheduling affect parental stress and parenting time in a national sample of hourly service-sector workers when including all measures simultaneously.
In addition, we examine how schedule control may moderate the association between variability, non-standard timing, and advance notice in predicting parental stress and parenting time. Very little research has taken this kind of interactive approach, with two notable exceptions. Lambert et al. (2019) examine the interaction of volatility and limited control and short notice and the interaction of short notice and limited control, finding generally null effects, but some evidence of heightened job insecurity and institutional distrust among workers with irregular schedules and limited control. Drawing on nationally representative data from the US General Social Survey, Cho et al. (2024) find that schedule predictability and worker control over hours moderate the relationship between work hour variability and work-family conflict. In particular, the moderating effect of schedule predictability decreases with increased variability, while the effect of worker control over hours is greatest when work hour variability is high. The authors also find gender differences in the results; the insulating effect of worker control over hours is especially pronounced for women. While work-family conflict is linked to parental stress and time with children, this article does not specifically analyze a sample of working parents and does not examine parenting time or stress.
Data and Methods
Survey Data
We draw on survey data collected by the Shift project from hourly workers at large service sector firms in the United States. These data were collected using a novel sampling and recruitment approach. The Shift Project identifies a target population of hourly workers employed at large firms in the service sector. The Shift Project then uses targeted Facebook and Instagram advertisements to recruit service-sector workers at specific employers who are identifiable in Meta’s ad targeting platform. In this way, Meta’s ad targeting platform is used both as a quasi-sampling frame and a device to recruit potential survey respondents by delivering paid advertisements to their feeds. This platform is the core of Meta’s business and contains detailed information on users that can aid in constructing a sampling frame. In our use, we construct employer-specific “audiences” of workers at each of 150 of the largest service sector firms in the United States. We then deliver targeted survey recruitment advertisements to each firm-specific audience. Users who click on the link in the advertisement are then directed to a Qualtrics site to consent and complete the online survey. This approach allows us to construct large scale survey data from a population of vulnerable service-sector workers that contains detailed novel measures of work scheduling and parenting outcomes. However, it is important to recognize the limitations of this approach as well.
Selection into Facebook/Instagram use could bias this approach. However approximately 80% of the working-age American population is active on Facebook or Instagram (Greenwood, Perrin, and Duggan 2016; Schneider and Harknett 2021), on par with telephone-based methods (Christian et al. 2010). Survey recruitment using this general approach is increasingly common and generates data that is well benchmarked to nationally representative surveys, generally (Zhang et al. 2020), and in the case of the Shift data specifically (Schneider and Harknett 2019c). The Shift Project data have been used extensively in several other studies (Choper et al. 2022; Harknett, Schneider, and Wolfe 2020; Luhr et al. 2022; Schneider and Harknett 2019a; Schneider and Harknett 2021).
We use the Shift Project survey data collected over 6 survey waves between the Spring of 2021 and the Fall of 2023 from a pooled cross-sectional sample of 71,142 completed surveys from hourly workers employed at 1 of 150 of the largest food and retail companies. We limit the sample to 11,230 parents of children 18 and under, after which the sample contains workers at 140 of the largest food and retail companies, since there are 10 firms for which none of the surveyed workers are parents. Children of parents in the sample include either biological or social and either co-resident or non-coresident, although most parents in the sample co-reside with their children. After restricting to non-missing on all relevant variables, the final harmonized sample for the main pooled models presented here comprises N = 5,292.
In our main models, described below, we use an unweighted list-wise-deleted sample. But, we also assess the sensitivity of our results to both weighting and imputation. We construct weights using the American Community Survey sample of workers in the same industries and occupations of the Shift Project firms, restricted to occupations that each comprise at least 0.5% of workers in those industries (Table A5). We further restrict the sample used for weighting to workers who are parents, weighting on age, gender, race, and industry. To impute the data, we apply the bootstrapping-based approach developed by Honaker, King, and Blackwell (2011) to the weighted sample, producing 10 implicates.
Key Variables
Dependent Variables – Parenting Outcomes
We measure parental stress and parenting time with composite indices, each derived from four survey items.
The “parental stress” index variable is a composite measure of parental stress constructed from responses to 4 survey items, and higher values of this variable indicate higher levels of parental stress. We asked respondents their level of agreement with the following statements: (1) “Being a parent is harder than I thought it would be”; (2) “I feel trapped by my responsibilities as a parent”; “I find that taking care of my child/children is much more work than pleasure”; (4) “I often feel tired, worn out, or exhausted from raising a family.” Respondents were asked to indicate whether they strongly agree, agree, disagree, or strongly disagree with each statement. Observations were reverse-coded and then summed to create the composite measure. The resulting composite parental stress index is on a scale from 0 to 12, where a value of 0 indicates that responses to all 4 items were “strongly disagree” (i.e. no parental stress), and a value of 12 indicates that responses to all 4 items were “strongly agree” (i.e. maximum parental stress).
The “parenting time” index variable is a composite measure of how often respondents engaged in certain activities with their children in the past month, constructed from responses to 4 survey items. Higher values of this variable indicate fewer occurrences of these activities, i.e. less parenting time. We asked respondents: (1) “In the past month, how often did you have a meal with your child/children?”; (2) “In the past month, how often did you and your child/children work on homework or read a book together?”; (3) “In the past month, how often did you and your child/children participate in indoor activities together (such as arts and crafts or board games)?”; (4) “In the past month, how often did you and your child/children participate in outdoor activities together (like going for a walk or to a playground)?” Respondents were asked to indicate whether their answer to each of the four questions in the past month was never, 1–2 times, once a week, several times a week, or every day. Observations were reverse-coded and then summed to create the composite measure. The resulting composite parenting time index is on a scale from 0 to 16, where a value of 0 indicates that responses to all 4 items were “never” (i.e. no parenting time), and a value of 16 indicates that responses to all 4 items were “every day” (i.e. maximum parenting time).
To assess the reliability and internal consistency of the parental stress and parenting time measures, we implement the Cronbach’s alpha test. We test the two indices separately, with the sample for each test restricted to nonmissing on the parental stress index (N = 8,516) and the parenting time index (N = 8,018) respectively. We find that the parental stress index has a reliability coefficient of 0.76 and that the parenting time index has a reliability coefficient of 0.84 (Table A2), both of which indicate acceptable reliability.
Independent Variables – Dimensions of Scheduling
We operationalize five dimensions of work scheduling, following the conceptual distinctions in Lambert and Fugiel (2023). First, we operationalize duration, defined as the number of hours that workers engage in paid work in a given time period, using self-reported usual weekly hours. Second, we measure timing with indicator variables for various schedule types: regular day, regular evening, regular night, rotating, and split shifts. We also leverage an intensive margin (number of instances) measure of exposure to clopening shifts – in which workers work a “closing” shift and then return to work an “opening” shift within 11 hours – as an additional measure of timing.1 Third, we measure variability by constructing a work-hour volatility variable defined as the difference in self-reported greatest and least weekly hours in the past month, divided by greatest weekly hours in the past month. We also include exposure to variable shifts – a schedule that changes from day to day – as a measure of variability in timing of work hours. Fourth, for predictability, we use intensive (number of instances) margin measures of exposure to timing changes, shift cancellations, and oncall shifts,2 as well as the amount of advance notice of their schedule that the worker received. Finally, we operationalize control as the degree to which the employee (relative to the employer) decides their total work hours as well as their shift start and end times, in line with a key recommendation from Lambert and Fugiel (2023) for measuring schedule control over both timing and amount of hours. We omit congruence since we do not have a measure of this construct in our data.
Controls
We adjust for a large set of potential confounders, including demographic characteristics (race, gender, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, employer fixed effects, and month and year fixed effects. In addition to nonmissingness on all covariates, the sample for all models is also restricted to (1) the number of reported occurrences for each of the following –clopenings, timing changes, oncall shifts, and canceled shifts – must individually be less than or equal to the reported total number of shifts;3 (2) the reported total number of shifts in the last month must be at or below the 97th percentile of the distribution across the sample. The additional two sample restrictions eliminate likely erroneous data (in which the number of unstable scheduling events is greater than the number of shifts worked) and extreme outliers on the number of shifts variable, respectively.
Analytical Approach
Main Models
We first estimate a series of OLS regression models describing the association between each of the dimensions of scheduling and our outcome measures: parental stress and parenting time. The models take, in turn, each parenting outcome as the dependent variable and include the full set of scheduling independent variables in each model.
We implement a step-wise model specification in which covariates are progressively added. The base model (Model 1) includes only month and year fixed effects. Model 2 adds demographic characteristics, Model 3 adds work characteristics, Model 4 adds state fixed effects, and Model 5 adds employer fixed effects. Each step-wise model includes the full set of scheduling variables. The sample for all models is restricted to nonmissing on all variables included in any of the models to ensure a consistent sample size. In total, we implement 10 models (2 dependent variables x 5 model specifications).
Next, we test whether scheduling control (captured by our measure of decision power over shift starting and stopping times) moderates the associations between timing (captured by our measure of non-standard schedules), schedule variability (captured by our measure of working a variable schedule), or predictability (captured by our measure of advance notice). We interact schedule control with each of these items, using the same model specification as above.
In supplementary models (Figures A1 and A2), we assess whether the results vary by parent gender and whether or not parents co-reside with their children. Additionally, we assess the sensitivity of our results to conditioning on the age composition of children (Figures A3-A6). We also address concerns about missing data by estimating our main model specifications on weighted and imputed data (Tables A3 and A4).
Results
Descriptive Data
Table 1 presents descriptive results on the extent of exposure to work schedule instability. We see that the parents in our sample worked an average of 37 hours per week. In terms of timing, the majority of parents in our sample worked regular day (36%) or rotating (15%) shifts, but 8% reported an evening shift and 9% reported a night shift. Approximately 40% of parents in our sample reported experiencing at least one clopening shift in the past month.
Table 1:
Descriptive Statistics for Independent Variables
| Mean (SD)/Percent | ||
|---|---|---|
| All respondents | Parents of children 18 and under | |
|
| ||
| Duration | ||
| Usual weekly hours | 35.01 (12.35) | 37.44 (12.26) |
| Timing | ||
| Schedule type | ||
| Regular day | 32 | 36 |
| Regular evening | 9 | 8 |
| Regular night | 9 | 9 |
| Rotating shifts | 16 | 15 |
| Split shifts | 2 | 1 |
| Other | 1 | 2 |
| # clopenings | ||
| 0 | 62 | 61 |
| 1 | 9 | 8 |
| 2 | 10 | 10 |
| 3 | 6 | 7 |
| 4 | 5 | 6 |
| 5+ | 6 | 7 |
| Variability | ||
| Work-hour volatility | 0.32 (1.05) | 0.30 (0.37) |
| Schedule type | ||
| Variable shifts | 31 | 29 |
| Predictability | ||
| Amount of advance notice | ||
| 0–2 days | 16 | 17 |
| 3–6 days | 14 | 14 |
| 1–2 weeks | 29 | 29 |
| 2–3 weeks | 22 | 19 |
| 3+ weeks | 20 | 21 |
| # timing changes | ||
| 0 | 42 | 44 |
| 1 | 13 | 11 |
| 2 | 17 | 17 |
| 3 | 11 | 11 |
| 4 | 6 | 6 |
| 5+ | 10 | 11 |
| # oncalls | ||
| 0 | 75 | 74 |
| 1 | 6 | 5 |
| 2 | 7 | 7 |
| 3 | 4 | 4 |
| 4 | 2 | 3 |
| 5+ | 6 | 8 |
| # canceled shifts | ||
| 0 | 88 | 89 |
| 1 | 6 | 5 |
| 2 | 3 | 4 |
| 3 | 1 | 1 |
| 4 | 1 | 0 |
| 5+ | 1 | 1 |
| Control | ||
| Control over shift start and end times | ||
| Entirely decided by employer/external factors | 48 | 51 |
| Decided by employer with worker input | 35 | 31 |
| Decided by worker within limits | 13 | 14 |
| Entirely decided by worker | 3 | 4 |
| Worker control over total weekly hours | ||
| None | 19 | 23 |
| Very little | 18 | 17 |
| A little | 17 | 16 |
| A moderate amount | 25 | 23 |
| A lot | 21 | 21 |
| N | 71,142 | 11,230 |
In terms of schedule variability, 29% of parents in our sample reported exposure to variable shifts, i.e., variability in timing of their work hours. To measure variability in amount of work, we define work-hour volatility as the relative difference in weekly hours worked, i.e. the difference of greatest weekly hours and least weekly hours, divided by greatest weekly hours. The parents in our sample experienced an average work-hour volatility of approximately 0.3, indicating fluctuations of 30% between the weeks that they worked the most and least hours in the past month. This means that the average parent in our sample would work as little as 30 hours in one week and as many as 43 hours in another week in a given month.
