Work schedules may fuel precariousness among U.S. workers by undermining perceptions of security, both economic and societal. Volatile hours, limited schedule input, and short advance notice are all dimensions of precarious work schedules. Our analyses suggest that scheduling practices that introduce instability and unpredictability into workers' lives undermine perceptions of security in unique ways for hourly and salaried workers. Although the data suggest that precarious scheduling practices are widespread in the labor market, workers who are black, young, and without a college degree appear to be at highest risk. The findings highlight the importance of examining constellations of scheduling practices and considering the direction of work-hour fluctuations when investigating the ramifications of today's scheduling practices for quality of employment and quality of life.

Keywords

work schedules, job quality, economic insecurity, institutional trust

Employer scheduling practices are part of broader societal transformations in which a growing proportion of social and economic risk is being shouldered by individuals and families rather than firms and government (Appelbaum and Batt 2014; Hacker 2006; Kalleberg 2011; Kalleberg and Vallas 2017; Lambert 2008; Standing 2011; Weil 2014). Across industries, frontline managers have adopted scheduling practices designed to keep labor flexible, facilitating their ability to meet their firm's accountability requirements that restrict outlays for labor. These practices—varying the number and timing of employees' work hours, providing little advance notice of work hours, and offering employees limited input into their work schedule—can undermine job quality by structuring instability and unpredictability into employees' work and personal lives, and if paid by the hour, earnings as well. In this article, we examine the prevalence of a broad set of scheduling practices in the U.S. labor market and consider their ramifications for workers' experiences of insecurity, both economic and societal. [End Page 218]

Work schedules are a defining feature of the quality of employment. As John Robinson and his colleagues note, "Variations in the number of hours that individuals spend working provide important evidence in comparisons of the quality of employment across occupations, countries, and time" (2002, 44). Scholars lament that the decline of collective bargaining in the United States, coupled with increasing inequality in returns to human capital, is exacerbating stratification within the workforce, with spoils concentrated among the few (Kalleberg 2011; Standing 2011). The spoils of labor are wages and hours, and some workers are rich with both but others face a scarcity of hours at poverty wages (Jacobs and Gerson 2004; Schor 1993; Golden 2016; McCrate 2017). In addition to the sheer number of hours, other aspects of work schedules differentiate jobs in terms of their quality. For example, nonstandard timing jobs, which require work outside conventional nine-to-five weekday hours, are widespread and interfere with family roles and worker well-being (Presser 2003; Staines and Pleck 1983). In addition, short advance notice of work hours, regardless of timing or number, makes it difficult for workers to predict when they will need to work, complicating their ability to manage both work and nonwork responsibilities (Clawson and Gerstel 2014; Henly and Lambert 2014; Schneider and Harknett 2016).

Particularly relevant to the current focus is recent research suggesting that fluctuations in weekly hours—within the same job—may be a key source of increasing income volatility among U.S. households and thus a source of economic insecurity. Since 2013, "lack of money/low wages" has competed with "health care costs" as Americans' top response to the question "What is the most important financial problem facing your family today?," surpassing concerns about unemployment or job loss by a 2:1 ratio (Gallup 2018a). Many households are strapped for cash. According to the 2016 Survey of Household Economics and Decisionmaking, fewer than half (48 percent) of adults in the United States have the cash on hand to cover an emergency requiring $400, and 30 percent report that they are either finding it difficult to get by or are just getting by (Board of Governors of the Federal Reserve System 2017).

Instability of work hours may help explain why some households face difficulty paying bills. Recent research demonstrates that hour variations are an important contributor to growing household income volatility in the United States (Farrell and Greig 2016; Finnigan 2018; Gottschalk and Moffitt 2009; Morduch and Schneider 2017). Researchers reason that unstable work hours are thus also a likely source of insecurity both objectively, in terms of spurring financial crises, and subjectively, in terms of fostering uncertainty. Although compelling evidence of these ramification is provided by targeted and qualitative research (Edin and Shaefer 2016; Morduch and Schneider 2017), only recently have representative surveys provided data on the magnitude of work-hour fluctuations and the prevalence of other scheduling practices that may make household finances, and family life, not only unstable but also unpredictable. Notably, new items in recent rounds of the 1997 National Longitudinal Survey of Youth (NLSY97) suggest that hour volatility, short advance notice, and employer-driven schedule control are common among U.S. workers (Lambert et al. 2019). The NLSY97 only surveys young adults (in their mid-twenties to early thirties), however, and does not include measures of perceived economic or societal insecurity. It thus remains unknown how widespread work-hour volatility and other precarious scheduling practices are in the broader U.S. workforce and whether they can help explain workers' experiences of insecurity.

In this article, we capitalize on new and existing questions in the General Social Survey (GSS), a nationally representative survey of U.S. residents, to advance understanding of the prevalence, distribution, and ramifications for insecurity of several dimensions of work schedules that contribute to or detract from job quality. Drawing on new questions in the 2016 GSS that gauge the magnitude of weekly hour fluctuations, length of advance notice, and worker input into weekly hours, we first examine whether problematic aspects of work schedules are differentially experienced by vulnerable subgroups of workers and address the possibility that workers may "age out" of precarious work schedules. We then pool multiple survey years, from 2002 to 2014, to examine the relationship [End Page 219] between fundamental aspects of work schedules and perceived insecurity. These analyses set the stage to consider how the more detailed aspects of work schedules captured in the 2016 GSS contribute new insight into the relationships between scheduling practices and insecurity. Throughout, we differentiate workers paid by the hour versus a salary given that earnings are a direct function of hours for the former but not the latter. Although our primary focus is on economic insecurity, we also explore the possibility that precarious work schedules fuel distrust in major institutions, which we conceptualize as a marker of societal insecurity.

work schedules and perceived insecurity

How work hours are related to perceptions of insecurity may depend on the extent to which time and money are linked in workers' minds. Jeffrey Pfeffer and Sanford DeVoe explain that "Time and money are particularly well-connected in people's minds when they are paid by the hour because their income is then a direct function of the number of hours they work multiplied by their rate of pay" (Pfeffer and DeVoe 2012, 56). Time and money may be more loosely connected among those paid by a salary. By definition, a salary provides financial stability by smoothing income when demand or effort dips. Using data from both the 2002 GSS and the 1988 National Survey of Families and Households, DeVoe and Pfeffer find that income and hourly status interact in explaining perceived happiness, with a significantly larger relationship among workers paid by the hour as compared to those paid in other ways (2009). It seems reasonable to conclude that because the fortunes of hourly workers are more directly tied to number of hours worked, so might be their perceptions of insecurity, both financial and societal.

Even so, fluctuating hours may not always undermine security, even among hourly workers. The ramifications of work-hour fluctuations likely depend on their direction and magnitude; hours can surge up as well as plummet down (Lambert et al. 2019). Ramifications may also depend on whether work-hour fluctuations are by choice, and whether they are predictable. Moreover, as explained in the following section, in some circumstances, fluctuating hours may actually enhance job security, for both workers paid by salary and by the hour (Lehndorff and Voss-Dahm 2005; Perlow 2012), and scheduling practices may unsettle more than family economics. We take these complexities into account in considering how the scheduling practices of key focus here—fluctuating weekly hours, advance notice, schedule input, and irregular timing—may help explain workers' perceptions of economic and societal insecurity.

Fluctuating Hours and Economic Insecurity

Recent research provides evidence that instability in work hours may help account for the growing volatility in U.S. household income. For example, using records of financial transactions recorded by JPMorgan Chase between 2012 and 2015, Diane Farrell and Fiona Greig decompose month-to-month variation in total credits to personal bank accounts into within-job volatility (for example, the amount deposited by an employer), between-job volatility (for example, lapses in paychecks, moves to new employers), and volatility in other sources of income (for example, public programs, retirement plans, additional job) (2016). Variation in earnings within the same job, as opposed to variation accompanying job loss or mobility, accounted for the overwhelming majority of month-to-month variation in labor income. Although their data do not reveal how workers are paid (by hour, salary, or other), Farrell and Greig reason that "volatility in paycheck amounts among weekly paid jobs could therefore be driven by variation in hours worked" (2016, 28). Jonathan Morduch and Rachel Schneider's analysis of data from the U.S. Financial Diaries study, which recorded the financial transactions of 235 low-income and middle-income families across a one-year period (2012 to 2013), offers additional evidence that variation in earners' take-home pay is a central driver of household income volatility; approximately half (47 percent) of month-to-month household income volatility was traced to fluctuations in earnings from the same job (2017). Most recently, Ryan Finnigan shows that working variable hours has become more prevalent since the Great Recession; the cumulative probability of hourly workers responding that [End Page 220] their "hours vary" increased from 37 to 47 percent when comparing the four years before, 2004 to 2007, to the four years after, 2008 to 2012, the Great Recession (2018). Finnigan also finds that this increase in variable work hours largely accounts for the significant increase in earnings instability observed between the same time periods.

Volatility in work hours may not, however, dictate perceived financial insecurity. The meaning of fluctuating hours depends on the nature of the volatility. Small week-to-week fluctuations may do little to inform workers' perceived financial security, and whether work hours surge above or fall below usual or full-time hours makes a difference for household finances. Morduch and Schneider find that, on average, the families in their sample experienced 2.5 months with income at least 25 percent above and 2.5 months with income at least 25 percent below their annual monthly average (2017). Using data from the NLSY97 (2012 to 2013), Susan Lambert and her colleagues find that the majority of variation in weekly hour fluctuations during a one-month period was due to surges above usual hours (2019). Thus, the extent to which fluctuating work hours exacerbate or reduce workers' assessments of financial insecurity may depend on their magnitude and direction, especially among hourly workers whose earnings are most closely tied to the number of hours they work.

