As executives wrestle with return-to-office policies, there’s no shortage of data available to help inform their decisions.
Since the Covid-19 pandemic pushed many companies to test out a large-scale work-from-home experiment, researchers have been tracking how global working patterns have evolved. But not all data is created equal, researchers at MIT Sloan School of Management discovered, when they set out to track work-from-home trends among Americans in 2020.
The Bureau of Labor Statistics may have underestimated the rate of Americans working remotely during the first year of the pandemic by up to 25 percentage points, according to a working paper recently published in the National Bureau of Economic Research and authored by the MIT researchers, along with scholars at Stanford, Princeton, and the University of Pennsylvania. The paper points to differences in research methodologies that affect our understanding of remote work rates in the US.
It’s all in the methodology. One major reason for the government’s conservative estimate lies in the way the BLS worded its survey questions in the Current Population Survey, according to Hong-Yi TuYe, a fourth-year PhD candidate at MIT who coauthored the paper. Between May 2020 and September 2022, the government asked CPS respondents if they’d worked from home “because of the coronavirus pandemic.” In doing so, the survey excluded respondents who were already working remotely prior to Covid-19 or working remotely for a reason other than the pandemic, he explained.
“It’s that sort of explicit phrasing that’s a bit narrow that could lead to a different measure with the government’s capture, whereas we’re being as broad as possible with a definition of remote work,” TuYe told HR Brew.
In collaboration with Gallup, Ye and his colleagues at MIT conducted a survey and found more than half of the continuously employed workforce (53.5%) worked from home at least some of the time in October 2020. In contrast, the BLS reported that 21.2% of employed Americans worked remotely that month. Other measures of remote work, including the Survey of Working Arrangements and Attitudes and Real-Time Population Survey, similarly recorded higher rates of telework than the BLS measurements, the paper finds.
Quick-to-read HR news & insights
From recruiting and retention to company culture and the latest in HR tech, HR Brew delivers up-to-date industry news and tips to help HR pros stay nimble in today’s fast-changing business environment.
Another oft-cited government measure from the American Time Use Survey may have underestimated pre-pandemic levels of remote work by around 3 percentage points, as the survey excluded self-employed workers, who tend to work remotely, according to the NBER paper.
Changes on the horizon. The BLS survey was intentionally designed to solely capture Americans who worked remotely for reasons related to the pandemic, Stacey Standish, a representative with the agency’s press office, told HR Brew via email.
Nevertheless, the BLS started posing broader questions about remote work in the Current Population Survey last October, asking respondents if they’d worked from home at any time in the last week and removing the phrase “because of the coronavirus pandemic.” These changes were made to improve the relevancy of the questions, according to Standish. The new data is still being processed and reviewed, and hasn’t yet been published, she said.
For HR decision-makers seeking to better understand the current prevalence of remote work in the US, Ye recommends looking at several different surveys and digging into their methodology to best understand what exactly they’re capturing.
“Be a little bit careful about how you’re comparing them,” TuYe said, “because the official government surveys measure one thing, we’re measuring some other thing,” and other surveys may change their methodology over time.
“Being more careful about reading the methodological section of these surveys is important,” TuYe said. “And not just taking any of these numbers at face value.”