Money Never Sleeps – Who Sleeps and Earns the Most
When it comes to success, we all know the stereotype: the successful don’t sleep. Whether it’s a tech CEO leading a risky new startup or a famous author working on her latest best-seller, everyone believes that earning more means working more, and working more inevitably means sleeping less. It’s a trade-off most of us expect to make eventually, if we’re not making it already.
Yet when it comes to scientific data to back up this idea, there’s often not much to be found. At Tuck, we wanted to know: how does the sleep/work trade-off really work? Is it really true—statistically true—that sleeping less means earning more? Might there be exceptions to the rule? And how do specific careers compare with one another in this regard?
Without clear answers, we decided to take scientific matters into our own hands. To get to the bottom of these questions, we went straight to one of the most authoritative sources available: the Bureau of Labor Statistics’ 2016 American Time Use Survey. Published annually since 2004, the American Time Use Survey gives analysts, journalists, and social scientists a comprehensive look at how Americans spend their days—including those huge time-takers, work and sleep.
The BLS dataset for 2016 included data from about 10,000 respondents. We used this data to run a regression analysis, which gave us the relationship between hours worked and hours slept we were looking for (read about our complete methodology below).
The results were fascinating. On the one hand, we confirmed some hunches (such as that those in legal professions would, on average, be the highest paid while working the most and sleeping the least). On the other, some results were less predictable. Scientists and architects saw quite high weekly pay, while getting close to average sleep, for instance. And teachers, who you might expect to be both low earners and low sleepers, fell almost exactly at the average in both categories. Though the rule was generally true—the more you work, the more you earn, and yes, the less you sleep—there were a few big exceptions to be found. What you’ll learn about coders will definitely surprise you.
These kinds of comparisons between professions were the real fruit of our analytical labor. Finally, we compiled them into gorgeous charts and assembled them into a handy infographic, to make reading and understanding these findings as pleasant and informative impossible. Check it out, find your job on the charts, and see how your own work and sleep numbers compare to those of other careers.
The data used for this report was sourced from the Bureau of Labor Statistics’ (BLS) American Time Use Survey for 2016, a survey with roughly 10,000 respondents. We hope to gain some insight about American sleep patterns from the survey. To do this, data has been compiled from three sources:
1) The ATUS Respondent data file, which contains employment and wage information about respondents
2) The ATUS Roster data file, which contains respondents’ basic information such as age and gender
3) The ATUS Activity Summary data file, which contains total time slept per 24 hour period of the ‘diary day’ during which activities are recorded
Data for the linear model is reformatted and modified only to include employed respondents with basic wage and hours worked information (e.g. respondents who did not provide that information were not included) reducing the dataset to 3165 respondents, and moderately limiting the generalizability of any results to the population at large.
Data for the occupation averages is also reformatted to only include respondents with wage, hours worked, and occupation code information, and surveyed on weekdays, limiting the set to just over 1500 respondents. Differences in the the ‘employed’ vs ‘not employed’ sets may be attributable to subjective states of employment (full time / part time, contractor etc) that are not accounted for in this analysis.
With additional resources, the BLS data could be re-weighted to accommodate for generalizability but for now the report should be treated as an exploratory analysis.
Independent variables were added one by one (a hierarchical regression) to the model according to how well correlated they were to the dependent variable (minutes slept) and a qualitative assessment of relevance.
The order of inclusion is as listed in the table below.
At each step, variables were assessed for significance; no variables were dropped from the analysis as all showed as significant. VIF gave no indication of issues of multicollinearity at any stage in the regression.
WEEKLY HOURS WORKED
The Raw Numbers:
WEEKLY WAGE AVERAGE
WEEKLY WAGE STD
WEEKLY HOURS AVERAGE
WEEKLY HOURS STD
SLEEP AVERAGE (minutes)
SLEEP STD (minutes)
Personal care and service occupations
Food preparation and serving related occupations
Healthcare support occupations
Building and grounds cleaning and maintenance occupations
Farming, fishing, and forestry occupations
Office and administrative support occupations
Transportation and material moving occupations
Sales and related occupations
Community and social service occupations
Installation, maintenance, and repair occupations
Education, training, and library occupations
Construction and extraction occupations
Arts, design, entertainment, sports, and media occupations
Protective service occupations
Healthcare practitioner and technical occupations
Business and financial operations occupations
Life, Physical, and social science occupations
Architecture and engineering occupations
Computer and mathematical science occupations
If you appreciate what you’ve learned here about sleep and labor, please feel free to share our graphic as well as the data as you wish. We simply ask that you link back to this page to credit Tuck as the author of this content.
To Share this Image On Your Site Via an Embed Code Use this