The relationship between fluctuating hours and job insecurity is also complex. Workers who incur fluctuations in work hours may do so as a means of demonstrating their commitment to an employer. In many hourly jobs, open availability—the willingness and ability to incur fluctuations in work hours, such as surges and shortfalls—has become a valued form of human capital as managers strive to implement labor flexibility to control outlays for labor (Carré and Tilly 2017; Haley-Lock and Ewert 2011; Lambert 2008; Lehndorff and Voss-Dahm 2005). For example, a study of a national U.S. retailer indicates that hourly sales associates who put constraints on their availability received fewer hours than their more available counterparts, controlling for work-hour preferences (Lambert, Haley-Lock, and Henly 2012). Several studies document that being willing and able to incur fluctuations in work hours is also highly valued in many salaried jobs. For example, Leslie Perlow's research demonstrates how organizations create norms among software engineers and consultants that foster a culture of working through deadlines and answering off-hour calls from bosses and clients (1997, 2012). Face time is used as a marker of employee performance, with career penalties incurred for those unable or unwilling to work on demand and at short notice. Whether penalties and payoffs are real or imagined, perceived job insecurity is typically conceptualized as a subjective uncertainty (see, for example, De Witte and Näswall 2003; Lee, Bobko, and Chen 2006) that has both cognitive and affective components (Huang et al. 2012). In today's workplaces, then, incurring fluctuating hours may be experienced as protection against job insecurity rather than a marker of it, by both hourly and salaried workers.

Schedule Input and Economic Insecurity

Whether fluctuations in the number of work hours exacerbate or mitigate economic insecurity may depend on who determines them. Work hours that are the result of workers' preferences are likely to be experienced as flexibility by workers (Jones 2017; Matos and Galinsky 2011). But when variation is employer driven, workers are more likely to experience it as uncertainty, and we posit, as a source of insecurity (Clawson and Gerstel 2014; Fugiel and Lambert 2019; Henly, Shaefer, and Waxman 2006; Lambert et al. 2019; McCrate 2012). Recent vignette studies—both a controlled experiment of applicants for jobs in a call center and a web-based representative panel—suggest that workers place a great deal of value on avoiding employer control over the timing of work shifts (Mas and Pallais 2016). To our knowledge, the 2016 GSS is the first nationally representative survey in the United States to include items about the extent of employee input into the number and timing of work hours.

In this article, we examine how input into both the number and timing of hours varies by worker characteristics, including age, gender, race, education, and occupation; we then explore how both types of schedule input may help explain perceived economic insecurity [End Page 221] alone and in combination with work-hour volatility. It seems reasonable to expect that lack of input would contribute to financial insecurity, especially when work hours are highly volatile. Whether lack of employee input is related to job insecurity is less clear. On the one hand, employer-driven scheduling may lead workers to feel that they are fulfilling necessary business functions; on the other hand, having no input into their hours may make them feel undervalued and expendable. We do not anticipate that relationships between economic insecurity and schedule input into number or timing of hours will vary for hourly and salaried workers because we do not have a theoretical reason to expect that control is more or less important to insecurity based on how a worker is paid, net of other differences.

Advance Notice and Economic Insecurity

Schedule unpredictability concerns the difficulty workers have anticipating when they will and will not work. Research demonstrates that unpredictability is associated with elevated levels of work stress and work-to-life conflict, and for workers paid by the hour, unpredictable hours mean unpredictable earnings (Clawson and Gerstel 2014; Henly and Lambert 2014). The further in advance workers know when they will need to work, the more certain they can be of hours and earnings. Thus, lengthy advance notice may foster a sense of financial security and short notice may undermine it. Lengthy advance notice may also foster job security by enabling workers to manage personal responsibilities in ways that do not interfere with work responsibilities (Henly and Lambert 2014). The Quality of Work Life supplement to the 2014 GSS and recent rounds of the NLSY97 have included comparable questions on advance schedule notice. In the 2014 GSS, fully 40 percent of hourly workers reported that they know when they will need to work a week or less in advance, the shortest length of notice included as a response category. The percentage of workers reporting a week or less notice is somewhat lower in the NLSY97 (Round 16), with hourly workers (30.3 percent) more likely to receive a week or less notice than nonhourly workers (22.5 percent). The 2016 GSS includes refined response categories that allows us to unpack a week or less advance notice into a day or less, two to three days, or four to seven days. We examine how length of advance notice varies by worker and occupational characteristics and then explore how short advance notice may help explain perceived economic insecurity alone and in combination with hour volatility and schedule input. We anticipate that short notice will increase the probability of financial and job insecurity, especially in the context of hour volatility and low input, and that the relationship to financial insecurity will be especially strong for workers in hourly jobs—because of the additional ramifications of short notice for financial budgeting.

Work Schedules and Societal Insecurity

Precarious scheduling practices also have potential ramifications for noneconomic forms of insecurity. As Richard Sennett observes, "What's peculiar about uncertainty today is that it exists without any looming historical disaster; instead it is woven into the everyday practices of a vigorous capitalism" (1998, 31). Nothing is more everyday in capitalism than work hours and schedules. We posit that the uncertainty introduced into daily life through precarious scheduling practices may fatigue workers' trust broadly. We explore this possibility by examining the relationship between precarious scheduling practices and workers' distrust of societal institutions.

Confidence in many major institutions—from Congress to the clergy—has declined since the 1970s (Gallup 2018b; Pew Research Center 2017). For example, when asked how much confidence "you, yourself have" in Congress, in 1973 (following the Watergate crisis), 42 percent of those polled responded "a great deal" or "quite a lot," but only 11 percent responded similarly in 2018 (Gallup 2018b). Similarly, although 65 percent of poll participants reported confidence in organized religion in 1973, only 38 percent did so in 2018.

Explanations for the increase in institutional distrust are varied. Some authors point to uncertainties introduced through increasing globalization, others stress the dismantling of local community power (for a review, see Abramson and Inglehart 1995); others emphasize increasing income inequality and a belief [End Page 222] in meritocracy (Hayes 2012) or processes of social modernization (Dalton 2005). Building on Sennett (1998), we explore the possibility that distrust in institutions may also spring from more mundane uncertainties. Regardless of how paid, unstable, unpredictable work hours over which workers have little control may introduce uncertainty into the core of work and family life, shaking confidence that societal institutions act in the best interests of people like them.

research questions

  1. 1. What is the prevalence of precarious work schedules in the U.S. labor market? How are distinct features of work schedules distributed across population subgroups by personal and job characteristics? Do workers age out of precarious work schedules?

  2. 2. How are fundamental features of work schedules related to financial insecurity, job insecurity, and institutional distrust?

  3. 3. How are more nuanced features of work schedules related to economic (financial and job) and societal (distrust in institutions) insecurity? Specifically, does the magnitude and direction of work-hour fluctuations, short advance notice and lack of schedule input—alone and in combination—explain perceived insecurity? Do relationships vary for workers paid by the hour and a salary?

methods

The General Social Survey, begun in 1972, is a cross-sectional, nationally representative personal-interview survey of adults age eighteen years or older living in the United States. It uses an equal probability, multistage cluster sample design for selecting housing units in the entire United States.

Data and Sample

We conduct two sets of analyses. First, we combine GSS surveys from 2002, 2006, 2010, and 2014 to examine the relationships between fundamental aspects of work schedules and financial insecurity, job insecurity, and distrust in institutions; the sample includes respondents who were currently in the wage and salary work-force in those years (N = 3,564). Second, we use the 2016 GSS, which incorporated the Fluctuating Work Hours Module (developed by authors Susan Lambert and Julia Henly) in the 3rd ballot of the core survey. This sample is composed of respondents in the wage and salary work-force who responded to the 2016 Module (N = 525). The 2016 data are not included in the combined multiyear data set because questions asking about type of schedule and input into timing are different in 2016, as detailed in our section on measures. We weight variables to improve population representation.

Analytic Approach

To address the first research question, we present descriptive statistics on fundamental aspects of scheduling available in the GSS prior to 2016 using the multiyear pooled data set and on more nuanced scheduling features using data from the 2016 Fluctuating Work Hour Module. With 2016 data, we present personal (gender, race, age, education) and occupational (part time or full time, occupation, union status, level of earnings) subgroup differences. To address the remaining research questions, we estimate a series of linear (indices of financial insecurity and institutional distrust) and logistic (dichotomous variable indicating job insecurity) regressions that sequentially introduce different aspects of work schedules, alone and then in combination. It is not our goal to estimate the relative contribution of each dimension but rather to examine how distinct dimensions and constellations may be differentially related to different types of insecurity, as discussed in the literature review. These models include a block of control variables that capture worker characteristics shown to be associated with labor market opportunities and outcomes, including age (age2), race, gender, education, household income, number of children, spouse and whether spouse works, and whether respondent works more than one job.1 We estimate models separately for hourly and salaried [End Page 223] workers.2 The data are cross-sectional, and our findings can only offer associational knowledge.

We also conduct sensitivity analyses that add occupation, industry, union status, and relative earnings into the main models. To avoid model misspecification, we do not enter these variables in our primary models; conceptually, one path through which occupation, industry, and union membership may affect economic insecurity is through scheduling practices. Sensitivity analyses also include indices capturing workers' overall optimism and hopefulness for achieving life goals to take into account the possibility that workers' individual outlook may color their assessment of their work schedule and also their finances and confidence in institutions. The addition of these variables does little to change our results, and we thus summarize the findings from the sensitivity analyses in the online appendix.

Measures of Insecurity

We define the dependent and key independent variables we employ in the regressions.3 Distributions on all variables are presented in table 1 and table 2, broken out for hourly and salary workers.

Financial insecurity is an index on which higher scores indicate greater insecurity. The index averages respondents' assessments of satisfaction with their current standard of living, their prospects of improving their standard of living, and how they think their standard of living compares to others in America and to their parents (items were standardized before averaging because questions used different response scales). Although this index has modest reliability (alpha = 0.51 in the multiyear and 0.54 in the 2016 GSS), each item added to reliability (that is, subtracting any item would have lowered estimated reliability). Moreover, the results of an exploratory factor analysis are consistent with a one-factor solution; each of these items met the convention of loading on a single factor at the level of 0.4 or higher. Perhaps most important, the overall pattern of results is substantively the same, showing only minor variations, when the individual items rather than the index are used (results available from the authors).

Job insecurity is a dichotomous variable that differentiates workers reporting job insecurity (coded 1) or not (coded 0). It is measured differently in the multiyear and 2016 data. In the multiyear data, it is measured by a single question that asks how true it is that "The job security is good." Workers are coded as job insecure if they respond "not too true" or "not at all true." In the 2016 data, job insecurity combines responses from two survey questions that ask about the extent to which workers worry about the possibility of losing their job and how difficult or easy it would be to find a job as good as their current one. Workers are coded as experiencing job insecurity if they say they worry at least a little about losing their job and they also think it will be fairly or very difficult to find another one.