In terms of schedule predictability, the parents in our sample varied in the amount of advance notice they received. Approximately 30% of parents reported 1–2 weeks of advance notice, while approximately 30% reported less than a week of advance notice, and approximately 40% received more than 2 weeks of advance notice. More than half of the parents in our sample reported at least one timing change in the previous month. About one in four reported being scheduled for at least one oncall shift, and more than 10% had a shift cancelled by their employer at least once in the month prior to survey. Table 1 also shows that the experience of these aspects of schedule instability is more varied than a dichotomous measure would suggest. We find substantial variation in the degree of exposure to these unstable and unpredictable scheduling practices. While 11% of parents experienced just one timing change in the last month, 17% experienced two, 11% 3, 6% 4 changes, and 11% experienced five or more such changes. We see a similar distribution across the other measures – oncall, clopening, and cancelled shifts, even as each of these practices is less common on the extensive margin. Finally, parents reported limited work schedule control, with 40% of parents reporting having no control or very little control over their schedules, and with more than 50% of parents reporting that their schedule was entirely determined by their employer or by external factors outside of their control.
Table 1 also demonstrates that parents with children 18 and under are just as likely to experience these unstable and unpredictable scheduling practices, both on the intensive and extensive margins, as their childless counterparts.
Alongside significant exposure to work schedule instability, parents also reported significant parental stress and limited time with children. We show the components of the two scales that form our two dependent variables – parental stress and parenting time – in Table 2. We see that the mean parental stress composite index was 5.20 (on a 0–12 scale). Large fractions of respondents reported agreeing (62%) or strongly agreeing (28%) that “being a parent is harder than I thought it would be” and substantial shares agreed (39%) or strongly agreed (14%) that they “often feel tired, worn out, or exhausted from raising a family,” though feeling that care of child was more work than pleasure and feeling trapped in parental responsibilities were less common.
Table 2:
Descriptive Statistics for Dependent Variables
| Mean (SD)/Percent of Group | |
|---|---|
|
| |
| Parental stress (composite index) | 5.20 (2.73) |
| Being a parent is harder than anticipated | |
| Strongly disagree | 1 |
| Disagree | 10 |
| Agree | 62 |
| Strongly agree | 28 |
| Feel trapped by parenting responsibilities | |
| Strongly disagree | 37 |
| Disagree | 40 |
| Agree | 16 |
| Strongly agree | 7 |
| Caring for child(ren) is more work than pleasure | |
| Strongly disagree | 34 |
| Disagree | 47 |
| Agree | 15 |
| Strongly agree | 5 |
| Often feel tired from parenting responsibilities | |
| Strongly disagree | 17 |
| Disagree | 29 |
| Agree | 39 |
| Strongly agree | 14 |
| N | 8,516 |
|
| |
| Parenting time (composite index) | 7.89 (4.10) |
| Have meal with children, in past month | |
| Never in past month | 6 |
| 1–2 times in past month | 11 |
| Once a week | 17 |
| Several times a week | 38 |
| Every day | 28 |
| Read or do homework with children, in past month | |
| Never in past month | 26 |
| 1–2 times in past month | 17 |
| Once a week | 17 |
| Several times a week | 27 |
| Every day | 13 |
| Play indoors with children, in past month | |
| Never in past month | 22 |
| 1–2 times in past month | 26 |
| Once a week | 22 |
| Several times a week | 23 |
| Every day | 8 |
| Play outdoors with children, in past month | |
| Never in past month | 19 |
| 1–2 times in past month | 30 |
| Once a week | 24 |
| Several times a week | 21 |
| Every day | 6 |
| N | 8,018 |
The mean parenting time composite index was 7.89 (on a 0–16 scale). Here, respondents most often reported having meals with their children, with 66% reporting doing so at least several times a week. Engaging in homework or reading or playing with their children (indoors and outdoors) were reported with substantially less frequency.
Main Models
In Table 3, we examine the association between each of the five dimensions of work scheduling (as operationalized by our 10 measures of schedules) and parental stress. Positive coefficients indicate that more exposure to schedule instability is associated with higher levels of parenting-specific stress.
Table 3:
Parental Stress Predicted by Dimensions of Scheduling
| Parental Stress | |||||
|---|---|---|---|---|---|
| M1 | M2 | M3 | M4 | M5 | |
|
| |||||
| Duration | |||||
| Usual weekly hours | −0.01** | −0.01** | −0.01* | −0.01* | −0.01* |
| Timing | |||||
| Schedule type (ref. Regular day) | |||||
| Regular evening | −0.13 | −0.19 | −0.14 | −0.15 | −0.21 |
| Regular night | −0.16 | −0.23+ | −0.19 | −0.21 | −0.24+ |
| Rotating shifts | −0.13 | −0.17 | −0.14 | −0.14 | −0.15 |
| Split shifts | 0.48 | 0.41 | 0.43 | 0.46+ | 0.42 |
| Other | −0.19 | −0.15 | −0.18 | −0.12 | −0.14 |
| Clopenings | −0.03 | −0.05+ | −0.05+ | −0.05+ | −0.04 |
| Variability | |||||
| Work-hour volatility | 0.15* | 0.05 | 0.05 | 0.07 | 0.05 |
| Schedule type (ref. Regular day) | |||||
| Variable shifts | 0.08 | 0.05 | 0.08 | 0.09 | 0.06 |
| Predictability | |||||
| Amount of advance notice (ref. 3+ weeks) | |||||
| 0–2 days | −0.29* | −0.30* | −0.26* | −0.23+ | −0.07 |
| 3–6 days | −0.09 | −0.10 | −0.07 | −0.06 | 0.12 |
| 1–2 weeks | −0.11 | −0.14 | −0.09 | −0.10 | 0.02 |
| 2–3 weeks | 0.03 | 0.03 | 0.04 | 0.05 | 0.08 |
| Timing changes | 0.05* | 0.03 | 0.03 | 0.03 | 0.03 |
| Oncall shifts | 0.06* | 0.05 | 0.05 | 0.04 | 0.05 |
| Canceled shifts | 0.16** | 0.15* | 0.15** | 0.15* | 0.15** |
| Control | |||||
| Start and end times (ref. Entirely decided by worker) | |||||
| Entirely decided by employer/external factors | −0.02 | 0.02 | 0.01 | 0.06 | 0.08 |
| Decided by employer with worker input | 0.05 | 0.03 | 0.04 | 0.11 | 0.14 |
| Decided by worker within limits | 0.08 | 0.09 | 0.09 | 0.15 | 0.15 |
| Worker control over total weekly hours | −0.09** | −0.10** | −0.10** | −0.10** | −0.10** |
|
| |||||
| N | 5292 | 5292 | 5292 | 5292 | 5292 |
| Month/year FE | ✓ | ✓ | ✓ | ✓ | ✓ |
| Demographic chars. | ✓ | ✓ | ✓ | ✓ | |
| Work chars. | ✓ | ✓ | ✓ | ||
| State FE | ✓ | ✓ | |||
| Employer FE | ✓ | ||||
Notes:
p<.10,
p<.05,
p<.01 All models include robust standard errors clustered at the respondent level. Sample for each model restricted to (1) # of occurrences of each of clopenings, timing changes, oncall shifts, and canceled shifts ≤ total # of shifts; (2) ≤ 97th percentile of # of shifts; (3) nonmissing on parental stress, parenting time, all independent variables, and all control variables to ensure a consistent sample size. All models include month/year fixed effects.
Step-wise model specification:
Model 1 – base model, including only the month/year fixed effects.
Model 2 – adds demographic characteristics (race, gender, age, education, school enrollment, marital status, income)
Model 3 – adds work characteristics (job tenure, hourly wage, managerial status)
Model 4 – adds state fixed effects
Model 5 – adds employer fixed effects
As reported in Table 3, we find that usual weekly work duration is significantly negatively associated with parental stress. When we examine predicted values from the fully-controlled model, Model 5, we observe that that compared to parents at the mean for work hours within the analytic sample (35.6 hours/week) who have a parental stress score of 5.3, parents 1.5 SD below that mean (18.8 hours/week) have a score of 5.4, while parents 1.5 SD above that mean (52.4 hours/week) have a score of approximately 5.1, a difference of 0.3, or approximately 6% difference in the parental stress measure (Figure 1). This relationship is robust across model specifications and supports our expectation that sufficient work hours are associated with less parental stress by way of less economic anxiety.
Figure 1: Predicted Values of Parental Stress across Observed Range of Scheduling Measures.
Notes: Predictive margins of parental stress computed at each level of categorical variables and indicated fixed values of continuous variables, using estimates of fully-controlled models with the following controls: month/year fixed effects, demographic characteristics (race, gender, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, employer fixed effects, robust standard errors clustered at the respondent level.
Turning to timing, we find that relative to regular day shifts, split shifts are associated with higher levels of parental stress, though this relationship is only significant in Model 4. In contrast, the coefficients on non-standard but regular shifts – including regular evening and regular night shifts are negative, if non-significant, suggestive of less stress. Similarly, the association between exposure to clopening shifts and increased parental stress is not significant after including key controls. Work-hour volatility is associated with higher levels of parental stress in M1, though the coefficient is not significant after the addition of controls. The association between variable shifts (relative to regular day shifts) and parental stress is positive, but not significant.
While the associations between timing and variability and parental stress are then generally null, we find a strong associations between one measure of predictability and parental stress. Specifically, exposure to canceled shifts is strongly and significantly associated with higher levels of parental stress. As seen in Figure 1, parents who reported no exposure to canceled shifts had a predicted parental stress score of 5.2, while parents who reported the maximum of 5 or more cancellations had a score of 6.0, an increase in parental stress more than 15%. This aligns with our expectation that last-minute schedule changes result in parental stress due to work-family conflict. However, the results for timing changes, oncall shifts, and amount of advance notice are inconclusive.
Finally, while we find no evidence that control over shift timing is associated with parental stress, we do find that control over total weekly hours is significantly associated with lower levels of parental stress. This latter finding perhaps points to greater control over work duration preventing underemployment, thus reducing parental stress relating to economic insecurity. Examining the predicted values of parental stress, we see in the bottom right of Figure 1 that parental stress decreases with greater control over total weekly hours. Compared to parents who report “a little” control over their total weekly hours who have a predicted parental stress score of 5.3, parents who report no control over their total hours predicted parental stress score of 5.5, while parents who report “a lot” of control over their total hours predicted parental stress score of 5.1, a difference of 0.4, indicating 7% lower parental stress.
Overall, we find that increased parental stress is significantly associated with lack of schedule predictability only in terms of exposure to shift cancellations. We also find a statistically significant relationship between work duration and reduced parental stress, which could suggest that having adequate weekly hours reduces parental stress, by way of decreased anxiety about underemployment. This is further supported by the fact that worker control over total weekly hours is strongly associated with lower levels of parental stress. We do not find substantial evidence that non-standard schedules, control over shift timing, or work-hour volatility are associated with parental stress.
