Which of the following professionals is most likely to have nonstandard work hours?

A large body of research conducted over the past few decades has explored how nonstandard schedules (i.e., employment during nights, evenings, and at weekends) affect partnership quality (hereafter, PQ). The overwhelming consensus is that, by engaging a worker in paid labor during periods of the day and week typically reserved for family life, nonstandard schedules can restrict his or her ability to participate fully in the roles and rhythms of a conjugal relationship. This, in turn, can have negative impacts on partnership cohesion and stability. Compelling evidence of this comes from studies linking nonstandard schedules to a higher risk of union dissolution and conflict (Davis et al., 2008; Kalil et al., 2010; Presser, 2000; White and Keith, 1990) and decreased partnership satisfaction (Maume and Sebastian, 2012; Presser, 2003).

However, despite their discordance with the traditional temporal organization of family life, findings on the effects of nonstandard schedules on PQ are not all negative. A second strand of research has documented that nonstandard schedules can be beneficial to partnerships if they represent a strategic choice by the household. Such was found, for example, in the case of dual-earner parents selecting into desynchronized schedules to more equally distribute childcare between partners (e.g., Begall et al., 2015; Täht and Mills, 2012). Other research has highlighted that nonstandard schedules can promote work-family balance if they come with a high degree of flexibility and predictability, one of the reasons individuals often self-select into these arrangements (e.g., Staines and Pleck, 1986).

In attempting to reconcile these two conflicting strands of findings – one highlighting the negative effects of nonstandard schedules on PQ, the other highlighting the positive impacts of these arrangements – it is important to acknowledge the many channels of selection underlying employment in nonstandard schedules. Work schedule arrangements like night, evening, and weekend work are far from randomly distributed across the population. Rather, entrance into these schedules is influenced by complex and competing preferences and constraints tied to characteristics of an individual's household situation and occupation – factors that are likely not independent of outcomes relating to family cohesion. Moreover, a distinction must be made between those who choose for a variety of reasons to work nonstandard schedules and those on whom these arrangements are imposed by the employer or the nature of the job. Failing to account for these selection mechanisms will produce biased and inconsistent findings and undermine the ability to draw conclusions on the causal effect of nonstandard schedules on PQ.

This study builds on previous research on the effects of nonstandard schedules on PQ by attempting to characterize and correct for these selection processes. Competing theories of positive and negative selection are proposed based on the contention that there will be heterogeneous returns to partnerships from nonstandard schedules (in terms of penalties and benefits) that will depend on how partners have selected into these arrangements. To test these assertions, a potential outcomes framework is adopted, and Mahalanobis Distance Matching techniques are employed on 2010–2016 UK Household Longitudinal Study data to simulate blocked randomization and reduce baseline selection biases. To identify patterns of heterogeneous treatment effects bias, I compare the effect of a nonstandard schedule on PQ amongst those who actually work such arrangements to the effect that would be observed if otherwise equivalent individuals in standard schedules instead worked nonstandard schedules. Analysis is oriented around two main questions. First, net of baseline selection biases, are nonstandard schedules associated with lower PQ for the worker and/or her partner as previous research would suggest? Second, is there evidence of positive or negative selection into these arrangements amongst men or women that could have biased previous results?

Positive selection

Following principles of the rational choice framework, one thesis is that individuals choose to enter a nonstandard schedule only when the expected utility is greater than the expected cost. This is consistent with Becker's (1965) theory of the allocation of time, where decisions within the home surrounding the timing of paid work are grounded in rational deliberations of utility maximization for the actor. Through this lens, individuals choose to engage in paid employment during periods of the

Data

To test these hypotheses, this study uses Waves 2 to 7 (2010–2016) of the UK Household Longitudinal Study (UKHLS). The UKHLS is the largest household panel dataset in Europe, providing a nationally representative probability sample of over 40,000 households in the UK. Given the extensive scope of the UKHLS questionnaire, several variables (including those of interest to the current study) are only measured every other wave of data collection. Consequently, Waves 2, 4, and 6 (collected between

Results

Fig. 1, Fig. 2 plot standardized mean differences (SMD) in the unmatched sample between nonstandard and standard hour and day workers, respectively; for a table of SMDs, see Appendix F. At first glance, one can see some prominent patterns that portend negative selection into nonstandard schedules. For example, relative to control subjects, those engaged in nonstandard hours and days occupy jobs with significantly lower occupational status (ISEI), tend to be in non-salaried positions with longer

Sensitivity analyses

Several covariates included in the matching matrix could plausibly be conceived as causally downstream from nonstandard schedules; thus, matching on these factors may have obscured a portion of the treatment effect. Appendix J presents plotted ATEs estimated from models where potential mediators have been omitted. These are separated into: (1) theorized mediators of a negative effect of nonstandard schedules on PQ (household income, job hours, and whether respondent works a second job); and (2)

Discussion

The current study adopted a potential outcomes framework to test whether the individuals most likely to select into nonstandard work schedules – based on their relative position within the multivariate space of measured confounders – differ not only in their baseline PQ but also in the size and direction of the partnership returns they receive from working these arrangements. Net of the confounding effects of occupational, self-, and socioeconomic selection, I explored first whether nonstandard

Funding

This research was supported by the European Research Council grants 615603 (SOCIOGENOME) and 835079 (CHRONO) in addition to Nuffield College, University of Oxford.

Acknowledgments

I am grateful for the constructive comments received from anonymous reviewers and the Editorial team. I also express gratitude to Melinda Mills, Christiaan Monden, Man Yee Kan, Kadri Täht, Laura Langner, and John Ermisch for providing helpful comments on earlier versions of this manuscript.

Cited by (3)

Recommended articles (6)