Institutional distrust is an index, with higher scores indicating greater institutional distrust. The index averages responses to whether "you would say you have a great deal of confidence, only some confidence, or hardly any confidence" in each of a set of institutions. Our index averages together confidence in major companies, education, the executive branch of government, the U.S. Supreme Court, Congress, [End Page 224] banks and financial institutions, organized labor, the press, and banks and financial institutions (alpha = 0.72 in both the multiyear and 2016 GSS).

Measures of Work Schedules in Multiyear GSS Data

Hours worked last week is the total number of hours workers reported working at all jobs in the past week; workers not working because of illness or vacation are asked to report typical hours. Note that this question asks about hours worked at all jobs; 18 percent of hourly workers and 14 percent of salaried workers reported they held more than one job.

Workers are asked whether they usually work a day, afternoon, night, split, irregular–on-call, or rotating shift. Irregular shift is coded 1 if working an irregular–on-call shift and 0 otherwise. Nonregular timing is coded 1 if working anything other than a day, afternoon, or night shift and 0 otherwise.

Little or no input into timing of hours is coded 1 if workers responded rarely or never to a question asking how often they are allowed to change their starting and quitting time on a daily basis.

Irregular with no input is coded 1 when irregular shift is equal to 1 and lack of input into timing is equal to 1, and 0 otherwise.

Nonregular with no input is coded 1 when nonregular timing is equal to 1 and lack of input into timing is also equal to 1, and 0 otherwise.

Measures of Work Schedules in 2016 GSS

Usual hours is the number of hours respondents reported they typically work each week at all jobs; 14 percent of hourly workers and 11 percent of salaried workers in the 2016 GSS reported holding more than one job.

Relative instability measures the magnitude of fluctuations in weekly work hours during the past month, conditioned on usual work hours. It is developed from three survey questions: usual work hours, the most hours worked a week in the past month (including overtime and work at home and other places), and the least hours worked a week in the past month (not including weeks with vacation or sick time). These questions refer to work at all jobs. This measure of magnitude can be likened to a coefficient of variation as it norms the absolute difference of most and least hours by average or usual work hours: [most–least] ÷ usual. For example, an eight-hour difference between most and least weekly hours may have different consequences for workers who usually work forty-eight hours versus sixteen hours a week. In the first case, variation in hours is 17 percent of a worker's usual hours (relative instability ratio of 0.17) whereas in the second case, variation amounts to 50 percent of usual hours (relative instability ratio of 0.50).

Direction of work-hour volatility is assessed with three variables that calculate the proportion of the difference between greatest and fewest weekly hours that indicates an hour shortfall (below usual hours) or an hour surge (above usual and above full-time hours). For example, the proportion of variation due to a surge above full time (which we define as working more than forty-five hours a week) is calculated as [greatest hours–forty-five] ÷ [greatest-fewest]. To conserve space, significant findings related to the direction of volatility are reported in the text but not tables.

Little or no work-hour input is a dichotomous variable that differentiates workers who indicated that the total number of hours they work each week is "decided by my employer with little or no input from me" (coded 1), from workers who indicated more input (coded 0).

No input into work timing is a dichotomous variable that is coded 1 if workers chose the response "Starting and finishing times are decided by my employer and I cannot change them on my own" and 0 if they responded that they can decide the times they work within certain limits or completely on their own.

Nonregular timing is a dichotomous variable that is comparable, but not identical, to the nonregular timing variable constructed using the multiyear data. In the 2016 GSS, we coded workers 0 on nonregular timing if they chose "I have a regular schedule or shift (daytime, evening, or night)" when asked about their usual work schedule in their main job. They are coded 1 if they chose either of the other two alternatives of working a schedule or shift that regularly changes ("for example, from days to evenings or to nights") or one "where working [End Page 225] times are decided at short notice" by their employer.

Week or less notice is a dichotomous variable that differentiates workers who indicated they know what days and hours they will need to work seven or fewer days in advance (coded 1) from workers who reported longer notice or that their schedule never changes (coded 0).

Volatility plus little input into number is coded 1 when relative instability is at least 0.25 and little or no work-hour input is coded 1; other workers are coded 0.

Volatility plus short notice is coded 1 when relative instability is at least 0.25 and week or less notice is coded 1; other workers are coded 0.

Volatility plus little input into timing is coded 1 when relative instability is at least 0.25 and no input into work timing is coded 1; other workers are coded 0.

Short notice plus little input into timing is coded 1 when week or less notice is coded 1 and no input into timing is coded 1; other workers are coded 0.

Short notice plus little input into number is coded 1 when week or less notice is coded 1 and little or no work-hour input is coded 1; other workers are coded 0.

results

In the following section, we present findings on the prevalence of precarious work schedules, first using the multiyear GSS data and then the more nuanced data on work schedules available in the 2016 GSS. We then present findings on the association of these work schedule variables with economic insecurity and distrust in institutions, using the multiyear and the 2016 data.

Prevalence and Distribution of Precarious Scheduling Practices

To address the first research question, we first present the prevalence of the fundamental features of work schedules for workers paid a salary and by the hour, using multiple years of the GSS. We then examine the full set of more nuanced features of work schedules available in the 2016 GSS. We discuss key differences in the distribution of each dimension of work schedules for workers paid a salary and by the hour, and for the 2016 data, how they are distributed across workers with distinct personal and job characteristics. Overall, the data suggest that today's labor market is highly stratified in terms of how much input workers have into the timing and number of work hours, how widely weekly work hours fluctuate, and how far in advance workers know when they will need to work.

Fundamental Features of Work Schedules Using Multiyear GSS Data Set

Using data across multiple years of the GSS, we examine the number of work hours reported in the week prior to the survey, the extent of input into the starting and finishing times of work, and the percentage of respondents who report having a schedule with nonregular work timing (see table 1). Overall, respondents report working an average of 41.6 hours in the prior week, and hourly workers report statistically significantly fewer hours (38.9) than salaried workers (45.6). Hourly workers also report statistically significantly less input into their work schedule timing, with 60 percent of hourly and 37 percent of salaried workers reporting that their employer decides their starting and finishing times without their input. Work schedules that are mostly nonregular, meaning employees work primarily irregular, on-call, split, or rotating shifts, are not highly prevalent in the sample, but nevertheless, characterize an important minority of workers' schedules (9.8 percent of salaried workers and 15.3 percent of hourly workers). The hourly-salary difference is reduced to nonsignificance when the definition of nonregular timing is restricted to only those respondents who work irregular or on-call shifts.

Nonregular timing may be especially difficult for workers without input into work schedule timing. These data suggest, however, that only a small minority of workers have both limited input into the timing of their work schedule and an irregular or on-call shift, although experiencing both together is somewhat more common for hourly workers and when the definition of nonregular is expanded to include irregular, on-call, split, and rotating shifts. The multiyear data illustrate important subgroup differences beyond the hourly-salary distinction [End Page 226]

Table 1. Descriptive Statistics of Variables Included in Regressions with Multiyear Data Source: Authors' calculations based on combined waves of the General Social Survey (), years 2002, 2006, 2010, and 2014. Note: Percentages and means (standard deviations). Significance difference between hourly and salary workers. aIncome in constant dollars, inflation-adjusted to year 2000. ^p < .1; *p < .05; **p < .01
Click for larger view
View full resolution
Table 1.

Descriptive Statistics of Variables Included in Regressions with Multiyear Data

Source: Authors' calculations based on combined waves of the General Social Survey (Smith et al. 2018), years 2002, 2006, 2010, and 2014.

Note: Percentages and means (standard deviations). Significance difference between hourly and salary workers.

aIncome in constant dollars, inflation-adjusted to year 2000.

^p < .1; *p < .05; **p < .01

[End Page 227]

Table 2. Descriptive Statistics of Variables Included in Regressions, 2016 GSS Source: Authors' calculations based on the 2016 General Social Survey (). Note: Percentages and means (standard deviations). Significance difference between hourly and salary workers. ^p &lt; .1; *p &lt; .05; **p &lt; .01
Click for larger view
View full resolution
Table 2.

Descriptive Statistics of Variables Included in Regressions, 2016 GSS

Source: Authors' calculations based on the 2016 General Social Survey (Smith et al. 2018).

Note: Percentages and means (standard deviations). Significance difference between hourly and salary workers.

^p < .1; *p < .05; **p < .01

[End Page 228] in the distribution of these fundamental features of work schedules (see table A1). For brevity, we do not describe these differences here, and instead elaborate important subgroup trends in our discussion of the 2016 descriptive statistics.

Features of Work Schedules in the 2016 GSS

The 2016 data allow us to supplement the work schedule variables available in the multiyear data with items that assess work-hour fluctuations, employee input into the number of hours they work, and the advance notice workers receive about their work schedule. In addition to providing for the first time these more nuanced assessments of work schedule dimensions in a nationally representative sample of U.S. workers, the 2016 data include measures of the usual hours of work each week, work schedule timing, and nonregular hours. These measures are worded slightly differently than prior waves, but tap into the same constructs as those reported in table 1. The overall descriptive statistics of all work schedule variables in the 2016 data are presented in table 2.

Work-Hour Fluctuations

The 2016 data indicate salaried workers report usually working 44.8 hours per week, relative to 38.4 hours per week among workers paid by the hour, a statistically significant difference. As shown in table 3, the overwhelming majority of both hourly (79.2 percent) and salaried (81.6 percent) workers reported at least some fluctuations in weekly work hours during the one month queried in the 2016 GSS. The absolute difference between the most and least number of hours worked during a week of the month averaged more than a full day's work, at 13.2 hours. On average, hours fluctuated by 35 percent of what workers report as their usual hours, with hourly workers (0.37) reporting slightly more relative instability than salaried workers (0.33), though the difference is not statistically significant. The prevalence and magnitude of fluctuating hours observed in the 2016 GSS is comparable to that found in the NLSY97 (Round 16) data; more than 70 percent of the early-career employees in the NLSY97 report some fluctuations in weekly hours (Lambert et al. 2019). In both the GSS (0.35) and NLSY07 (0.34), the relative instability ratios among hourly workers are enough to suggest that fluctuations in number of work hours may play a substantial role in helping account for the earnings volatility observed in recent research, such as the reported 20 percent month to month (Farrell and Greig 2016).