We estimate the same set of models, shown in Table 4, to examine the association between parenting time with children and exposure to each type of schedule instability. Here, positive coefficients indicate more time spent with children in key developmental activities.
Table 4:
Parenting Time with Children Predicted by Dimensions of Scheduling
| Parenting Time with Children | |||||
|---|---|---|---|---|---|
| M1 | M2 | M3 | M4 | M5 | |
|
| |||||
| Duration | |||||
| Usual weekly hours | −0.04** | −0.05** | −0.04** | −0.04** | −0.04** |
| Timing | |||||
| Schedule type (ref. Regular day) | |||||
| Regular evening | −1.01** | −1.10** | −1.00** | −1.00** | −0.90** |
| Regular night | −0.80** | −0.87** | −0.77** | −0.76** | −0.75** |
| Rotating shifts | −0.53** | −0.58** | −0.50** | −0.51** | −0.47** |
| Split shifts | −0.72+ | −0.77* | −0.71+ | −0.69+ | −0.62 |
| Other | −0.33 | −0.31 | −0.35 | −0.33 | −0.32 |
| Clopenings | −0.13** | −0.15** | −0.15** | −0.15** | −0.15** |
| Variability | |||||
| Work-hour volatility | −0.01 | −0.17+ | −0.16 | −0.14 | −0.12 |
| Schedule type (ref. Regular day) | |||||
| Variable shifts | −0.66** | −0.77** | −0.70** | −0.70** | −0.71** |
| Predictability | |||||
| Amount of advance notice (ref. 3+ weeks) | |||||
| 0–2 days | −0.59** | −0.53** | −0.44* | −0.49* | −0.51* |
| 3–6 days | −0.09 | −0.12 | −0.05 | −0.11 | −0.16 |
| 1–2 weeks | −0.41* | −0.41* | −0.32+ | −0.37* | −0.40* |
| 2–3 weeks | 0.07 | 0.03 | 0.07 | 0.05 | −0.03 |
| Timing changes | −0.07* | −0.11** | −0.11** | −0.10** | −0.09* |
| Oncall shifts | −0.01 | −0.03 | −0.02 | −0.02 | −0.00 |
| Canceled shifts | 0.14+ | 0.15* | 0.16* | 0.16* | 0.16* |
| Control | |||||
| Start and end times (ref. Entirely decided by worker) | |||||
| Entirely decided by employer/external factors | −0.73* | −0.67+ | −0.70+ | −0.71* | −0.73* |
| Decided by employer with worker input | −0.46 | −0.51 | −0.51 | −0.55 | −0.54 |
| Decided by worker within limits | −0.39 | −0.39 | −0.41 | −0.46 | −0.50 |
| Worker control over total weekly hours | 0.30** | 0.29** | 0.29** | 0.30** | 0.30** |
|
| |||||
| N | 5292 | 5292 | 5292 | 5292 | 5292 |
| Month/year FE | ✓ | ✓ | ✓ | ✓ | ✓ |
| Demographic chars. | ✓ | ✓ | ✓ | ✓ | |
| Work chars. | ✓ | ✓ | ✓ | ||
| State FE | ✓ | ✓ | |||
| Employer FE | ✓ | ||||
Notes:
p<.10,
p<.05,
p<.01 All models include robust standard errors clustered at the respondent level. Sample for each model restricted to (1) # of occurrences of each of clopenings, timing changes, oncall shifts, and canceled shifts ≤ total # of shifts; (2) ≤ 97th percentile of # of shifts; (3) nonmissing on parental stress, parenting time, all independent variables, and all control variables to ensure a consistent sample size. All models include month/year fixed effects.
Step-wise model specification:
Model 1 – base model, including only the month/year fixed effects.
Model 2 – adds demographic characteristics (race, gender, age, education, school enrollment, marital status, income)
Model 3 – adds work characteristics (job tenure, hourly wage, managerial status)
Model 4 – adds state fixed effects
Model 5 – adds employer fixed effects
As reported in Table 4, in terms of duration of work hours and parenting time with children, we find that usual weekly work duration is significantly associated with reduced parenting time with children. When parents work longer hours, they less frequently engage in core parenting activities with their children. Taking predicted values from the fully-controlled model (Model 5), we see in the top-left corner of Figure 2 that compared to parents at the mean for work hours within the analytic sample (35.6 hours/week) who have a parenting time score of 7.9, parents 1.5 SD below that mean (18.8 hours/week) have a score of 8.6, while parents 1.5 SD above that mean (52.4 hours/week) have a score of approximately 7.1, a difference of 1.5, or a 17% lower parenting time score.
Figure 2: Predicted Values of Parenting Time across Observed Range of Scheduling Measures.
Notes: Predictive margins of parenting time with children computed at each level of categorical variables and indicated fixed values of continuous variables, using estimates of fully-controlled models with the following controls: month/year fixed effects, demographic characteristics (race, gender, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, employer fixed effects, robust standard errors clustered at the respondent level.
While we found little evidence that shift timing was related to parental stress, we find much stronger evidence that parenting time is patterned by shift timing. Relative to regular day shifts, all alternative schedule types are strongly and significantly associated with less parenting time, which supports our hypothesis that non-standard work hours reduce parenting time. These results are robust across model specifications. As shown in Figure 2, while parents who work a regular day schedule have a predicted parenting time index of 8.3, parents who work a regular evening schedule have a predicted index of 7.4 (11% decrease in the parenting time index), and those who work a regular night schedule have a predicted index of 7.5 (10% decrease in the parenting time index). Parents who work rotating and split shifts have slightly higher predicted parenting time indices of 7.8 and 7.7 respectively, but still far lower than those who work a regular day schedule (6% and 7% decreases in parenting time, respectively). This points to the lack of time available to spend with children for parents whose work hours fall outside of the traditional school day, but also how variation in shift timing may upset parenting routines. We also find that frequent clopening shifts are strongly associated with reduced parenting time. Examining the predicted values in Figure 2, we see strong evidence of a dose-response relationship between frequency of clopenings and reduced parenting time. Parents who do not report any exposure to clopening shifts have a predicted parenting time score of 8.0, while parents who report even just two clopening shifts have a notably lower predicted score of 7.7. Parents who report 5 or more clopening shifts have a score of just 7.3, or nearly 10% less parenting time lower than those who report no clopenings.
While work-hour volatility is not significantly associated with parenting time with children, variability in shift timing is significantly associated with decreased parenting time. Compared to parents who work a regular day shift (predicted parenting time score of 8.3), parents who work variable shifts have a predicted parenting time index of 7.6, or more than 8% lower parenting time. The fact that variable shifts have a stronger negative effect on parenting time compared to rotating shifts suggests that more frequent variation in shift timing – changing from day to day – is more consequential.
In contrast to hour volatility, there is strong and consistent evidence across measures that schedule unpredictability is associated with less time with children. Exposures to timing changes is strongly and significantly associated with less parenting time. Parents who report zero exposure to timing changes have predicted parenting time scores of 8, while those who report 5 or more have scores of 7.6. Relative to having 3 or more weeks of advance notice, having less than 2 weeks, and especially less than 2 days, notice of work schedules is significantly negatively associated with parenting time. Parents who report 3 or more weeks of advance notice have a predicted parenting time score of 8.1, while parents who report less than 2 days of advance notice have a score of 7.6, a difference of more than 6% (Figure 2). This aligns with our theory that last-minute schedule changes disrupt family routines, thus reducing parenting time with children. Interestingly, exposure to canceled shifts is associated with increased parenting time, across model specifications. Parents who report 5 or more cancellations have a predicted parenting time score of 8.6, which is more than 10% higher than the predicted parenting time score of parents who report no exposure to canceled shifts (7.8). This aligns with our theory that parenting time with children is partially driven by total time availability.
Finally, in terms of control, relative to respondents whose schedules are entirely worker-decided, workers with less control over their shift start and end times are significantly likely to report less parenting time with children. Parents who report having full control over their shift timing have a very high predicted parenting time score of 8.5, while those who report that their schedules are entirely decided by their employer or by external factors have a score of only 7.8, or 8% less parenting time. Meanwhile, having more control over total weekly hours is significantly associated with more parenting time with children, and predicted parenting time progressively increases at higher levels of control over total hours. Parents with no control over their total weekly hours have a predicted parenting time score of 7.2, while those with “a lot” of control have a score of 8.4, an increase of nearly 17% (Figure 2).
Overall, we find robust associations between non-standard and variable timing, unpredictability, and lack of control and reduced parenting time with children. This provides evidence of the time-based strain and work-family conflict mechanisms linking work schedule instability to reduced parenting time. We also observe a small but statistically significant negative relationship between work duration and parenting time, which is also consistent with the idea that the amount of time parents spend caring for their children is also driven by overall time availability.
Moderation of Non-Standard Timing, Variability, and Predictability by Schedule Control
In Figure 3, we show the results from a set of models which follow the same specification as above, but now include interactions between worker control over setting shift starting and stopping times and measures that capture other key dimensions of work scheduling to predict our key outcomes.
Figure 3: Estimated Moderating Effect of Schedule Control on Associations of Schedule Type and Advance Notice with Parenting Outcomes.
Notes: Error bars indicate 95% confidence intervals. All models include robust standard errors clustered at the respondent level. Each model run with each indicated variable fully interacted with either schedule type or amount of advance notice, as indicated. Sample for each model restricted to (1) # of occurrences of each of clopenings, timing changes, oncall shifts, and canceled shifts ≤ total # of shifts; (2) ≤ 97th percentile of # of shifts; (3) nonmissing on parental stress, parenting time, all independent variables in the main analysis as well as those included in these models, and all control variables to ensure a consistent sample size. All models include month/year fixed effects, demographic characteristics (race, gender - in co-residence models, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, and employer fixed effects.
For parental stress, we first examine the interaction of non-standard timing and schedule control. Compared to when the employer entirely controls the scheduling process, the associations between non-standard shifts (vs. regular day shifts) are less negative when workers have at least some control and workers have significant control over scheduling. While signed in the expected direction, these interactions are only statistically significant for regular evening shifts interacted with some employee control (B = −0.75; p < 0.05), regular evening shifts interacted with significant employee control (B= −0.95; p < 0.05), and rotating shifts interacted with significant employee control (B = −1.26; p < 0.01) (Table A1). The interactions between variable timing and control are also in the expected direction, with more control muting the effects of variable shifts, but the coefficients are not significant. Similarly, short notice appears less harmful for parental stress when coupled with employee input, but these coefficients are not statistically significant either.
For parental time with children, we see a largely similar pattern emerge. Partial or significant worker control over scheduling (as compared with total employer control) moderates the associations between non-standard timing and variable timing, with the coefficients on the interactions generally positively signed, but never significant. For predictability, short notice is moderated in the expected direction when workers’ have significant control over schedules, but these coefficients are also not significant, and the sign is not consistent for worker partial control (Figure 3).
Sub-Analyses by Parent Gender and Co-residential Status
We also examine the potential moderating effect of parent gender and co-residential status by estimating the fully-controlled models from our main analysis, with parent gender and, alternatively, co-residence interacted with the key independent variable in each model. As reported in Figure A1, we find that parent gender does not have a significant moderating effect on the associations between parental stress and any of the scheduling dimensions, except for worker control over weekly hours. For parenting time, we find that parent gender does not moderate the associations with any of the schedule instability measures (Figure A2). Similarly, we do not find any evidence of a moderating effect of co-residence status on the associations between schedule instability and parental stress (Figure A1). As reported in Figure A2, co-residence status does not moderate any of the associations between parenting time and schedule instability except for working split shifts, relative to regular day shifts. This is unsurprising, given that non-resident parents are a minority in our sample. This may be in part attributable to the relatively small sample size of non-resident parents, as well as that we only measure co-residence with children age 14 and under.
