Every spring and fall, a furious debate ignites across the country: Should we do away with daylight saving time?
Beyond anecdotes of preferences amongst various professions and household situations, the debate raises real questions for policymakers to consider. Does the time shift improve certain types of productivity? Does it save on power usage? Does it make people more or less physically active?
According to new research from Duke University, the answer to the last question is neither. When comparing the number of steps tracked on Fitbits worn by thousands of people across four states, researchers found no appreciable difference in the daily number of steps taken before and after a time change.
The research did, however, find a measurable shift in steps being taken between morning and evening. This shift, they said, likely reflects people adjusting their schedules to match the new daylight availability.
The results appear online April 23 in the journal Nature Health .
"This study is important because it shows how large-scale data collection efforts funded by the government can generate findings that can be used to make evidence-based decisions around policy," said Jessilyn Dunn, associate professor of biomedical engineering at Duke.
In the study, Dunn and PhD students from across Duke's campus pulled data from the All Of Us Research Program. Funded and administered by the National Institutes of Health, the program pulls Fitbit data volunteered by over 50,000 people across the country every year. Its goal is to provide an enormous dataset to fuel thousands of studies just like this one.
Given that most of Arizona does not change their clocks along with the rest of the contiguous United States (neither does Hawaii), the group decided to focus their attention on the "four corner" states of Arizona, Colorado, New Mexico and Utah.
"Picking these states allowed us to put together more of a natural experiment design," said Hayoung Jeong, a PhD student in Dunn's laboratory and co-first author of the paper. "Rather than randomly assigning people into the study, we were able to frame this question in a more statistically sound method."
Choosing this area of the country provided the researchers with a natural control population. While these states are each large on their own, their respective inhabitants are not wildly different from one another in aspects of climate, types of jobs, or other variables that could affect their daily step routines. And because most of Arizona does not change its clocks along with the other three, the researchers could compare differences between them.
To help with the data analysis, Dunn and Jeong turned to Srikar Katta, a PhD student advised by Alex Volfovsky, associate professor of statistical science and computer science at Duke. To refine their statistics even more, the group compressed the step data into three groups: morning, evening or full day. They then homed in on one week before and after the time changes that took place in the fall of 2021 through the spring of 2023.
While the researchers did not find any appreciable difference in the number of steps people took before and after a time change, they did see changing patterns in the times people were taking those steps. For example, if someone is working a job from 7:00 am to 7:00 pm, they likely do not have the ability to change their exercise routine. But if their work schedule is more flexible, then the time change might affect their day-to-day decisions.
"This is information that wearables give you that nothing else can, unless you are explicitly watching them," said Katta, who is also a co-first author of the paper. "Research has shown that tens or hundreds of extra steps a day can affect your health, so how daylight saving time affects those steps is important to a lot of people."
While the researchers did not have access to the Fitbit wearers' employment data, they did have demographics like age, annual income and the perceived walkability of their environment. And these were the variables in which the small changes to daily routines they found were most apparent. For example, younger people, those with higher earnings, and those living in places with readily available sidewalks or trails were more likely to change their routines.
The results seem to indicate that, were daylight saving time to be abolished, choosing daylight or standard time would not have a meaningful effect on people's activity levels. They also show that keeping the semiannual switch won't either. There does, however, appear to be an aspect of fairness at play for those with less flexible schedules.
"Overall, changing times does not seem to be the public health benefit that some think that it is, at least in terms of activity levels," Dunn said. "But we should probably be looking more closely at whether these time changes are inconveniencing some groups but not others."
This work was supported by the National Science Foundation (2339669), the National Institutes of Health (F31HL179990-0), and the Google Cloud Research Credits Program.
CITATION: "Impact of Daylight Saving Time on Physical Activity Patterns." Hayoung Jeong, Srikar Katta, Will Ke Wang, Alexander Volfovsky, Jessilyn Dunn. Nature Health, 2026. DOI: 10.1038/s44360-026-00115-z