Subgroups vary on number of usual hours and the magnitude of fluctuations in work hours in ways that mostly align with socioeconomic status (for example, lower-status workers report fewer usual hours and more fluctuations) but not always. Differences between hourly and salaried workers are shown for each dimension of work schedules in table 3. Consistent with research (for a review, see Frazis and Stewart 2014), salaried workers report working more hours than hourly workers, and this difference is especially pronounced for whites, higher educated workers, and workers in professional, service, and sales occupations. The difference in usual hours worked is especially large between hourly and salaried workers not covered by a union contract. Workers twenty-six and younger, especially those in hourly jobs, report the largest absolute difference between the most and least number of weekly hours of any group; they are also among the highest in terms of relative instability. The magnitude of fluctuating hours among part-time hourly workers, whether measured in absolute or relative terms, is greater than among full-time hourly workers. Although relative instability varies significantly by race within both hourly and salary groups (significant levels not shown in table), it is white hourly workers who report the greatest relative instability.

Overall, the findings reported in table 3 suggest that a substantial proportion of workers across the U.S. workforce experience sizable fluctuations in weekly work hours, regardless of personal and job characteristics. But knowing the magnitude of the fluctuations may not be enough to understand the conditions under which fluctuations serve to mitigate or exacerbate financial insecurity.

Employee Input into Timing of Weekly Hours

Table 4 indicates that almost two-thirds (64.5 percent) of hourly workers and one-third (33.6 percent) of salaried workers report that the [End Page 229]

Table 3. Descriptives of Usual Hours and Measures of Hour Fluctuations, 2016 GSS Source: Authors' calculations based on the 2016 General Social Survey (). Note. Percentages and means (standard deviations). All percentages are weighted. Total sample size is 525,334 for hourly workers and 191 for salaried workers. ^p &lt; .1; *p &lt; .05; **p &lt; .01
Click for larger view
View full resolution
Table 3.

Descriptives of Usual Hours and Measures of Hour Fluctuations, 2016 GSS

Source: Authors' calculations based on the 2016 General Social Survey (Smith et al. 2018).

Note. Percentages and means (standard deviations). All percentages are weighted. Total sample size is 525,334 for hourly workers and 191 for salaried workers.

^p < .1; *p < .05; **p < .01

[End Page 231] starting and finishing times of their work shifts are decided by their employer. Only 3.5 percent of hourly workers report that they are entirely free to decide their starting and finishing times, versus almost one-fifth (18.1 percent) of salaried workers. Despite these statistically significant differences in schedule input between hourly and salaried workers in the sample overall, several subgroups of workers, even when paid by salary, nevertheless report little to no input into their work schedules. For example, as with their hourly counterparts, large shares of salaried African Americans (64.1 percent), workers twenty-six and younger (49.9 percent), and workers with a high school education or less (49.6 percent) report that starting and finishing times are decided by their employer. Service, sales, and related fields also provide salaried workers limited input into their schedule (60.5 percent). In addition, more than 70 percent of workers in jobs covered by a union contract also report that their start and end times are controlled by their employer whether salaried or paid by the hour. It may be that the timing of shifts in union jobs has been a matter of collective bargaining, but workers see it as solely employer driven because employers are the ones who construct the work schedule.

Employee Input into Number of Weekly Hours

Table 5 shows that almost half of hourly workers and more than one-third of salaried workers report that the total number of hours they work each week is decided by their employer with little or no input from them (see table 5, figures 1 and 2). Salaried workers are more than twice as likely as hourly workers (36.6 versus 16.7 percent) to report controlling the number of hours they work each week either freely or within limits set by their employer. White workers paid by salary are significantly less likely than workers of other races or ethnicities, regardless of how paid, to report little or no input into the number of hours they work (significance levels for comparisons across race not reported on table). At almost 60 percent, black hourly workers are particularly likely to report not having a say into the number of hours they work. The lack of work-hour control that black hourly workers report suggests that the lower relative instability in weekly work hours they incur relative to white workers (see table 3) may be more a matter of employer practice than employee choice.

Input into the number of weekly hours varies in important ways for hourly and salaried workers across occupations and by union status. Workers with the least input into their hours are in hourly administrative support jobs (52 percent) and service and sales jobs (47.2 percent), and construction, production, and transportation jobs whether paid by the hour (53.2 percent) or by salary (50.3 percent). Regarding union status, it is non-unionized hourly workers (48.2 percent) and unionized salaried workers (57.9 percent) who report the least input into the number of hours they work. Again, the unionized salaried workers may have input through collective bargaining that is not reflected in their subjective reports of their input into work hours.

Overall, like fluctuations in work hours, the distribution of work-hour input both for timing and number of hours across subgroups suggests that although a substantial proportion of today's workers may not control the starting and finishing times of their workday and may have limited input into the number of hours they work each week, some groups are clearly at higher risk than others, in ways that mostly mirror broader stratification in the labor market and society.

Advance Schedule Notice

The descriptive trends regarding the advance schedule notice that workers receive indicate marked bifurcation in the labor market. On the one hand, almost half of workers (48.2 percent overall) report either that they know their work schedule four weeks or more in advance or that their schedule never changes, but on the other hand, more than one-third (36 percent) report one week or less advance notice in their work schedule (see figures 3 and 4 and table 6). The relative consistency in workers' reports in the 2016 GSS, the 2014 GSS, and Round 16 of the NLSY97 offers some confidence in these estimates of advance schedule notice, and suggests that experiencing a week or less schedule notice is the norm for an important minority of workers in today's labor market.

The 2016 GSS further breaks down the "one week or less" category (see table 6). These data [End Page 232]

Table 4. Descriptives of Input into Starting and Finishing Times of Work, 2016 GSS Source: Authors' calculations based on the 2016 General Social Survey (). Note: Percentages and means (standard deviations). All percentages are weighted. Total sample size is 525,334 for hourly workers and 191 for salaried workers. ^p &lt; .1; *p &lt; .05; **p &lt; .01
Click for larger view
View full resolution
Table 4.

Descriptives of Input into Starting and Finishing Times of Work, 2016 GSS

Source: Authors' calculations based on the 2016 General Social Survey (Smith et al. 2018).

Note: Percentages and means (standard deviations). All percentages are weighted. Total sample size is 525,334 for hourly workers and 191 for salaried workers.

^p < .1; *p < .05; **p < .01

[End Page 233]

Table 5. Desccriptives of Employee Input into Number of Weekly Work Hours, 2016 GSS Source: Authors' calculations based on the 2016 General Social Survey (). Note: Percentages and means (standard deviations). All percentages are weighted. Total sample size is 525,334 for hourly workers and 191 for salaried workers. ^p &lt; .1; *p &lt; .05; **p &lt; .01
Click for larger view
View full resolution
Table 5.

Desccriptives of Employee Input into Number of Weekly Work Hours, 2016 GSS

Source: Authors' calculations based on the 2016 General Social Survey (Smith et al. 2018).

Note: Percentages and means (standard deviations). All percentages are weighted. Total sample size is 525,334 for hourly workers and 191 for salaried workers.

^p < .1; *p < .05; **p < .01

[End Page 235]

Figure 1. Hourly Employees: Input into Number of Hours Source: Authors' calculations based on the General Social Survey ().
Click for larger view
View full resolution
Figure 1.

Hourly Employees: Input into Number of Hours

Source: Authors' calculations based on the General Social Survey (Smith et al. 2018).

Figure 2. Salaried Employees: Input into Number of Hours Source: Authors' calculations based on the 2016 General Social Survey ().
Click for larger view
View full resolution
Figure 2.

Salaried Employees: Input into Number of Hours

Source: Authors' calculations based on the 2016 General Social Survey (Smith et al. 2018).

show that subgroups of workers are subject to (or in some cases may choose) even less than one week's notice of their work hours. In fact, almost half of those with one week or less notice report that they know when they need to work only a day or less in advance. Hourly workers who are male (28.0 percent), Hispanic (29.7 percent), or have at most a high school degree (20.3 percent) are especially likely to have advance notice in their work schedule of a day or less. Hourly union workers are also subject to last-minute scheduling (25.6 percent). Despite the overall high rates of short notice among hourly workers, salaried workers are not immune to last-minute notice, particularly salaried workers of color; about a quarter of black [End Page 236]

Figure 3. Hourly Employees: Advance Notice Source: Authors' calculations based on the 2016 General Social Survey ().
Click for larger view
View full resolution
Figure 3.

Hourly Employees: Advance Notice

Source: Authors' calculations based on the 2016 General Social Survey (Smith et al. 2018).

Figure 4. Salaried Employees: Advance Notice Source: Authors' calculations based on the 2016 General Social Survey ().
Click for larger view
View full resolution
Figure 4.

Salaried Employees: Advance Notice

Source: Authors' calculations based on the 2016 General Social Survey (Smith et al. 2018).

(25.3 percent) and Hispanic (23.7 percent) workers paid by salary report knowing when they will need to work one day or less in advance. Moreover, an especially high percentage of both hourly (42.3 percent) and salaried (45.4 percent) workers in construction, production, and transportation jobs only know their schedules one day or less in advance.

Intersection of the Dimensions of Work Schedules

As reported earlier, the multiyear data indicate that only a small minority of workers experience the combination of limited input into the timing of their work and working a nonregular (irregular, on-call, rotating, or split shift) schedule. However, the 2016 data indicate that an important percentage of workers experience both [End Page 237]

Table 6. Descripitves of Advance Schedule Notice, 2016 GSS Source: Authors' calculations based on the 2016 General Social Survey (). Note: Percentages and means (standard deviations). All percentages are weighted. Total sample size is 525,334 for hourly workers and 191 for salaried workers. ^p &lt; .1; *p &lt; .05; **p &lt; .01
Click for larger view
View full resolution
Table 6.

Descripitves of Advance Schedule Notice, 2016 GSS

Source: Authors' calculations based on the 2016 General Social Survey (Smith et al. 2018).

Note: Percentages and means (standard deviations). All percentages are weighted. Total sample size is 525,334 for hourly workers and 191 for salaried workers.

^p < .1; *p < .05; **p < .01

[End Page 239] limited input and either short notice or hour volatility. As reported in table 2, 16.2 percent of workers overall report little or no input into the number of hours they work while also working highly fluctuating hours, and more than one-fifth report having no input into the timing of their work in combination with either a volatile work schedule or a week or less notice. Notably, more than one-fourth (25.1 percent) of workers report the "double whammy" of both a volatile work schedule and one week or less of advance notice. In all cases, the likelihood of potentially problematic combinations is greater among hourly workers.