Sub-Analyses by Child Age
To assess whether the findings vary by child age, we estimate our main models stratified by grouped child age (under 5, age 5–9, age 10–14, and age 15–18). As seen in Figures A3 and A4, out of 20 comparisons, none are significantly different. However, we are hesitant to interpret these substantively given the risks of multiple comparisons. We therefore estimate the models using interactions between child age group and scheduling dimensions, using child age 15–18 as the reference category. As reported in Figures A5 and A6, we find significant moderating effects of having at least one child age 0–4 on the associations between work-hour volatility and parental stress. We also find that having at least one child age 10–14 moderates the association between parental stress and working regular night shifts, as well as between parental stress and the lowest amount of advance notice (0–2 days) (Figure A5). Most notably, we find that having at least one child age 5–9 moderates the effect of a limited relative lack of worker control over shift timing on parenting time with children (Figure A6).
Supplemental Analysis on Weighted and Imputed Data
We address potential concerns about sample selection bias through poststratification weighting and imputation. We estimate our preferred fully-controlled models with two different approaches applied to our analytic sample: (1) weighted using the ACS sample of currently employed workers in the same industries and occupations of the Shift Project sample (Table A5), restricting to workers in occupations that each comprise at least 0.5% of the workers in the represented industries, who also are parents, weighting on age, gender, race, and industry; (2) multiply imputed in addition to weighted using the bootstrapping-based approach designed by Honaker, King, and Blackwell (2011) to produce 10 implicates. As reported in Tables A3 and A4, the results are quite robust to weighting and imputation. Although the associations attenuate slightly in magnitude with imputation and some lose significance, we observe no differences in estimates across the various samples, except for the association between work-hour volatility and parental stress, which is positive and insignificant in the unweighted model, but negative and significant in the weighted model, though the estimate loses significance with imputation. The same overarching patterns seen in our main models still persist.
Discussion
Children’s lives are starkly stratified by the stresses faced by their parents and limits that their parents face on investing in developmental childcare. Prior work has recognized that these two crucial family processes are in turn produced at the intersection of family and working lives. Scholarship increasingly recognizes the role of parental schedules in shaping these processes. Important prior work has examined how the timing of parental schedules, especially related to non-standard shifts, is associated with parental stress (Joshi and Bogen 2007) and parenting time (Han and Fox 2011; Pilarz and Awkward-Rich 2024; Pilarz et al. 2020), finding increases in stress and somewhat mixed results for parenting time. The duration of parental work time is not strongly associated with parental stress in prior work (Nomaguchi and Johnson 2016), but is negatively associated with parenting time (Pilarz and Awkward-Rich 2024; Yeung et al. 2001).
However, parental schedules are complex and can vary across multiple key dimensions, beyond just timing and duration. We draw on recent conceptual work by Lambert and Fugiel (2023) who discern distinct dimensions of scheduling, including timing and duration, but also variability in the amount of work hours, the predictability of schedules, and control over schedules. These dimensions have been much less studied as predictors of parental stress and parenting time in prior work. But, it is exactly these dimensions that appear to capture the contemporary experience of work time for millions of hourly workers in the United States. Beyond the fact of working or of working at night or on the weekend, workers, especially hourly workers in the service sector, are exposed to work schedules that are unstable and unpredictable, that vary from day to day and week to week and are often changed with little notice.
Our understanding of the role of work schedules in shaping parental stress and parenting time has not kept pace with these rapid developments in the world. Likely constrained by a lack of data that measures these parenting processes alongside detailed multi-dimensional measures of work scheduling, we have little evidence that allows us to carefully compare how different dimensions of work scheduling are associated with these two important family processes.
We draw on new data collected by The Shift Project, to examine the association between parental exposure to five different dimensions of work scheduling – duration, timing, variability, predictability, and control – as captured with new detailed measures, and parental stress and parenting time. We find that work duration reduces parental stress, perhaps via economic security, but also decreases parenting time, likely through a time-availability pathway. Although schedule timing is not associated with parental stress, we find strong evidence that non-standard and irregular timing reduces parenting time with children, especially using our novel intensive margin measure of exposure to clopening shifts. This is likely due to increased time-based strain and work-family conflict. While we find no evidence that schedule variability, as measured by work-hour volatility, is associated with either outcome, we find a strong negative association between variable shift timing and parenting time. Lack of predictability is strongly associated with higher levels of parental stress and with reduced parenting time. In particular, parents exposed to frequent last-minute timing changes and oncall shifts report less time with children, as do parents who receive less than 2 weeks’ advance notice of their schedule. Cancelled shifts are strongly associated with increased parental stress, but not with less time. Finally, we find that greater worker control over schedule timing and weekly duration increases parenting time, while worker control specifically over total weekly duration reduces parental stress.
Our work extends the well-developed literature on non-standard scheduling and parental stress and parenting time to describe the consequences of a new, even more precarious temporal order at work and puts these different dimensions of scheduling in direct comparison. We add to the small but growing literature on the consequences of parental exposure to unstable and unpredictable schedules on child-care arrangements and child outcomes. We advance the more general literature on the effects on worker economic and social well-being by demonstrating the persistent effects of schedule instability on parental stress and time with children, especially in developmentally important parent-child interactions.
We also contribute to nascent research that examines interactions between these different dimensions of scheduling. We test if control over schedule timing moderates the associations between non-standard timing, variable timing, and predictability. While we find directionally expected coefficients, they are not generally statistically significant. This may reflect a lack of sufficient power or may reflect the reality that, in the service sector, even workers with input into starting and stopping times, but who have non-standard, variable, and short-notice schedules are rarely actually free of employer control (Wood 2020). Nevertheless, this remains an important area for future conceptual development and empirical work.
This points to both a strength and limitation of our work. These results are based on a narrow slice of the US workforce – hourly workers employed at large, generally low-wage, firms in the food service and retails sectors. The processes that we describe here may not generalize for the broader population of workers for whom these scheduling dimensions may have different lived meanings and who may be better able to draw on other resources, such as higher incomes, to buffer the negative consequences of poor scheduling quality. Indeed, in work that explicitly compares hourly and salaried workers, Lambert et al. (2019), find significant heterogeneity in effects on workers’ own outcomes.
While this limitation is important, it is also the case that the population of workers studied here very closely resembles the population of workers at the center of recent labor organizing efforts, such as the Fight for $15 and campaigns at Starbucks and Amazon, as well as of recent labor standards policy making. For this reason, this work has the potential to usefully inform ongoing advocacy and policy work to regulate work schedule quality. Fair Work Week laws passed in San Francisco, Seattle, New York City, Chicago, and other jurisdictions around the country require large service sector employers, exactly the employers for whom members of our sample work, to provide 2 weeks of advance notice and regulate last-minute timing changes, cancelled shifts, clopening shifts, and oncall shifts (Lyden 2020; Wolfe et al. 2018). These laws are especially impactful for working parents, who are often among the targeted beneficiaries of clauses that specify accommodations for employees with caregiving responsibilities (Vogtman and Schulman 2016). While these laws share much in common, they also vary in their specific details. For instance, laws in Oregon and Philadelphia initially did not require 14 days’ notice and the law in San Francisco does not regulate clopening shifts. Our results first reinforce the importance of regulations targeting schedule predictability. Amount of advance notice, timing changes, oncall shifts, and cancelled shifts are among the most important predictors of parental stress and parenting time that we assess. Additionally, our results suggest that at least when it comes to parenting time, regulating clopening shifts and requiring at least two weeks’ of advance notice are key provisions given their strong associations with parenting time. Finally, our results demonstrate the positive impact that greater schedule control can have for working parents and support policies requiring employers to include worker input in scheduling. These implications are especially salient in light of employer opposition and legal challenges to fair work week legislation (Lyden 2020).
That said, the existing set of Fair Work Week laws are complex and have proven challenging to fully implement (Harknett, Schneider, and Irwin 2021; Lambert and Haley 2021; Petrucci et al. 2022). The challenges of implementation, compliance, and enforcement appear most acute when it comes to access to hours provisions, and perhaps most tractable when it comes to providing at least two-weeks’ advance notice and reducing last-minute timing changes. Our work suggests that there is consonance between the tractable and the beneficial in so far as aspects of predictability – specifically advance notice and fewer timing changes - are important for the family processes that we study.
Acknowledgements:
We acknowledge Connor Williams, Dylan Nguyen, Jeremy Mopsick, and Ale Soto’s excellent research assistance. We are grateful for useful feedback from participants in the Shift Project as well as from Afshin Zilanawala. A previous version of this paper was presented at the 2024 meetings of the Population Association of America. Research reported in this publication was supported by the Bill and Melinda Gates Foundation (INV-002665), the National Institute on Aging (R01AG066898), the Robert Wood Johnson Foundation (74528), and the William T. Grant Foundation (188043). The content is solely the responsibility of the authors and does not necessarily represent the official views of these foundations. The findings and conclusions contained within are those of the authors and do not necessarily reflect positions or policies of these foundations. The authors have no conflicts of interest to declare that are relevant to the content of this article. The data used in this study are not publicly available as they contain personally identifying information, including geographic location (US state-level) and employer name, obtained under informed consent from Shift Project participants. Data alongside code for the purposes of replication will be available from the authors upon request.
Appendix
Table A1:
Estimated Moderating Effect of Schedule Control on Associations of Schedule Type and Advance Notice with Parenting Outcomes
| Parenting Outcomes | ||||
|---|---|---|---|---|
| Parental Stress | Parenting Time | Parental Stress | Parenting Time | |
|
| ||||
| Shift timing control (ref. Entirely employer) | ||||
| X Schedule type (ref. Reg. day) | ||||
| Mostly employer X Variable shifts | −0.23 | 0.12 | ||
| Mostly employer X Reg. evening | −0.75* | 0.07 | ||
| Mostly employer X Reg. night | −0.45 | 0.08 | ||
| Mostly employer X Rotating shifts | −0.31 | 0.13 | ||
| Mostly employer X Split shifts | −0.14 | 0.02 | ||
| Mostly employer X Other | −0.43 | 1.33 | ||
| Mostly/all worker X Variable shifts | −0.39 | 0.02 | ||
| Mostly/all worker X Reg. evening | −0.95* | −0.33 | ||
| Mostly/all worker X Reg. night | −0.12 | 0.55 | ||
| Mostly/all worker X Rotating shifts | −1.26** | 0.31 | ||
| Mostly/all worker X Split shifts | −0.93 | −0.37 | ||
| Mostly/all worker X Other | 0.46 | 1.67 | ||
| Shift timing control (ref. Entirely employer) | ||||
| X Advance notice (ref. 3+ weeks) | ||||
| Mostly employer X 0–2 days | 0.21 | −0.68 | ||
| Mostly employer X 3–6 days | 0.24 | −0.65 | ||
| Mostly employer X 1–2 weeks | −0.17 | −0.07 | ||
| Mostly employer X 2–3 weeks | −0.05 | 0.40 | ||
| Mostly/all worker X 0–2 days | −0.08 | −0.21 | ||
| Mostly/all worker X 3–6 days | −0.04 | 0.54 | ||
| Mostly/all worker X 1–2 weeks | −0.25 | 0.44 | ||
| Mostly/all worker X 2–3 weeks | 0.29 | 0.20 | ||
|
| ||||
| N | 5292 | 5292 | 5292 | 5292 |
| Other scheduling vars. | ✓ | ✓ | ✓ | ✓ |
| Month/year FE | ✓ | ✓ | ✓ | ✓ |
| Demographic chars. | ✓ | ✓ | ✓ | ✓ |
| Work chars. | ✓ | ✓ | ✓ | ✓ |
| State FE | ✓ | ✓ | ✓ | ✓ |
| Employer FE | ✓ | ✓ | ✓ | ✓ |
Notes:
p<.10,
p<.05,
p<.01 All models include robust standard errors clustered at the respondent level. Sample restricted to (1) # of occurrences of each of clopenings, timing changes, oncall shifts, and canceled shifts ≤ total # of shifts; (2) ≤ 97th percentile of # of shifts; (3) nonmissing on parental stress, parenting time, all independent variables, and all control variables to ensure a consistent sample size. Model includes control over shift start and end time (reference category: entirely employer) fully interacted with schedule type (reference category: regular day shift) and with amount of advance notice (reference category: 3+ weeks) as indicated. All models include all other scheduling variables (usual weekly hours, # clopenings, work-hour volatility, # timing changes, # oncalls, # canceled shifts, worker control over weekly hours; schedule type and amount of advance notice if not already in the interaction), month/year fixed effects, demographic characteristics (race, gender, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, and employer fixed effects
Table A2:
Reliability Analysis of Parenting Outcome Measures
| Item-test Correlation | Item-rest Correlation | Avg. Inter-item Covariance | Reliability Coeff. (α) | |
|---|---|---|---|---|
|
| ||||
| Parental Stress Index | 0.30 | 0.76 | ||
| Being a parent is harder than anticipated | 0.61 | 0.42 | 0.41 | 0.77 |
| Feel trapped by parenting responsibilities | 0.83 | 0.65 | 0.25 | 0.66 |
| Caring for child(ren) is more work than pleasure | 0.80 | 0.61 | 0.28 | 0.68 |
| Often feel tired from parenting responsibilities | 0.81 | 0.59 | 0.26 | 0.69 |
|
| ||||
| Parenting Time Index | 0.88 | 0.84 | ||
| Have meal with children, in past month | 0.76 | 0.60 | 1.01 | 0.83 |
| Read or do homework with children, in past month | 0.85 | 0.69 | 0.80 | 0.79 |
| Play indoors with children, in past month | 0.86 | 0.73 | 0.82 | 0.77 |
| Play outdoors with children, in past month | 0.82 | 0.69 | 0.90 | 0.79 |
Notes: The Cronbach’s alpha test was implemented to assess reliability and internal consistency of parental stress and parenting time indices. Each of the two indices was tested separately, with the sample for each test restricted to nonmissing on the parental stress index (N = 8,516) and the parenting time index (N = 8,018) respectively. The average inter-item covariance and the reliability coefficient for each summative index are presented. For each subcomponent of each scale, the following values are presented: item-test correlation (correlation between that item and the overall index), item-rest correlation (correlation between that item and the index formed by all other items), average inter-item correlation excluding that item, and the reliability coefficient excluding that item.