Unpacking further the intersection of very short notice and schedule input may help discern whether unpredictable schedules are employer driven or due to employee preference. Analyses, not reported on the tables, suggest the former is the case among a larger proportion of hourly workers than salaried workers. For example, among salaried workers reporting a day or less of advance notice, only 18.2 percent also report they have little or no say in the number of hours they work, whereas 38.1 percent of hourly workers who report a day or less of notice report they have little say in the number of weekly hours. Strikingly, among white workers who report a day or less of notice, 22.8 percent say that they have little or no input into the number of hours they work, but among black workers reporting a day or less, 69.2 percent say they have little or no input into the number of weekly hours. These patterns provide further evidence that examining the intersection of different dimensions of work schedules may be necessary to discern their meaning in the lives of workers and families and their contributions to inequality, as suggested in prior research (Lambert et al. 2019; McCrate 2012, 2017; Presser 2003).

Do Workers "Age Out" of Precarious Work Schedules?

An examination of the descriptive data on schedule input and advance notice suggests that hourly workers may gain more predictability in their schedules as they age, whereas salaried workers may gain more control over schedule timing and the number of hours they work. For example, the percentage of workers who report a day or less notice increases with age among salaried workers but decreases among hourly workers. The opposite is true of "my schedule never changes;" the proportion of hourly workers who choose this response increases with age and the proportion of salaried workers decreases. On the other hand, salaried workers but not hourly workers, gain control over their work hours as they age. By age fifty-five, more than half of salaried workers report that they can decide starting and finishing times within certain limits or entirely (61.4 percent) and more than one-third of salaried workers (35.3 percent) report that they control the number of hours they work "within limits." Hour fluctuations also dissipate somewhat with age, but less so for hourly workers. As shown in table 3, fluctuations in work hours, as measured by relative instability, drop over the age of fifty-four, except for hourly workers sixty-five or older. Even so, the trends by age also indicate that a substantial proportion of workers, even those fifty-five or older, do not age out of fluctuating work hours.

Fundamental Features of Work Schedules and Perceived Insecurity: Multiyear GSS Data

To address the second research question, we use the multiyear GSS data in regression models to estimate the extent to which fundamental aspects of work schedules can help explain workers' economic insecurity and distrust in institutions, providing a foundation for assessing the contribution the more nuanced measures in the 2016 GSS may make to understanding the relationship between precarious scheduling practices and perceived insecurity.4

The findings pertaining to financial insecurity support the contention that, per Pfeffer and DeVoe, how workers are paid matters (2012). As shown in table 7 (panel A), across all models employing the full sample, hourly workers report significantly greater financial insecurity than salaried workers, even after controlling for [End Page 240] household income and composition. The sheer number of hours respondents report working in the prior week and whether respondents worked an irregular–on-call (or nonregular) schedule are not significantly related to financial insecurity for either hourly or salaried workers, suggesting that these basic work-hour conditions do not effectively differentiate the extent to which workers feel financially vulnerable overall.

Job insecurity shows a different pattern (table 7, panel B). As anticipated, hourly and salaried workers do not significantly differ in their overall assessments of job security, but they do vary in the aspects of work schedules that inform their perceptions of job insecurity. Among hourly workers, the more hours they work in the preceding week, the less insecure they feel in their job (table 7, panel B, models 2 through 7). Although working more hours may help protect hourly workers from a sense of job insecurity, it does not save them from the job insecurity that accompanies working an irregular schedule, especially when those hours are determined by the employer. An irregular or on-call schedule almost doubles the odds (OR = 1.89) of job insecurity among hourly workers, and triples the odds (OR = 3.23) when combined with a lack of input into start and end times, although caution is needed because only 5.4 percent of hourly workers report working an irregular or on-call schedule (see table 1). Working an irregular schedule is also positively associated with distrust in institutions alone (table 7, panel C, model 3) and especially when the irregularity of hours is determined by employer rather than the employee (table 7, panel C, model 6), but only among hourly workers. Working an irregular schedule or having non-regular hours does not appear to contribute to job insecurity or institutional distrust among salaried workers.

In sum, the pattern of relationships between these fundamental aspects of work schedules and economic and societal insecurity begin to flesh out themes we summarize in the literature review—the economic valuation of time, as marked by how workers are paid, seems to shape the meaning of work hours; working more hours may reduce feelings of job insecurity; and irregular or on-call hours, especially when under employer control, may have consequences beyond their economic ramifications.

Scheduling Practices, Economic Insecurity, and Institutional Distrust: 2016 GSS

In regard to the third research question, the new items in the 2016 GSS allow us to examine in greater detail the qualities of work schedules that may place workers at risk of economic insecurity and institutional distrust. All models control for the same set of personal characteristics included in the multiyear analyses.5 In these models, however, we control for respondents' reports of usual work hours rather than hours worked last week. As detailed in table 3, weekly work hours vary a great deal during a one-month period and thus hours worked last week could be the exception rather than the rule.6 We look at hourly and salaried workers separately to provide further insight into how the strength of the connection between work hours and earnings may shape the relationship between different dimensions of work schedules and perceptions of economic and societal insecurity.

Work Schedules and Financial Insecurity

The 2016 data provide additional information on the scheduling practices that contribute to greater financial insecurity among hourly workers, as observed in the multiyear data. Findings are consistent with our proposition that whether fluctuating hours contribute to or detract from job quality depends on the magnitude and direction of the volatility. As shown in table 8 (panel A, model 1), the magnitude of volatility in weekly work hours is negatively related to hourly workers' perceptions of financial [End Page 241]

Table 7. Regressions for Fundamental Characteristics of Work Schedules, Multiyear GSS Source: Authors' calculations based on multiyear General Social Survey data (). Note: N= Full = 3,564; salaried 2,121; hourly 1,443. ^p &lt; .1; *p &lt; .05; **p &lt; .01
Click for larger view
View full resolution
Table 7.

Regressions for Fundamental Characteristics of Work Schedules, Multiyear GSS

Source: Authors' calculations based on multiyear General Social Survey data (Smith et al. 2018).

Note: N= Full = 3,564; salaried 2,121; hourly 1,443.

^p < .1; *p < .05; **p < .01

[End Page 243]

Table 8. Regressions for New Dimensions of Work Schedules, 2016 GSS Source: Authors' calculations based on the 2016 General Social Survey (). Note: Standard errors (SE). Total sample size is 525,334 for hourly workers and 191 for salaried workers. ^p &lt; .1; *p &lt; .05; **p &lt; .01
Click for larger view
View full resolution
Table 8.

Regressions for New Dimensions of Work Schedules, 2016 GSS

Source: Authors' calculations based on the 2016 General Social Survey (Smith et al. 2018).

Note: Standard errors (SE). Total sample size is 525,334 for hourly workers and 191 for salaried workers.

^p < .1; *p < .05; **p < .01

insecurity. Further analyses of the direction of volatility (available from authors) indicates that surges above forty-five hours a week significantly lowers hourly workers' perceptions of financial insecurity; among the 81.6 percent of hourly workers incurring at least some fluctuations in hours, the larger the proportion of fluctuations above forty-five hours, the lower their financial insecurity (b = –0.311, p < .05). Among hourly workers, then, surging up into overtime seems to play a protective role when it comes to financial insecurity.

The opposite picture emerges for workers paid a salary. Although the number of hours worked was not significantly related to salaried workers' financial insecurity in the multiyear data, usual hours is positively associated with financial insecurity among salaried workers in all of the models specified with the 2016 GSS data. The more hours salaried workers report that they usually work, the greater their financial insecurity. This finding is consistent with research indicating that some salaried workers may work long hours out of fear that not doing so will lower their chances for advancement (Perlow 2012).

In addition to the sheer number of work hours, lack of control over the timing (but not the number) of work hours is positively associated with financial insecurity among salaried workers, a relationship also found in the multiyear data. The findings also suggest that less than a week of advance notice plays a substantial role in heightening perceptions of financial [End Page 244] insecurity among salaried workers, by itself (table 8, panel A, model 2) and in combination with volatile work hours (table 9, panel A, model 2), lack of input into start and end times (table 9, panel A, model 4), and lack of input into number of hours (table 9, panel A, model 5). Thus, although much research has focused on the importance of advance notice for hourly workers, these results suggest that short advance notice can undermine the financial security of workers paid a salary.

Work Schedules and Job Insecurity

In the multiyear data, the number of hours hourly workers report working in the last week is negatively associated with job insecurity across all models estimated. In the 2016 data, usual weekly hours is not significantly associated with job insecurity for either hourly or salaried workers (table 8, panel B), and the non-significance holds when hours worked last week, rather than usual hours, is entered in models (not in table). Although the sheer number of hours does not help explain job insecurity in this smaller 2016 data set, other aspects of work schedules do. As in the multiyear data, working a nonregular schedule (that is, a schedule other than a "regular day, afternoon, evening schedule") is positively associated with hourly worker job insecurity (table 8, panel B, model 5). Short advance notice also heightens job insecurity among hourly workers, either alone (table 8, panel B, model 2) or in combination with work-hour volatility (table 9, panel B, [End Page 245]

Table 9. Regressions for Combinations of Scheduling Practices, 2016 GSS Source: Authors' calculations based on the 2016 General Social Survey (). Note: Standard errors (SE). Total sample size is 525,334 for hourly workers and 191 for salaries workers. ^p &lt; .1; *p &lt; .05; **p &lt; .01
Click for larger view
View full resolution
Table 9.

Regressions for Combinations of Scheduling Practices, 2016 GSS

Source: Authors' calculations based on the 2016 General Social Survey (Smith et al. 2018).

Note: Standard errors (SE). Total sample size is 525,334 for hourly workers and 191 for salaries workers.

^p < .1; *p < .05; **p < .01

model 2). Together, these findings suggest that nonregular timing and lack of advance notice contribute to hourly workers' sense of job insecurity, especially when they work highly fluctuating hours.

Again, a different picture surfaces among salaried workers. Rather than contributing to job insecurity, working a nonregular schedule is negatively related to job insecurity among salaried workers (table 8, panel B, model 5). This relationship may reflect the value employers place on being willing and able to work outside standard hours, as discussed in the introduction. Were this the case, however, we would also expect to see a significant negative relationship between the magnitude of work-hour fluctuations and job insecurity, and we do not; although the relationship between relative instability and job insecurity is negative, it is not statistically significant (table 8, panel B, model 1). The only other dimension of work schedules significantly associated with job insecurity among salaried workers is that lack of input into the timing (not the number) is positively associated with job insecurity, though only at the p < .1 level. These results provide further evidence that the nature of relationships between work scheduling practices and perceived [End Page 246] insecurity are likely different for workers paid by hour and salary. Although working a non-regular schedule may impede a sense of job security among hourly workers, it seems to play a protective role among salaried workers.

Work Schedules and Distrust of Institutions

Analyses of the multiyear survey data suggest that working irregular or on-call shifts, especially when they are employer controlled, is associated with distrust in institutions among hourly workers only. Supplemental analyses of the 2016 data help fill in this picture. Although the overall magnitude of fluctuations in weekly hours is not statistically significant in explaining hourly workers' distrust in institutions (table 8, panel C, model 1), those with highly volatile hours (relative instability is at least 0.25) report greater distrust in institutions than those with less volatility (b = 0.106, p < .05, not in table). The direction of the fluctuation matters, however. Workers who report shortfalls in weekly hours (at least 25 percent less than their usual hours) report significantly greater institutional distrust than workers who report smaller shortfalls in weekly hours (available from authors).

Although the multiyear data reveal no significant [End Page 247] associations between work scheduling qualities and distrust in institutions among salaried workers, the 2016 GSS indicates that the more hours fluctuate, the greater salaried workers' distrust of institutions. But, like their hourly counterparts, it is salaried workers who experience shortfalls in hours (at least 25 percent less than their usual) who express the most distrust in societal institutions (available from the authors).

discussion

The sources of uncertainty are expanding in the U.S. labor market. This article considers employer scheduling practices as one determinant of uncertainty that is undermining the quality of jobs and quality of life in the United States. Recent research traces the relationship between fluctuations in weekly work hours to volatility in workers' earnings and household incomes and in turn, to financial insecurity and hardship. In this article, we look at these relationships from the workers' perspective by examining the ramifications of fluctuating work hours, alone and in combination with other scheduling practices, for perceived insecurity, both economic and societal. Our findings suggest that different dimensions of work schedules may serve to undermine, or bolster, hourly and salaried workers' perceptions of financial and job security and their trust in major institutions. Findings also suggest that workers who are black, young, and without a college degree are at highest risk of experiencing problematic combinations of scheduling practices.

The questions commonly used in national surveys to capture the nature of working time may underestimate the prevalence of problematic scheduling practices in today's U.S. labor market because they do not offer insight into several key dimensions—such as the magnitude and direction of work-hour fluctuations, length of advance notice, and amount of input into the number of weekly hours—that may be especially prevalent in today's workplaces. In particular, the new questions on the 2016 General Social Survey suggest that working fluctuating hours is significantly more common than captured by commonly used survey questions. For example, at most, one-fifth of workers would be identified as working fluctuating hours if based on the common question of schedule type (for example, irregular, on-call, or split or rotating shift), whereas more than three-fourths of workers gave different responses when reporting the greatest versus fewest number of hours they worked a week in the past month. For most, these fluctuations were not inconsequential, averaging more than a full day of work and approximately one-third of their usual weekly hours—all within just a one-month period.

The new 2016 GSS questions enabled us to update and unpack the nature of work schedules in the U.S. labor market further by also examining the length of advance schedule notice and input into both timing and hours. Like fluctuating hours, these additional aspects of work schedules have the potential to undermine job quality. We find that two-fifths of hourly and one-third of salaried workers report one week or less advance schedule notice, almost two-thirds of hourly and one-third of salaried workers report that the starting and finishing times of their work shifts are decided by their employer, and almost one-half of hourly and one-third of salaried workers report that employers entirely decide the number of hours they work. Limited advance notice and input into work hours are especially commonplace among black and Hispanic workers in hourly jobs; in addition, although a substantial proportion of women report having precarious work schedules, the data suggest that men are at even greater risk. The data provide evidence that some, but certainly not all, workers age out of precarious work schedules, with hourly workers gaining more predictability as they age and salaried workers gaining more control.

Our analyses reveal several themes worthy of future research. One key theme concerns the importance of taking into account how workers are paid when investigating the implications of work hours for economic insecurity. Findings lend support to Pfeffer and DeVoe's observation that the structure of compensation can prime the economic valuation of time (2012). In our multiyear survey analyses, we find that workers paid by the hour report greater financial insecurity than those paid a salary, after adjusting for covariates. The differing relationships we observe between number of weekly hours and [End Page 248] perceptions of economic insecurity among hourly versus salaried workers further highlight how employer practices that structure the tie between hours and earnings may alter workers' experiences of their work hours. Among hourly workers, the greater the number of usual weekly hours, the lower their job insecurity (multiyear GSS) and notably, work-hour surges above forty-five hours a week seem to further protect hourly workers from experiencing job insecurity (2016 GSS, not in table). This makes sense, given that earnings are a function of hours worked, and, if employers are following the law, hourly workers receive a premium when weekly hours exceed forty. Salaried workers do not receive such a premium, and in the 2016 GSS the number of hours salaried workers worked last week was positively related to financial insecurity across several models. Together, these findings suggest that working long hours may reduce economic insecurity among hourly workers, but may foster, or be a response to, economic insecurity among salaried workers. These analyses are associational, not causal; notably, salaried workers may be working longer hours because they feel insecure, rather than vice versa.

Another theme that emerges from our findings is that fluctuating hours are not always a marker of a poor-quality job. The ramifications of work-hour fluctuations depend on their magnitude and direction. In the 2016 GSS, the greater the volatility in weekly hours, the lower hourly workers' financial insecurity—especially when the majority of hour fluctuations were due to surges above forty-five hours a week. Thus, questions in surveys that ask about hour variations as a yes or no job characteristic or as a particular schedule type provide limited insight into the conditions under which fluctuating hours matter for workers' lives. Future research is needed to examine how the magnitude and direction of fluctuating hours are related to nonfinancial aspects of life. We find that working fluctuating hours is positively associated with workers' distrust of societal institutions, among both hourly (multiyear GSS) and salaried (2016 GSS) workers. Further analyses revealed that the direction of the fluctuations mattered, with distrust being highest among workers who reported a substantial shortfall in weekly hours. The magnitude and direction of work-hour fluctuations may matter for family life as well. For example, although surges in work hours may help protect workers from financial insecurity, they may complicate care-giving and create stress (Henly and Lambert 2014).

The findings of this study also highlight the usefulness of considering scheduling practices in combination with one another. In the multiyear data, the combination of working an irregular or on-call schedule plus little input into the timing of work increased the odds of job insecurity among hourly and salaried workers. In the 2016 data, although work-hour volatility is not significantly related to financial insecurity among salaried workers when examined alone, the combination of hour volatility plus either short advance notice or lack of input into schedule timing are both positively associated with these workers' perceptions of financial insecurity. These findings add to evidence that fluctuations in work hours can be experienced as flexibility or as instability, depending on whether they are employee versus employer driven (Fugiel and Lambert 2019; Henly, Shaefer, and Waxman 2006).

The results suggest that advance notice is a salient aspect of job quality not only among hourly workers, but among salaried workers as well. The budding literature on advance notice has focused on the implications of schedule unpredictability mostly among low-paid hourly workers (Henly and Lambert 2014; Schneider and Harknett 2016). Daniel Clawson and Naomi Gerstel's recent examination of predictability in health-care settings, however, reminds us that it is not that unpredictability is unimportant to higher-status workers but that they are better equipped to avoid it (2014). The results of this study confirm that the length of advance schedule notice is a marker of job quality in hourly jobs. Among hourly workers, less than a week's notice is positively related to financial insecurity, when accompanied by a lack of input into the timing of hours, and job insecurity, when accompanied by volatile hours. Our findings also suggest that length of advance notice is a marker of the quality of salaried positions as well. Among salaried workers, a week or less of notice is positively associated with financial [End Page 249] insecurity when examined alone and in combination with work-hour volatility and lack of input into both the number and timing of work hours.

The scheduling practices that undermine job quality can have consequences for communities and society. In this article, we explore the possibility that work schedules may be unsettling beyond their economic ramifications by undermining confidence in societal institutions. Our findings suggest that experiencing shortfalls in weekly hours may help cement the growing distrust in key societal institutions observed in the United States over several decades. This finding raises the possibility that growing rates of involuntary part-time employment may exacerbate American's distrust in institutions, and perhaps in one another. Such possibilities seem worthy of further investigation.

We remind readers that the observed associations, and the lack thereof, come from exploratory analyses. Although in some analyses we are able to capitalize on the larger sample afforded by pooling multiple years of the GSS, these years included only a handful of questions on work schedules. The 2016 GSS provided data on a richer set of work scheduling practices, however the one-year sample limited statistical power. Moreover, the very nature of our dependent variables set the bar high in terms of identifying the ramifications of scheduling practices. The questions making up our index of financial insecurity capture workers' assessment of overall economic well-being and standard of living, rather than the more tangible aspects of financial hardship that research suggests can occur when workers' hours vary at the behest of their employer and with little time to adjust expenses or budgets (Morduch and Schneider 2017). Examining the relationship between everyday scheduling practices and what seems a fairly distal outcome—confidence in societal institutions—is similarly ambitious.

Even with these conceptually ambitious measures, the results provide evidence that scheduling practices that introduce instability and unpredictability into workers' lives—volatile work hours, little input into the timing and number of hours, and short advance notice—undermine the quality of many American jobs. From a research standpoint, the findings attest to the merits of examining how different dimensions of work schedules, both fundamental and more nuanced aspects, operate in tandem to affect workers' assessments of their lives and livelihoods. The widespread prevalence of work-hour volatility, short advance notice, and limited input into the number and timing of hours—and the concentration of these among marginalized subgroups—suggest the merits of further study of the conditions under which, and for whom, these scheduling practices heighten insecurity and distrust. [End Page 250]

Susan J. Lambert

Susan J. Lambert is associate professor in the School of Social Service Administration at the University of Chicago.