Table A3:
Parental Stress Predicted by Dimensions of Scheduling, with Weighted and Imputed Data
| Parental Stress | |||
|---|---|---|---|
| Unweighted | Weighted | Weighted and Imputed | |
|
| |||
| Duration | |||
| Usual weekly hours | −0.01* | −0.01* | −0.02** |
| Timing | |||
| Schedule type (ref. Regular day) | |||
| Regular evening | −0.21 | −0.10 | 0.06 |
| Regular night | −0.24+ | 0.03 | 0.14 |
| Rotating shifts | −0.15 | 0.08 | −0.01 |
| Split shifts | 0.42 | 0.34 | 0.34 |
| Other | −0.14 | 0.58 | 0.14 |
| Clopenings | −0.04 | −0.04 | −0.04 |
| Variability | |||
| Work-hour volatility | 0.05 | −0.33+ | −0.18 |
| Schedule type (ref. Regular day) | |||
| Variable shifts | 0.06 | 0.14 | 0.14 |
| Predictability | |||
| Amount of advance notice (ref. 3+ weeks) | |||
| 0–2 days | −0.07 | −0.11 | −0.09 |
| 3–6 days | 0.12 | 0.13 | 0.07 |
| 1–2 weeks | 0.02 | 0.15 | 0.12 |
| 2–3 weeks | 0.08 | −0.14 | −0.12 |
| Timing changes | 0.03 | 0.00 | 0.01 |
| Oncall shifts | 0.05 | 0.05 | 0.06 |
| Canceled shifts | 0.15** | 0.25** | 0.20** |
| Control | |||
| Start and end times (ref. Entirely decided by worker) | |||
| Entirely decided by employer/external factors | 0.08 | 0.94* | 0.43 |
| Decided by employer with worker input | 0.14 | 1.05* | 0.56 |
| Decided by worker within limits | 0.15 | 1.06* | 0.75+ |
| Worker control over total weekly hours | −0.10** | −0.19** | −0.16** |
|
| |||
| N | 5292 | 5090 | 6245 |
| Month/year FE | ✓ | ✓ | ✓ |
| Demographic chars. | ✓ | ✓ | ✓ |
| Work chars. | ✓ | ✓ | ✓ |
| State FE | ✓ | ✓ | ✓ |
| Employer FE | ✓ | ✓ | ✓ |
Notes:
p<.10,
p<.05,
p<.01 All models include robust standard errors clustered at the respondent level. Sample for each model restricted to (1) # of occurrences of each of clopenings, timing changes, oncall shifts, and canceled shifts ≤ total # of shifts; (2) ≤ 97th percentile of # of shifts; (3) nonmissing on parental stress, parenting time, all independent variables in the main analysis as well as those included in these models, and all control variables to ensure a consistent sample size. Sample weighted to the overall population of workers in the appropriate occupations and industries who are also parents, and where indicated, multiply imputed in addition to weighted, and pooled across 10 imputations. All models include month/year fixed effects, demographic characteristics (race, gender, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, and employer fixed effects.
Table A4:
Parenting Time Predicted by Dimensions of Scheduling, with Weighted and Imputed Data
| Parenting Time | |||
|---|---|---|---|
| Unweighted | Weighted | Weighted and Imputed | |
|
| |||
| Duration | |||
| Usual weekly hours | −0.04** | −0.04** | −0.02* |
| Timing | |||
| Schedule type (ref. Regular day) | |||
| Regular evening | −0.90** | −0.72* | −0.37 |
| Regular night | −0.75** | −0.67+ | −0.41+ |
| Rotating shifts | −0.47** | −0.25 | 0.14 |
| Split shifts | −0.62 | 0.00 | 0.84 |
| Other | −0.32 | 1.86+ | 0.13 |
| Clopenings | −0.15** | −0.10+ | −0.09+ |
| Variability | |||
| Work-hour volatility | −0.12 | −0.18 | −0.11 |
| Schedule type (ref. Regular day) | |||
| Variable shifts | −0.71** | −0.55* | −0.10 |
| Predictability | |||
| Amount of advance notice (ref. 3+ weeks) | |||
| 0–2 days | −0.51* | −0.04 | −0.06 |
| 3–6 days | −0.16 | 0.17 | 0.15 |
| 1–2 weeks | −0.40* | −0.00 | −0.03 |
| 2–3 weeks | −0.03 | 0.13 | −0.03 |
| Timing changes | −0.09* | −0.11* | −0.10* |
| Oncall shifts | −0.00 | −0.00 | −0.03 |
| Canceled shifts | 0.16* | 0.07 | 0.18* |
| Control | |||
| Start and end times (ref. Entirely decided by worker) | |||
| Entirely decided by employer/external factors | −0.73* | −0.23 | 0.01 |
| Decided by employer with worker input | −0.54 | 0.22 | 0.41 |
| Decided by worker within limits | −0.50 | 0.32 | 0.50 |
| Worker control over total weekly hours | 0.30** | 0.35** | 0.20** |
|
| |||
| N | 5292 | 5090 | 6245 |
| Month/year FE | ✓ | ✓ | ✓ |
| Demographic chars. | ✓ | ✓ | ✓ |
| Work chars. | ✓ | ✓ | ✓ |
| State FE | ✓ | ✓ | ✓ |
| Employer FE | ✓ | ✓ | ✓ |
Notes:
p<.10,
p<.05,
p<.01 All models include robust standard errors clustered at the respondent level. Sample for each model restricted to (1) # of occurrences of each of clopenings, timing changes, oncall shifts, and canceled shifts ≤ total # of shifts; (2) ≤ 97th percentile of # of shifts; (3) nonmissing on parental stress, parenting time, all independent variables in the main analysis as well as those included in these models, and all control variables to ensure a consistent sample size. Sample weighted to the overall population of workers in the appropriate occupations and industries who are also parents, and where indicated, multiply imputed in addition to weighted, and pooled across 10 imputations. All models include month/year fixed effects, demographic characteristics (race, gender, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, and employer fixed effects.
Table A5:
Industries and Occupations of Employed Workers in the American Community Survey Sample Used for Weighting and Imputation, Comprising Workers in Both a Listed Industry and a Listed Occupation
| Industries | |
|---|---|
| Lumber and building material retailing | Furniture and home furnishings stores |
| Hardware stores | Radio, TV, and computer stores |
| Department stores | Eating and drinking places |
| Miscellaneous general merchandise stores | Drug stores |
| Grocery stores | Sporting goods, bicycles, and hobby stores |
| Auto and home supply stores | Miscellaneous retail stores |
| Gasoline service stations | Retail trade, n.s. |
| Apparel and accessory stores, except shoe | |
|
Occupations | |
| Supervisors and proprietors of sales jobs | Cooks, variously defined |
| Retail sales clerks | Cashiers |
| Waiter/waitress | Managers of food-serving and lodging establishments |
| Stock and inventory clerks | Misc. food prep workers |
| Customer service reps, investigators and adjusters, except insurance | Truck, delivery, and tractor drivers |
| Laborers outside construction | Dental laboratory and medical appliance technicians |
| Pharmacists | Kitchen workers |
| Bartenders | Janitors |
| Salespersons, n.e.c. | Bookkeepers and accounting and auditing clerks |
| Butchers and meat cutters | Shipping and receiving clerks |
| Secretaries | Waiter’s assistant |
| Bakers | |
Notes: Occupations restricted to those that each alone comprise at least 0.5% of workers in the represented industries.
Figure A1: Estimated Moderating Effects of Parent Gender and Co-residential Status on Associations of Dimensions of Scheduling and Parental Stress.
Notes: Error bars indicate 95% confidence intervals. All models include robust standard errors clustered at the respondent level. Each model run with each indicated variable fully interacted with either parent gender or co-residential status, as indicated. Sample for each model restricted to (1) # of occurrences of each of clopenings, timing changes, oncall shifts, and canceled shifts ≤ total # of shifts; (2) ≤ 97th percentile of # of shifts; (3) nonmissing on parental stress, parenting time, all independent variables in the main analysis as well as those included in these models, and all control variables to ensure a consistent sample size. All models include month/year fixed effects, demographic characteristics (race, gender - in co-residence models, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, and employer fixed effects. Stratification is by whether one or more of a given respondent’s children were in the indicated age group.
Figure A2: Estimated Moderating Effects of Parent Gender and Co-residential Status on Associations of Dimensions of Scheduling and Parenting Time.
Notes: Error bars indicate 95% confidence intervals. All models include robust standard errors clustered at the respondent level. Each model run with each indicated variable fully interacted with either parent gender or co-residential status, as indicated. Sample for each model restricted to (1) # of occurrences of each of clopenings, timing changes, oncall shifts, and canceled shifts ≤ total # of shifts; (2) ≤ 97th percentile of # of shifts; (3) nonmissing on parental stress, parenting time, all independent variables in the main analysis as well as those included in these models, and all control variables to ensure a consistent sample size. All models include month/year fixed effects, demographic characteristics (race, gender - in co-residence models, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, and employer fixed effects. Stratification is by whether one or more of a given respondent’s children were in the indicated age group.
Figure A3: Estimated Effects of Dimensions of Scheduling on Parental Stress, by Child Age.