Julia R. Henly

Julia R. Henly is professor in the School of Social Service Administration at the University of Chicago.

Jaeseung Kim

Jaeseung Kim is assistant professor in the College of Social Work at the University of South Carolina.

Direct correspondence to: Susan Lambert at slambert@uchicago.edu, University of Chicago, School of Social Service Administration, 969 E. 60th St., Chicago, IL 60637.
Table A1. Descriptives of Fundamental Scheduling Practices by Subgroups Source: Authors' calculations based on multiyear General Social Survey data (). Note: All percentages are weighted. N = 3,564; salaried 2,121; hourly 1,443. ^p &lt; .1; *p &lt; .05; **p &lt; .01
Click for larger view
View full resolution
Table A1.

Descriptives of Fundamental Scheduling Practices by Subgroups

Source: Authors' calculations based on multiyear General Social Survey data (Smith et al. 2018).

Note: All percentages are weighted. N = 3,564; salaried 2,121; hourly 1,443.

^p < .1; *p < .05; **p < .01

Table A2. Relationship Between Control Variables and Dependent Measures Source: Authors' calculations based on the General Social Survey (). ^p &lt; .1; *p &lt; .05; **p &lt; .01
Click for larger view
View full resolution
Table A2.

Relationship Between Control Variables and Dependent Measures

Source: Authors' calculations based on the General Social Survey (Smith et al. 2018).

^p < .1; *p < .05; **p < .01

references

Abramson, Paul, and Ronald Inglehart.1995. Value Change in Global Perspective. Ann Arbor: University of Michigan Press.
Appelbaum, Eileen, and Rosemary Batt. 2014. Private Equity at Work: When Wall Street Manages Main Street. New York: Russell Sage Foundation.
Board of Governors of the Federal Reserve System. 2017. "Report on the Economic Well-Being of U.S. Households in 2016." Accessed April 5, 2018. https://www.federalreserve.gov/publications/files/2016-report-economic-well-being-us-households-201705.pdf.
Carré, Francoise, and Chris Tilly. 2017. Where Bad Jobs Are Better: Why Retail Jobs Differ across Countries and Companies. New York: Russell Sage Foundation.
Clawson, Dan, and Naomi Gerstel. 2014. Unequal Time: Gender, Class, and Family in Employment Schedules. New York: Russell Sage Foundation.
Dalton, Russell J. 2005. "The Social Transformation of Trust in Government." International Review of Sociology 15(1): 133–54.
DeVoe, Sanford E., and Jeffrey Pfeffer. 2009. "When Is Happiness About How Much You Earn? The Effect of Hourly Payment on the Money–Happiness Connection." Personality and Social Psychology Bulletin 35(12): 1602–18.
De Witte, Hans, and Katharina Näswall. 2003. "'Objective' versus 'Subjective' Job Insecurity: Consequences of Temporary Work for Job Satisfaction and Organizational Commitment in Four European Countries." Economic and Industrial Democracy 24(2): 149–88.
Edin, Kathryn J., and H. Luke Shaefer. 2016. $2.00 a Day: Living on Almost Nothing in America. New York: First Mariner Books.
Farrell, Diane, and Fiona Greig. 2016. "Paychecks, Paydays, and the Online Platform Economy: Big Data on Income Volatility." New York: J. P. Morgan Chase. Accessed March 15, 2018. https://www.jpmorganchase.com/corporate/institute/document/jpmc-institute-volatility-2-report.pdf.
Finnigan, Ryan. 2018. "Varying Weekly Work Hours and Earnings Instability in the Great Recession." Social Science Research 74 (August): 96–107.
Frazis, Harley, and Jay Stewart. 2014. "Is the Workweek Really Overestimated?" Monthly Labor Review 6 (June): 1–15.
Fugiel, Peter J., and Susan J. Lambert. 2019. "On-Demand and On-Call Work in the United States." In Zero Hours and On-Call Work in Anglo-Saxon Countries, edited by Michelle O'Sullivan, Jonathan Lavelle, Juliette McMahon, Lorraine Ryan, Caroline Murphy, Thomas Turner, and Patroc Gunnigle. New York: Springer Publishing.
Gallup Organization. 2018a. "Economy." Polls Between 2013 and 2018. Accessed April 13, 2018. http://news.gallup.com/poll/1609/Consumer-Views-Economy.aspx.
———. 2018b. "Confidence in Institutions." Polls Between 1937 and 2018. Accessed July 5, 2018. https://news.gallup.com/poll/1597/confidence-institutions.aspx.
Golden, Lonnie. 2016. "Still Falling Short on Hours and Pay: Part-time Work Becoming New Normal." Washington, D.C.: Economic Policy Institute. Accessed February 27, 2019. https://www.epi.org/publication/still-falling-short-on-hours-and-pay-part-time-work-becoming-new-normal.
Gottschalk, Peter, and Robert Moffitt. 2009. "The Rising Instability of U.S. Earnings." Journal of Economic Perspectives 23(4): 3–24.
Hacker, Jacob. 2006. The Great Risk Shift: The Assault on American Jobs, Families, Health Care, and Retirement and How You Can Fight Back. New York: Oxford University Press.
Haley-Lock, Anna. and Stephanie Ewert. 2011. "Serving Men and Mothers: Workplace Practices and Workforce Composition in Two U.S. Restaurant Chains and States." Community, Work & Family 14(4): 387–404.
Hayes, Christopher. 2012. Twilight of the Elites: America after Meritocracy. New York: Random House.
Henly, Julia R., and Susan J. Lambert. 2014. "Unpredictable Work Timing in Retail Jobs: Implications for Employee Work-life Outcomes." Industrial and Labor Relations Review 67(3): 986–1016.
Henly, Julia R., H. Luke Shaefer, and R. Elaine Waxman. 2006. "Nonstandard Work Schedules: Employer-and Employee-Driven Flexibility in Retail Jobs." Social Service Review 80(4): 609–34.
Huang, Guo-hua, Xiongyinig Niu, Cynthia Lee, and Susan J. Ashford. 2012. "Differentiating Cognitive and Affective Job Insecurity: Antecedents and Outcomes." Journal of Organizational Behavior 33(6): 752–69.
Jacobs, Jerry A., and Kathleen Gerson. 2004. The Time Divide: Work, Family and Gender Inequality. Cambridge, Mass.: Harvard University Press.
Jones, Kerry. 2017. "The Most Desirable Employee Benefits." Harvard Business Review, February 15, 2017. Accessed February 27, 2019. https://hbr.org/2017/02/the-most-desirable-employee-benefits.
Kalleberg, Arne L. 2011. Good Jobs, Bad Jobs: The Rise of Polarized and Precarious Employment Systems in the United States, 1970s to 2000s. New York: Russell Sage Foundation.
Kalleberg, Arne L., and Steven P. Vallas. 2017. Precarious Work. Bingley, UK: Emerald Publishing.
Lambert, Susan J. 2008. "Passing the Buck: Labor Flexibility Practices that Transfer Risk onto Hourly Workers." Human Relations 61(9): 1203–27.
Lambert, Susan J., Anna Haley-Lock, and Julia R. Henly. 2012. "Schedule Flexibility in Hourly Jobs: Unanticipated Consequences and Promising Directions." Community, Work & Family 15(3): 293–315.
Lambert, Susan J., Julia R. Henly, Peter Fugiel, and Joshua Choper. 2019. "The Magnitude and Meaning of Work Hour Volatility among Early-Career Employees in the US." Unpublished manuscript, University of Chicago.
Lee, Cynthia, Philip Bobko, and Zhen Xiong Chen. 2006. "Investigation of the Multidimensional Model of Job Insecurity in Two Countries." Applied Psychology: An International Review 55(4): 167–95.
Lehndorff, Steffen, and Dorothea Voss-Dahm. 2005. "The Delegation of Uncertainty: Flexibility and the Role of the Market in Service Work," In Working in the Service Sector: A Tale from Different Worlds, edited by Gerhard Bosh and Steffen Lehndorff. New York: Routledge.
Mas, Alexandre, and Amanda Pallais. 2016. "Valuing Alternative Work Arrangements." NBER working paper no. 22708. Cambridge, Mass.: National Bureau of Economic Research. Accessed February 27, 2019. http://ww.nber.org/papers/w22708.
Matos, Kenneth, and Ellen Galinsky. 2011. "Workplace Flexibility in the United States: A Status Report." New York: Families and Work Institute.
McCrate, Elaine. 2012. "Flexibility for Whom? Control Over Work Schedule Variability in the U.S." Feminist Economics 18(1): 39–72.
———. 2017. "Unstable and On-Call Work Schedules in the United States and Canada." Conditions of Work and Employment Series no. 99. Geneva: International Labour Office. Accessed February 27, 2019. https://www.ilo.org/wcmsp5/groups/public/---ed_protect/---protrav/---travail/documents/publication/wcms_619044.pdf.
Morduch, Jonathan, and Rachel Schneider. 2017. The Financial Diaries: How American Families Cope in a World of Uncertainty. Princeton, N.J.: Princeton University Press.
Perlow, Leslie A. 1997. Finding Time: How Corporations, Individuals, and Families Can Benefit from New Work Practices. Ithaca, N.Y.: Cornell University Press.
———. 2012. Sleeping with Your Smartphone. Cambridge, Mass.: Harvard Business School Publishing.
Pew Research Center. 2017. "Sharp Partisan Divisions in View of National Institutions." Washington, D.C.: Pew Research Center, U.S. Politics & Policy. Accessed February 27, 2019. http://www.people-press.org/2017/07/10/sharp-partisan-divisions-in-views-of-national-institutions.
Pfeffer, Jeffrey, and Sanford E. DeVoe. 2012. "The Economic Evaluation of Time: Organizational Causes and Individual Consequences." Research in Organizational Behavior 32(1): 47–62.
Presser, Harriet B. 2003. Working in a 24/7 Economy: Challenges for American Families. New York: Russell Sage Foundation.
Robinson, John P., Alain Chenu, and Anthony S. Alvarez. 2002. "Measuring the Complexity of Hours at Work: The Weekly Work Grid." Monthly Labor Review 125(4): 44–54.
Schneider, Daniel, and Kristen Harknett. 2016. "Schedule Instability and Unpredictability and Worker Health and Wellbeing." Working paper. Washington, D.C.: Washington Center for Equitable Growth. Accessed February 27, 2019. http://cdn.equitablegrowth.org/wp-content/uploads/2016/09/12135618/091216-WP-Schedule-instability-and-unpredictability.pdf.
Schor, Juliet. 1993. The Overworked American: The Unexpected Decline of Leisure. New York: Basic Books.
Sennett, Richard. 1998. The Corrosion of Character: The Personal Consequences of Work in the New Capitalism. New York: W. W. Norton.
Smith, Tom W., Michael Davern, Jeremy Freese, and Michael Hout. 2018. General Social Surveys, 1972–2016 [machine-readable data file]. Chicago: NORC, University of Chicago. Data available from the GSS Data Explorer website at http://gssdataexplorer.norc.org (accessed April 8, 2019).
Staines, Graham L., and Joseph H. Pleck. 1983. The Impact of Work Schedules on the Family. Ann Arbor, Mich.: Institute for Social Research.
Standing, Guy. 2011. The Precariat: The New Dangerous Class. London: Bloomsbury Academic.
Weil, David. 2014. The Fissured Workplace: Why Work Became So Bad for So Many and What Can Be Done to Improve It. Cambridge, Mass.: Harvard University Press.