Notes: Error bars indicate 95% confidence intervals. All models include robust standard errors clustered at the respondent level. Sample for each model restricted to (1) # of occurrences of each of clopenings, timing changes, oncall shifts, and canceled shifts ≤ total # of shifts; (2) ≤ 97th percentile of # of shifts; (3) nonmissing on parental stress, parenting time, all independent variables in the main analysis as well as those included in these models, and all control variables to ensure a consistent sample size. All models include month/year fixed effects, demographic characteristics (race, gender, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, and employer fixed effects. Stratification is by whether one or more of a given respondent’s children were in the indicated age group.
Figure A4: Estimated Effects of Dimensions of Scheduling on Parenting Time with Children, by Child Age.
Notes: Error bars indicate 95% confidence intervals. All models include robust standard errors clustered at the respondent level. Sample for each model restricted to (1) # of occurrences of each of clopenings, timing changes, oncall shifts, and canceled shifts ≤ total # of shifts; (2) ≤ 97th percentile of # of shifts; (3) nonmissing on parental stress, parenting time, all independent variables in the main analysis as well as those included in these models, and all control variables to ensure a consistent sample size. All models include month/year fixed effects, demographic characteristics (race, gender, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, and employer fixed effects. Stratification is by whether one or more of a given respondent’s children were in the indicated age group.
Figure A5: Estimated Moderating Effect of Child Age on Associations of Dimensions of Scheduling and Parental Stress.
Notes: Error bars indicate 95% confidence intervals. All models include robust standard errors clustered at the respondent level. Each model run with each indicated variable fully interacted with child age, with 15–18 as the reference category. Sample for each model restricted to (1) # of occurrences of each of clopenings, timing changes, oncall shifts, and canceled shifts ≤ total # of shifts; (2) ≤ 97th percentile of # of shifts; (3) nonmissing on parental stress, parenting time, all independent variables in the main analysis as well as those included in these models, and all control variables to ensure a consistent sample size. All models include month/year fixed effects, demographic characteristics (race, gender, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, and employer fixed effects.
Figure A6: Estimated Moderating Effect of Child Age on Associations of Dimensions of Scheduling and Parenting Time.
Notes: Error bars indicate 95% confidence intervals. All models include robust standard errors clustered at the respondent level. Each model run with each indicated variable fully interacted with child age, with 15–18 as the reference category. Sample for each model restricted to (1) # of occurrences of each of clopenings, timing changes, oncall shifts, and canceled shifts ≤ total # of shifts; (2) ≤ 97th percentile of # of shifts; (3) nonmissing on parental stress, parenting time, all independent variables in the main analysis as well as those included in these models, and all control variables to ensure a consistent sample size. All models include month/year fixed effects, demographic characteristics (race, gender, age, education, school enrollment, marital status, income), work characteristics (job tenure, hourly wage, managerial status), state fixed effects, and employer fixed effects.
Footnotes
We leverage a change to the Shift Project surveys that was first implemented in the Spring of 2021 to capture the “intensive margin” of exposure, with frequency of clopenings as a measure of timing and frequencies of timing changes, oncall shifts, and canceled shifts as measures of predictability. For questions about exposure to clopening shifts, timing changes, oncall shifts, and canceled shifts, respondents who indicated “yes” were then asked to select an option for how many times in the past month they experienced the relevant unstable scheduling event, up to “5 or more.” These detailed schedule instability variables are therefore capped at 5 occurrences. Because the question asking for the frequency of each unstable scheduling event is only presented to those who had a value of “Yes” for the dichotomous indicator variable for whether they experience that unstable scheduling event at all in the past month, all those who answered “No” to those questions were assigned a value of 0 for the corresponding numerical variable.
See Footnote 1.
In the Spring 2021 survey wave and earlier, respondents were first asked to provide a number response when asked about the number of shifts they worked in the past month, followed by a question with a list of category options for responses. The order of these two questions was reversed in subsequent survey waves. Therefore, to arrive at a harmonized measure of number of shifts, we use the categorical variable value (recoded to the median of the category) unless a numerical value is available in the Spring 2021 data. For subsequent waves, we use the numerical value unless it is missing, in which case we use the categorical variable value (recoded to the median of the category).
Contributor Information
Siri Neerchal, Harvard University, Department of Sociology, Cambridge, MA, USA.
Daniel Schneider, Harvard University, John F. Kennedy School of Government, Cambridge, MA, USA.
References
- Agrawal Tara, Tracy Jean Farrell Elaine Wethington, and Devine Carol M.. 2018. “‘Doing Our Best to Keep a Routine:’ How Low-Income Mothers Manage Child Feeding with Unpredictable Work and Family Schedules.” Appetite 120 (January): 57–66. 10.1016/j.appet.2017.08.010. [DOI] [PubMed] [Google Scholar]
- Ananat Elizabeth O., and Gassman-Pines Anna. 2021. “Work Schedule Unpredictability: Daily Occurrence and Effects on Working Parents’ Well-Being.” Journal of Marriage and Family 83 (1): 10–26. 10.1111/jomf.12696. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ananat Elizabeth O., Gassman-Pines Anna, and Fitz-Henley John A.. 2022. “The Effects of the Emeryville Fair Workweek Ordinance on the Daily Lives of Low-Wage Workers and Their Families.” RSF: The Russell Sage Foundation Journal of the Social Sciences 8 (5): 45–66. 10.7758/RSF.2022.8.5.03. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bardasi Elena, and Francesconi Marco. 2004. “The Impact of Atypical Employment on Individual Wellbeing: Evidence from a Panel of British Workers.” Social Science & Medicine 58 (9): 1671–88. 10.1016/S0277-9536(03)00400-3. [DOI] [PubMed] [Google Scholar]
- Barnett Rosalind Chait, and Gareis Karen C.. 2007. “Shift Work, Parenting Behaviors, and Children’s Socioemotional Well-Being: A Within-Family Study.” Journal of Family Issues 28 (6): 727–48. 10.1177/0192513X06298737. [DOI] [Google Scholar]
- Baxter Jennifer. 2011. “Flexible Work Hours and Other Job Factors in Parental Time with Children.” Social Indicators Research 101 (2): 239–42. 10.1007/s11205-010-9641-4. [DOI] [Google Scholar]
- Bell David N. F., and Blanchflower David G.. 2021. “Underemployment in the United States and Europe.” ILR Review 74 (1): 56–94. 10.1177/0019793919886527. [DOI] [Google Scholar]
- Bell David N.F., and Blanchflower David G.. 2019. “The Well-Being of the Overemployed and the Underemployed and the Rise in Depression in the UK.” Journal of Economic Behavior & Organization 161 (May): 180–96. 10.1016/j.jebo.2019.03.018. [DOI] [Google Scholar]
- Bolino Mark C., Kelemen Thomas K., and Matthews Samuel H.. 2021. “Working 9-to-5? A Review of Research on Nonstandard Work Schedules.” Journal of Organizational Behavior 42 (2): 188–211. 10.1002/job.2440. [DOI] [Google Scholar]
- Boll Christina, and Schüller Simone. 2023. “The Economic Well-Being of Nonresident Fathers and Custodial Mothers Revisited: The Role of Paternal Childcare.” Journal of Family and Economic Issues 44 (4): 836–53. 10.1007/s10834-022-09876-7. [DOI] [Google Scholar]
- Cai Julie. 2024. “Work Hours Volatility and Child Poverty: The Potential Mitigating Role of Safety Net Programs.” Social Forces 102 (3): 902–25. 10.1093/sf/soad109. [DOI] [Google Scholar]
- Carré Françoise, and Tilly Christopher. 2017. Where Bad Jobs Are Better: Retail Jobs Across Countries and Companies. Russell Sage Foundation. [Google Scholar]
- Cho Hyojin, Susan J Lambert Emily Ellis, and Henly Julia R. 2024. “How Work Hour Variability Matters for Work-to-Family Conflict.” Work, Employment and Society 38 (6): 1611–35. 10.1177/09500170231218191. [DOI] [Google Scholar]
- Choper Joshua, Schneider Daniel, and Harknett Kristen. 2022. “Uncertain Time: Precarious Schedules and Job Turnover in the US Service Sector.” ILR Review 75 (5): 1099–132. 10.1177/00197939211048484. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Christian Leah, Keeter Scott, Purcell Kristen, and Smith Aaron. 2010. “Assessing the Cell Phone Challenge.” Pew Research Center. [Google Scholar]
- Connelly Rachel, and Kimmel Jean. 2011. “The Role of Non-Standard Work Status in Parental Caregiving for Young Children.” Eastern Economic Journal 37 (2): 248–69. doi: 10.1057/eej.2010.45. [DOI] [Google Scholar]
- Craig Lyn, and Powell Abigail. 2012. “Dual-Earner Parents’ Work-Family Time: The Effects of Atypical Work Patterns and Non-Parental Childcare.” Journal of Population Research 29 (3): 229–47. doi: 10.1007/s12546-012-9086-5. [DOI] [Google Scholar]
- Deater-Deckard Kirby. 2004. Parenting Stress. Yale University Press. [Google Scholar]
- Deater-Deckard Kirby, and Panneton Robin, eds. 2017. Parental Stress and Early Child Development. Springer International Publishing. [Google Scholar]
- Fabricius William V., Sokol Karina R., Diaz Priscila, and Braver Sanford L.. 2012. “Parenting Time, Parent Conflict, Parent–Child Relationships, and Children’s Physical Health.” In Parenting Plan Evaluations. Oxford University Press. 10.1093/med:psych/9780199754021.003.0007. [DOI] [Google Scholar]
- Fligstein Neil, and Shin Taek-Jin. 2004. “The Shareholder Value Society: A Review of the Changes in Working Conditions and Inequality in the United States, 1976 to 2000.” In Social Inequality, edited by Neckerman K. Russell Sage Foundation. [Google Scholar]
- Fomby Paula, and Musick Kelly. 2018. “Mothers’ Time, the Parenting Package, and Links to Healthy Child Development.” Journal of Marriage and Family 80 (1): 166–81. doi: 10.1111/jomf.12432. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fugiel Peter J., and Lambert Susan J.. 2019. “On-Demand and On-call Work in the United States.” In Zero Hours and On-Call Work in Anglo-Saxon Countries, edited by O’Sullivan Michelle, Lavelle Jonathan, McMahon Juliette, Ryan Lorraine, Murphy Caroline, Turner Thomas, and Gunnigle Patroc. Springer Publishing. [Google Scholar]
- Gassman-Pines Anna. 2011. “Low-Income Mothers’ Nighttime and Weekend Work: Daily Associations With Child Behavior, Mother-Child Interactions, and Mood.” Family Relations 60 (1): 15–29. doi: 10.1111/j.1741-3729.2010.00630.x. [DOI] [Google Scholar]
- Genadek Katie R., and Hill Rachelle. 2017. “Parents’ Work Schedules and Time Spent with Children.” Community, Work & Family 20 (5): 523–42. doi: 10.1080/13668803.2017.1371672. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Golden Lonnie. 2015. “Irregular Work Scheduling and its Consequences.” Economic Policy Institute; Briefing Paper No. 394. [Google Scholar]
- Goldfarb Samantha, Tarver Will, and Sen Bisakha. 2014. “Family Structure and Risk Behaviors: The Role of the Family Meal in Assessing Likelihood of Adolescent Risk Behaviors.” Psychology Research and Behavior Management 53. doi: 10.2147/PRBM.S40461. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Greenwood Shannon, Perrin Andrew, and Duggan Maeve. 2016. “Social Media Update 2016.” Pew Research Center 11 (2): 1–18. [Google Scholar]
- Haas Linda. 1999. “Families and Work.” In Handbook of Marriage and the Family, edited by Sussman MB, Steinmetz SK, and Peterson GW. Springer Publishing. [Google Scholar]
- Hacker Jacob. 2006. The Great Risk Shift: The New Economic Insecurity and the Decline of the American Dream. Oxford University Press. [Google Scholar]
- Haley-Lock Anna, and Posey-Maddox Linn. 2016. “Fitting It All in: How Mothers’ Employment Shapes Their School Engagement.” Community, Work & Family 19 (3): 302–21. doi: 10.1080/13668803.2015.1023699. [DOI] [Google Scholar]
- Hammons Amber J., and Fiese Barbara H.. 2011. “Is Frequency of Shared Family Meals Related to the Nutritional Health of Children and Adolescents?” Pediatrics 127 (6): e1565–74. doi: 10.1542/peds.2010-1440. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Han Wen-Jui. 2005. “Maternal Nonstandard Work Schedules and Child Cognitive Outcomes.” Child Development 76 (1): 137–54. doi: 10.1111/j.1467-8624.2005.00835.x. [DOI] [PubMed] [Google Scholar]
- Han Wen-Jui, and Fox Liana E.. 2011. “Parental Work Schedules and Children’s Cognitive Trajectories.” Journal of Marriage and Family 73 (5): 962–80. doi: 10.1111/j.1741-3737.2011.00862.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Han Wen-Jui, and Miller Daniel P.. 2009. “Parental Work Schedules and Adolescent Depression.” Health Sociology Review 18 (1): 36–49. doi: 10.5172/hesr.18.1.36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Han Wen-Jui, Miller Daniel P., and Waldfogel Jane. 2010. “Parental Work Schedules and Adolescent Risky Behaviors.” Developmental Psychology 46 (5): 1245–67. doi: 10.1037/a0020178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Harknett Kristen, Schneider Daniel, and Irwin Véronique. 2021. “Improving Health and Economic Security by Reducing Work Schedule Uncertainty.” Proceedings of the National Academy of Sciences 118 (42): e2107828118. doi: 10.1073/pnas.2107828118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Harknett Kristen, Schneider Daniel, and Luhr Sigrid. 2022. “Who Cares If Parents Have Unpredictable Work Schedules?: Just-in-Time Work Schedules and Child Care Arrangements.” Social Problems 69 (1): 164–83. doi: 10.1093/socpro/spaa020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Harknett Kristen, Schneider Daniel, and Wolfe Rebecca. 2020. “Losing Sleep over Work Scheduling? The Relationship between Work Schedules and Sleep Quality for Service Sector Workers.” SSM - Population Health 12: 100681. doi: 10.1016/j.ssmph.2020.100681. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Henly Julia R., and Lambert Susan. 2005. “Nonstandard Work and Child Care Needs of Low- Income Parents.” In Work, Family, Health, and Well-Being, edited by Bianchi S, Casper L, and King R. Lawrence Erlbaum Associates. [Google Scholar]
- Henly Julia R., and Lambert Susan J.. 2014. “Unpredictable Work Timing in Retail Jobs: Implications for Employee Work–Life Conflict.” ILR Review 67 (3): 986–1016. doi: 10.1177/0019793914537458. [DOI] [Google Scholar]
- Henly Julia R., Luke Shaefer H, and Waxman Elaine. 2006. “Nonstandard Work Schedules: Employer- and Employee-Driven Flexibility in Retail Jobs.” Social Service Review 80 (4): 609–34. doi: 10.1086/508478. [DOI] [Google Scholar]
- Honaker James, King Gary, and Blackwell Matthew. 2011. “Amelia II: A Program for Missing Data.” Journal of Statistical Software 45 (7). doi: 10.18637/jss.v045.i07. [DOI] [Google Scholar]
- Hwang Woosang, and Jung Eunjoo. 2020. “Unpartnered Mothers’ Work-Family Conflict and Parenting Stress: The Moderating Effects of Nonstandard Work Schedules.” Journal of Family and Economic Issues 41 (1): 158–71. doi: 10.1007/s10834-019-09647-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jang Soo J. 2008. “Relationships among Perceived Work-Life Balance, Resources, and the Well-being of Working Parents.” Diss., Rutgers The State University of New Jersey. http://www.library.hbs.edu/intra/go/abi.html?url=http://search.proquest.com/dissertations-theses/relationships-among-perceived-work-life-balance/docview/304452874/se-2. [Google Scholar]
- Joshi Pamela, and Bogen Karen. 2007. “Nonstandard Schedules and Young Children’s Behavioral Outcomes Among Working Low-Income Families.” Journal of Marriage and Family 69 (1): 139– 56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kalil Ariel. 2015. “Inequality Begins at Home: The Role of Parenting in the Diverging Destinies of Rich and Poor Children.” In Families in an Era of Increasing Inequality, edited by Amato PR, Booth A, McHale SM, and Van Hook J. Springer International Publishing. [Google Scholar]
- Kalil Ariel, and Ryan Rebecca. 2020. “Parenting Practices and Socioeconomic Gaps in Childhood Outcomes.” The Future of Children 30 (2020): 29–54. doi: 10.1353/foc.2020.0004. [DOI] [Google Scholar]
- Kalil Ariel, Ryan Rebecca, and Corey Michael. 2012. “Diverging Destinies: Maternal Education and the Developmental Gradient in Time With Children.” Demography 49 (4): 1361–83. doi: 10.1007/s13524-012-0129-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kalleberg Arne. 2011. Good Jobs, Bad Jobs: The Rise of Polarized and Precarious Employment Systems in the United States, 1970s to 2000s. Russell Sage Foundation. [Google Scholar]
- Kauhanen Merja, and Nätti Jouko. 2015. “Involuntary Temporary and Part-Time Work, Job Quality and Well-Being at Work.” Social Indicators Research 120 (3): 783–99. doi: 10.1007/s11205-014-0617-7. [DOI] [Google Scholar]
- Kelly Erin L., and Moen Phyllis. 2007. “Rethinking the ClockWork of Work: Why Schedule Control May Pay Off at Work and at Home.” Advances in Developing Human Resources 9 (4): 487–506. doi: 10.1177/1523422307305489. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kelly Erin L., Moen Phyllis, and Tranby Eric. 2011. “Changing Workplaces to Reduce Work-Family Conflict: Schedule Control in a White-Collar Organization.” American Sociological Review 76 (2): 265–90. doi: 10.1177/0003122411400056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kendig Sarah M., and Bianchi Suzanne M.. 2008. “Single, Cohabitating, and Married Mothers’ Time With Children.” Journal of Marriage and Family 70 (5): 1228–40. doi: 10.1111/j.1741-3737.2008.00562.x. [DOI] [Google Scholar]
- Kim Jaeseung, Henly Julia R., Golden Lonnie M., and Lambert Susan J.. 2020. “Workplace Flexibility and Worker Well-Being by Gender.” Journal of Marriage and Family 82 (3): 892–910. doi: 10.1111/jomf.12633. [DOI] [Google Scholar]
- Kim Jaeseung, and Golden Lonnie. 2022. “Inadequacy Inequality: The Distribution and Consequences of Part-Time Underemployment in the US.” Community, Work & Family 25 (1): 84–111. doi: 10.1080/13668803.2021.1985433. [DOI] [Google Scholar]
- Kivimäki Mika, Jokela Markus, Nyberg Solja T., et al. 2015. “Long Working Hours and Risk of Coronary Heart Disease and Stroke: A Systematic Review and Meta-Analysis of Published and Unpublished Data for 603 838 Individuals.” The Lancet 386 (10005): 1739–46. doi: 10.1016/S0140-6736(15)60295-1. [DOI] [PubMed] [Google Scholar]
- Lambert Susan. 2008. “Passing the Buck: Labor Flexibility Practices That Transfer Risk onto Hourly Workers.” Human Relations 61 (9): 1203–27. [Google Scholar]
- Lambert Susan, and Fugiel Peter J.. 2023. “Updating Measures of Work Schedules in Federal Surveys.” 10.31235/osf.io/j3dk5 [DOI] [Google Scholar]
- Lambert Susan J., Fugiel Peter J., and Henly Julia R.. 2014. “Precarious Work Schedules among Early-Career Employees in the US: A National Snapshot.” EINet (Employment Instability, Family Well-Being, and Social Policy Network) at the University of Chicago. [Google Scholar]
- Lambert Susan J., and Haley Anna. 2021. “Implementing Work Scheduling Regulation: Compliance and Enforcement Challenges at the Local Level.” ILR Review 74 (5): 1231–57. doi: 10.1177/00197939211031227. [DOI] [Google Scholar]
- Lambert Susan J., and Henly Julia R.. 2014. “Measuring Precarious Work Schedules.” University of Chicago Crown School of Social Work, Policy, and Practice. [Google Scholar]
- Lambert Susan J., Henly Julia R., and Kim Jaesung. 2019. “Precarious Work Schedules as a Source of Economic Insecurity and Institutional Distrust.” RSF: The Russell Sage Foundation Journal of the Social Sciences 5 (4): 218. doi: 10.7758/rsf.2019.5.4.08. [DOI] [Google Scholar]
- Larson Reed, and Richards Maryse H.. 1991. “Daily Companionship in Late Childhood and Early Adolescence: Changing Developmental Contexts.” Child Development 62 (2): 284–300. [DOI] [PubMed] [Google Scholar]
- Lee Soomi, Hale Lauren, Berger Lawrence M., and Buxton Orfeu M.. 2019. “Maternal Perceived Work Schedule Flexibility Predicts Child Sleep Mediated by Bedtime Routines.” Journal of Child and Family Studies 28 (1): 245–59. doi: 10.1007/s10826-018-1262-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li Jianghong, Johnson Sarah E., Han Wen-Jui, Andrews Sonia, Kendall Garth, Strazdins Lyndall, and Dockery Alfred. 2014. “Parents’ Nonstandard Work Schedules and Child Well-Being: A Critical Review of the Literature.” The Journal of Primary Prevention 35 (1): 53–73. doi: 10.1007/s10935-013-0318-z. [DOI] [PubMed] [Google Scholar]
- Luhr Sigrid, Schneider Daniel, and Harknett Kristen. 2022. “Parenting Without Predictability: Precarious Schedules, Parental Strain, and Work-Life Conflict.” RSF: The Russell Sage Foundation Journal of the Social Sciences 8 (5): 24–44. doi: 10.7758/RSF.2022.8.5.02. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lyden Kelly M. 2020. “Predictive Scheduling Is Trending: Is Milwaukee Next? ” Marquette Benefits & Social Welfare Law Review 22 (1): 111–126. [Google Scholar]
- Masarik April S., and Conger Rand D.. 2017. “Stress and Child Development: A Review of the Family Stress Model.” Current Opinion in Psychology 13: 85–90. doi: 10.1016/j.copsyc.2016.05.008. [DOI] [PubMed] [Google Scholar]
- Milkie Melissa A., Mattingly Marybeth J., Nomaguchi Kei M., Bianchi Suzanne M., and Robinson John P.. 2004. “The Time Squeeze: Parental Statuses and Feelings About Time With Children.” Journal of Marriage and Family 66 (3): 739–61. doi: 10.1111/j.0022-2445.2004.00050.x. [DOI] [Google Scholar]
- Monna Berenice, and Gauthier Anne H.. 2008. “A Review of the Literature on the Social and Economic Determinants of Parental Time.” Journal of Family and Economic Issues 29 (4): 634–53. doi: 10.1007/s10834-008-9121-z. [DOI] [Google Scholar]