Online Appendix

Precarious Work Schedules as a Source of Economic Insecurity and Institutional Distrust

Susan J. Lambert, Julia R. Henly, and Jaeseung Kim

I. Primary Survey Questions

GSS Multi-year
Work hours last week

Last week were you working full time, part time, going to school, keeping house, or what? A. IF WORKING, FULL OR PART TIME: How many hours did you work last week, at all jobs?

Work timing input

How often are you allowed to change your starting and quitting times on a daily basis? 1= Often 2= sometimes 3=rarely 4=never

Non-regular work schedule

Which of the following best describes your usual work schedule?

  • Day shift

  • Afternoon shift

  • Night shift

  • Split shift

  • Irregular shift/on-call

  • Rotating shifts

Financial insecurity index

Four survey questions are used to construct the financial insecurity index.

Financial (Dis)satisfaction: So far as you and your family are concerned, would you say that you are pretty well satisfied with your present financial situation, more or less satisfied, or not satisfied at all?

No better than parents: Compared to your parents when they were the age you are now, Do you think your own standard of living now is much better, somewhat better, about the same, somewhat worse, or much worse than theirs was?

No better in future: The way things are in America, people like me and my family have a good chance of improving our standard of living --do you agree or disagree? (strongly agree, agree, neither, disagree, strongly disagree)

Relative family income: Compared with American families in general, would you say your family income is far below average, below average, average, above average, or far above average? (PROBE: Just your best guess.)

Job insecurity

Now I'm going to read you another list of statements about your main job. For each, please tell me if the statement is very true, somewhat true, not too true, or not at all true with respect to the work you do. K. The job security is good

Distrust of Institutions

Eight survey questions from the same stem question are used to construct the measure of distrust in institutions.

I am going to name some institutions in this country. As far as the people running these institutions are concerned, would you say you have a great deal of confidence, only some confidence, or hardly any confidence at all in them? Confidence in… major companies / in education/executive branch of federal government/organized labor / press / US supreme court/congress/banks and financial institution

GSS 2016
Usual Hours

How many hours a week do you usually work, at all paid jobs?

Relative instability (magnitude of fluctuating work hours):

Three survey questions are used to construct the measure of relative instability: (most-least)/usual.

  1. 1. How many hours a week do you usually work, at all paid jobs?

  2. 2. In the last month, what is the greatest number of hours you worked in a week, at all paid jobs? Please consider all hours, including any extra hours, overtime, work you did at home for your job, and time you spent on work that may not have been directly billable or compensated.

  3. 3. In the last month, what is the fewest number of hours you worked in a week, at all paid jobs? Please do not include weeks in which you missed some or all hours because of illness,

Advance notice

How far in advance do you usually know what days and hours you will need to work?

  • One day or less in advance

  • 2 to 3 days in advance

  • 4 to 7 days in advance

  • between 1 and 2 weeks in advance

  • between 3 and 4 weeks in advance

  • more than 4 weeks in advance

  • My schedule never changes.

Work timing input

Which of the following statements best describes how your working hours are decided? (By working hours we mean here the times you start and finish work, and not the total hours you work per week or month.)

  • Starting and finishing times are decided by my employer and I cannot change them on my own

  • I can decide the time I start and finish work, within certain limits, or

  • I am entirely free to decide when I start and finish work?

Work hour input

Which of the following statements best describes how your working hours are decided? In this question, working hours refers to the total number of hours you work each week, not the time you start and finish work each day.

  • The total number of hours I work each week is decided by my employer with little or no input from me

  • The total number of hours I work each week is decided by my employer but with my input.

  • I can decide how many hours I work each week, within certain limits.

  • How many hours I work a week depends on things outside of my control and outside of my employer's control.

Non-regular Work Schedule

Which of the following statements best describes your usual working schedule in your main job?

  • I have a regular schedule or shift (daytime, evening, or night)

  • I have a schedule or shift which regularly changes (for example, from days to evenings or to nights)

  • I have a schedule where daily working times are decided at short notice by my employer

Financial insecurity index

Same as Multi-Year GSS

Job insecurity

Two survey questions are used to construct the measure of job insecurity.

To what extent, if at all, do you worry about the possibility of losing your job? 1=worry great deal, 2=worry some extent, 3=worry a little, 4=don't worry

How difficult or easy do you think it would be for you to find a job at least as good as your current one? 1=very easy, 2=fairly easy, 3=neither, 4=fairy difficult, 5=very difficult

Distrust of Institutions

Same as Multi-Year GSS

II. Definitions of control variables

Age and age2 is a continuous variable indicating the worker's age in years.

In the multi-year data, race is captured with a set of dichotomous variables identifying workers as white, black, Hispanic, or other; white is the reference group. Given the smaller cell-sizes in the 2016 data, race is a dichotomous variable that is coded 1 for workers of color and 0 for whites.i

Female is coded 1; others are coded 0.

High school or less is coded 1 if the worker has at most a high school degree; workers with more education are coded 0.

Household income is a set of dichotomous variables (see table 1a and table 1b). Workers reporting incomes in the middle category of the distributions are the reference group.

No partner/spouse is a dichotomous variable that is coded 1 if the worker does not have a spouse or partner; Spouse doesn't work is coded 1 if the spouse/partner is not employed; the reference group with these two variables in the model is workers with a working spouse/partner.

Number of children 18 or younger is the total number of children in the household.

Multiple jobs is a dichotomous variable indicating whether (coded 1) or not (coded 0) the workers holds more than one job.

Low pay is a dichotomous variable that differentiates workers whose job earnings are in the lower third (coded 1) or upper two-thirds (coded 0) of the earnings distribution in the GSS sample (normed separately for hourly and salaried workers, and in the multi-year and 2016 data).

Union is a dichotomous variables that differentiates workers who said they belong to a labor union (coded 1) from those who do not (coded 0).

Occupation is reported in four categories: office/administrative/management/business; professional and related occupations; service and sales; and construction/production/transportation/natural resources.

III. Summary of results of sensitivity analyses

The key relationships that are statistically significant in the main models using both the multiyear and the 2016 GSS hold when subjected to several robustness tests that control for labor market factors (occupation, industry, union, and earnings) as well as individual characteristics (two measures of respondents' subjective outlook, i.e., optimism and hopeii). (Results are summarized in the online appendix.)

In some models, relationships become stronger with controls. For example, among salaried workers, short notice and lack of input into the starting and finishing times of the workday move from marginally to statistically significant at a p<.05 level when occupation and industry (and union status in the case of input into timing) are taken into account. Similarly, among salaried workers, the marginally significant relationship between financial insecurity and the 'double whammy' of having both short notice and limited input into timing becomes more significant when occupation and industry are controlled, and remains marginally significant with union membership and earnings controlled in the model. This relationship no longer holds, however, when the model is adjusted for the personal outlook variables (either hope or optimism).

The results from the multi-year data are robust to sensitivity tests as well, with two exceptions. For hourly workers, irregular shift becomes significantly related to financial insecurity when occupation and industry are controlled in the regressions. However, the marginally significant relationship between irregular shift and job insecurity that is observed in the main model is no longer significant with occupation and industry included.

Footnotes

The authors are grateful to the Russell Sage Foundation for supporting this research and to David Howe, Arne Kalleberg, and reviewers for helpful suggestions.

1. Definitions of control variables are included in the online appendix (https://www.rsfjournal.org/content/5/4/218/tab-supplemental).

2. Little data were missing on the independent and control variables in the multiyear and 2016 GSS. The most data were missing on household income, which had 7.8 percent missing in the multiyear data (and 14 percent missing on respondents' earnings) and 5.5 percent missing (16 percent missing on respondents' earnings) in the 2016 GSS. Missing data on the remaining independent and control variables ranged from 1 to 2 percent. Because of the low proportion of missing responses on any one variable, we simply filled in missing data on the independent and control variables with the mean, normed separately for hourly and salaried subsamples as appropriate.

3. An online appendix presents the survey questions included in the Fluctuating Work Hours Module as well as the questions on economic insecurity and institutional distrust (https://www.rsfjournal.org/content/5/4/218/tab-supplemental). We constructed numerous alternative measures of both financial insecurity and fluctuating work hours to estimate sensitivity of findings to measurement decisions. The significant and statistically non-significant findings reported in this article largely hold regardless of measure or model.

4. Table A2, panel A, reports associations between the control variables and financial insecurity, job insecurity, and institutional distrust included in analyses of the multiyear data.

5. Table A2, panel B, reports associations between control variables and financial insecurity, job insecurity, and institutional distrust in the 2016 data.

6. Analyses using "hours last week" rather than "usual hours" do not substantively change the parameter estimates of our key independent variables; the question on usual hours was not available in all of the years in the combined data set.

i. We exclude 'other' from the descriptive tables because of small ns.

ii. Optimism is an index that averages six questions that ask how much respondents agree or disagree that good things will happen to them personally (alpha = 0.70). Hope s an index that averages six questions asking respondents the extent to which they see themselves as successful, resourceful, and energetic (alpha = 0.83).

Additional Information

ISSN
2377-8261
Print ISSN
2377-8253
Pages
218-261
Launched on MUSE
2019-09-19
Open Access
Yes
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.