- Musick Kelly, and Meier Ann. 2012. “Assessing Causality and Persistence in Associations Between Family Dinners and Adolescent Well-Being.” Journal of Marriage and Family 74 (3): 476–93. doi: 10.1111/j.1741-3737.2012.00973.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Neece Cameron L., Green Shulamite A., and Baker Bruce L.. 2012. “Parenting Stress and Child Behavior Problems: A Transactional Relationship Across Time.” American Journal on Intellectual and Developmental Disabilities 117 (1): 48–66. doi: 10.1352/1944-7558-117.1.48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Neppl Tricia K., Senia Jennifer M., and Brent Donnellan M. 2016. “Effects of Economic Hardship: Testing the Family Stress Model over Time.” Journal of Family Psychology 30 (1): 12–21. doi: 10.1037/fam0000168. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nomaguchi Kei, and Johnson Wendi. 2016. “Parenting Stress Among Low-Income and Working- Class Fathers: The Role of Employment.” Journal of Family Issues 37 (11): 1535–57. doi: 10.1177/0192513X14560642. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Osterman Paul, and Shulman Beth. 2011. Good Jobs America: Making Work Better for Everyone. Russell Sage Foundation. [Google Scholar]
- Pelchat Diane, Bisson Jocelyn, Bois Caroline, and Saucier Jean-Francois. 2003. “The Effects of Early Relational Antecedents and Other Factors on the Parental Sensitivity of Mothers and Fathers.” Infant and Child Development 12 (1): 27–51. doi: 10.1002/icd.335. [DOI] [Google Scholar]
- Petrucci Larissa, Loustaunau Lola, Scott Ellen, and Stepick Lina. 2022. “Persistent Unpredictability: Analyzing Experiences with the First Statewide Scheduling Legislation in Oregon.” ILR Review 75 (5): 1133–58. doi: 10.1177/00197939211064902. [DOI] [Google Scholar]
- Pilarz Alejandra Ros. 2021. “Mothers’ Work Schedule Inflexibility and Children’s Behavior Problems.” Journal of Family Issues 42 (6): 1258–84. doi: 10.1177/0192513X20940761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pilarz Alejandra Ros, and Awkward-Rich Leah. 2024. “Mothers’ Work Schedules and Children’s Time with Parents.” Journal of Family and Economic Issues 45: 117–136. doi: 10.1007/s10834-023-09894-z. [DOI] [Google Scholar]
- Pilarz Alejandra Ros, Cuesta Laura, and Drazen Yonah. 2020. “Nonstandard Work Schedules and Father Involvement Among Resident and Nonresident Fathers.” Journal of Marriage and Family 82 (2): 587–604. doi: 10.1111/jomf.12627. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pilarz Alejandra Ros, and Hill Heather D.. 2017. “Child-Care Instability and Behavior Problems: Does Parenting Stress Mediate the Relationship?: Child-Care Instability and Behavior Problems.” Journal of Marriage and Family 79 (5): 1353–68. doi: 10.1111/jomf.12420. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ponnet Koen, Mortelmans Dimitri, Wouters Edwin, Karla Van Leeuwen Kim Bastaits, and Pasteels Inge. 2013. “Parenting Stress and Marital Relationship as Determinants of Mothers’ and Fathers’ Parenting.” Personal Relationships 20 (2): 259–76. doi: 10.1111/j.1475-6811.2012.01404.x. [DOI] [Google Scholar]
- Presser Harriet B. 1999. “Toward a 24-Hour Economy.” Science 284 (5421): 1778–9 http://www.library.hbs.edu/intra/go/abi.html?url=http://search.proquest.com/scholarly-journals/toward-24-hour-economy/docview/213557778/se-2. [Google Scholar]
- Presser Harriet B. 2000. “Nonstandard Work Schedules and Marital Instability.” Journal of Marriage and Family 62 (1): 93–110. doi: 10.1111/j.1741-3737.2000.00093.x. [DOI] [Google Scholar]
- Puff Jayme, and Renk Kimberly. 2014. “Relationships Among Parents’ Economic Stress, Parenting, and Young Children’s Behavior Problems.” Child Psychiatry Human Development 45 (6): 712–27. [DOI] [PubMed] [Google Scholar]
- Putnick Diane L., Bornstein Marc H., Hendricks Charlene, Painter Kathleen M., Suwalsky Joan T. D., and Andrew Collins W. 2010. “Stability, Continuity, and Similarity of Parenting Stress in European American Mothers and Fathers Across Their Child’s Transition to Adolescence.” Parenting 10 (1): 60–77. doi: 10.1080/15295190903014638. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Quittner Alexandra L., Glueckauf Robert L., and Jackson Douglas N.. 1990. “Chronic Parenting Stress: Moderating Versus Mediating Effects of Social Support.” Journal of Personality and Social Psychology 59 (6): 1266–1278. doi: 10.1037/0022-3514.59.6.1266. [DOI] [PubMed] [Google Scholar]
- Rapoport Benoît, and Bourdais Céline Le. 2008. “Parental Time and Working Schedules.” Journal of Population Economics 21 (4): 903–32. doi: 10.1007/s00148-007-0147-6. [DOI] [Google Scholar]
- Renk Kimberly, Roberts Rex, Roddenberry Angela, et al. 2003. “Mothers, Fathers, Gender Role, and Time Parents Spend with Their Children.” Sex Roles 48: 305–315. 10.1023/A:1022934412910. [DOI] [Google Scholar]
- Rönkä Anna, Malinen Kaisa, Eija Sevón Riitta-Leena Metsäpelto, and May Vanessa. 2017. “Positive Parenting and Parenting Stress Among Working Mothers in Finland, the UK and the Netherlands: Do Working Time Patterns Matter?” Journal of Comparative Family Studies 48 (2): 175–96. doi: 10.3138/jcfs.48.2.175. [DOI] [Google Scholar]
- Roxburgh Susan. 2004. ““There Just Aren’t Enough Hours in the Day’: The Mental Health Consequences of Time Pressure.” Journal of Health and Social Behavior 45 (2): 115–31. doi: 10.1177/002214650404500201. [DOI] [PubMed] [Google Scholar]
- Sanik Margaret Mietus. 1993. “The Effects of Time Allocation on Parental Stress.” Social Indicators Research 30 (2–3): 175–84. doi: 10.1007/BF01078726. [DOI] [Google Scholar]
- Sayer Liana C., Bianchi Suzanne M., and Robinson John P.. 2004. “Are Parents Investing Less in Children? Trends in Mothers’ and Fathers’ Time with Children.” American Journal of Sociology 110 (1): 1–43. doi: 10.1086/386270. [DOI] [Google Scholar]
- Schneider Daniel, and Harknett Kristen. 2019a. “Consequences of Routine Work-Schedule Instability for Worker Health and Well-Being.” American Sociological Review 84 (1): 82–114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schneider Daniel, and Harknett Kristen. 2019b. “It’s About Time: How Work Schedule Instability Matters for Workers, Families, and Racial Inequality.” Shift Project Research Brief. [Google Scholar]
- Schneider Daniel, and Harknett Kristen. 2019c. “What’s to Like? Facebook as a Tool for Survey Data Collection.” Sociological Methods & Research 51 (1): 108–140. doi: 10.1177/0049124119882477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schneider Daniel, and Harknett Kristen. 2021. “Hard Times: Routine Schedule Unpredictability and Material Hardship among Service Sector Workers.” Social Forces 99 (4): 1682–1709. doi: 10.1093/sf/soaa079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schneider Daniel, Hastings Orestes P., and LaBriola Joe. 2018. “Income Inequality and Class Divides in Parental Investments.” American Sociological Review 83 (3): 475–507. [Google Scholar]
- Scott Jacqueline, and Alwin Duane F.. 1989. “Gender Differences in Parental Strain: Parental Role or Gender Role?” Journal of Family Issues 10 (4): 482–503. [Google Scholar]
- Shelleby Elizabeth C. 2018. “Economic Stress in Fragile Families: Pathways to Parent and Child Maladjustment.” Journal of Child and Family Studies 27 (12): 3877–86. doi: 10.1007/s10826-018-1232-z. [DOI] [Google Scholar]
- Sherman Jennifer, and Harris Elizabeth. 2012. “Social Class and Parenting: Classic Debates and New Understandings: Social Class and Parenting.” Sociology Compass 6 (1): 60–71. doi: 10.1111/j.1751-9020.2011.00430.x. [DOI] [Google Scholar]
- Simon Robin W. 1992. “Spouse, Parent, and Worker: Gender, Multiple Role Involvements, Role Meaning, and Mental Health.” Diss., Indiana University. http://www.library.hbs.edu/intra/go/abi.html?url=http://search.proquest.com/dissertations-theses/spouse-parent-worker-gender-multiple-role/docview/304009139/se-2. [Google Scholar]
- Small Stephen A., Eastman Gay, and Cornelius Steven. 1988. “Adolescent Autonomy and Parental Stress.” Journal of Youth and Adolescence 17 (5): 377–91. doi: 10.1007/BF01537880. [DOI] [PubMed] [Google Scholar]
- Sparks Kate, Cooper Cary, Fried Yitzhak, and Shirom Arie. 1997. “The Effects of Hours of Work on Health: A Meta-analytic Review.” Journal of Occupational and Organizational Psychology 70 (4): 391–408. doi: 10.1111/j.2044-8325.1997.tb00656.x. [DOI] [Google Scholar]
- Van der Heijden Franciëlla, Poortman Anne-Rigt, and van der Lippe Tanja. 2016. “Children’s Postdivorce Residence Arrangements and Parental Experienced Time Pressure.” Journal of Marriage and Family 78 (2): 468–81. doi: 10.1111/jomf.12283. [DOI] [Google Scholar]
- Rijswijk Van, Karen Marrie H. J. Bekker, Rutte Christel G., and Croon Marcel A.. 2004. “The Relationships Among Part-Time Work, Work-Family Interference, and Well-Being.” Journal of Occupational Health Psychology 9 (4): 286–95. doi: 10.1037/1076-8998.9.4.286. [DOI] [PubMed] [Google Scholar]
- Virtanen Marianna, and Kivimäki Mika. 2018. “Long Working Hours and Risk of Cardiovascular Disease.” Current Cardiology Reports 20 (11): 123. doi: 10.1007/s11886-018-1049-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Vogtman Julie, and Schulman Karen. 2016. “Set Up to Fail: When Low-Wage Work Jeopardizes Parents’ and Children’s Success.” National Women’s Law Center. [Google Scholar]
- Walther Anna K., and Pilarz Alejandra Ros. 2024. “Associations between Parental Precarious Work Schedules and Child Behavior Problems among Low-income Families.” Journal of Marriage and Family 86 (3): 551–73. doi: 10.1111/jomf.12933. [DOI] [Google Scholar]
- Ward Kaitlin P., and Lee Shawna J.. 2020. “Mothers’ and Fathers’ Parenting Stress, Responsiveness, and Child Wellbeing among Low-Income Families.” Children and Youth Services Review 116: 105218. doi: 10.1016/j.childyouth.2020.105218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wight Vanessa R., Raley Sara. B., and Bianchi Suzanne. M.. 2008. “Time for Children, One’s Spouse and Oneself among Parents Who Work Nonstandard Hours.” Social Forces 87 (1): 243–71. doi: 10.1353/sof.0.0092. [DOI] [Google Scholar]
- Wolfe Julia, Jones Janelle, and Cooper David. 2018. ‘Fair Workweek’ Laws Help More than 1.8 Million Workers. Economic Policy Institute. [Google Scholar]
- Wood Alex J. 2020. Despotism on Demand: How Power Operates in the Flexible Workplace. Cornell University Press. [Google Scholar]
- Wood Clare. 2002. “Parent–Child Pre–School Activities Can Affect the Development of Literacy Skills.” Journal of Research in Reading 25 (3): 241–58. doi: 10.1111/1467-9817.00173. [DOI] [Google Scholar]
- Jean Yeung, W., Sandberg John F., Davis-Kean Pamela E., and Hofferth Sandra L. 2001. “Children’s Time With Fathers in Intact Families.” Journal of Marriage and Family 63 (1): 136–54. doi: 10.1111/j.1741-3737.2001.00136.x. [DOI] [Google Scholar]
- Zhang Baobao, Mildenberger Matto, Howe Peter D., Marlon Jennifer, Rosenthal Seth A., and Leiserowitz Anthony. 2020. “Quota Sampling Using Facebook Advertisements.” Political Science Research and Methods 8 (3): 558–64. doi: 10.1017/psrm.2018.49. [DOI] [Google Scholar]