In 2004, urban planner Lawrence Frank and his colleagues published a study that became an influential piece of the research connecting transportation, community design and health. Their cross-sectional analysis of 10,878 Atlanta-area participants found that each additional hour spent in a car per day was associated with 6% higher odds of obesity, while each additional kilometer walked was associated with roughly 5% lower odds.
Two decades later, Frank and Jacob Carson revisited that work in a 2026 commentary in the American Journal of Preventive Medicine. Their argument is that cleaner vehicle technology changes an important environmental cost of driving, but it does not remove the sedentary time built into car travel.
That distinction matters. The original 6% figure did not come from a trial comparing electric vehicles with gasoline cars. Frank’s electric-versus-gas comparison is a 2026 interpretation of the older car-time finding: propulsion changes, but the driver is still sitting. In a January 8 University of California report, Frank applied the same 6% figure to car time in an electric vehicle.
What the 2004 study actually measured
The original study came out of the Strategies for Metropolitan Atlanta’s Transportation and Air Quality project, or SMARTRAQ. Its published methods say that body mass index, minutes spent in cars, kilometers walked and demographic variables were derived through a regional travel survey involving 10,878 participants, while researchers also measured features of the built environment around participants’ homes.
Frank and Carson’s 2026 retrospective adds useful detail about the transportation side of the project. Reported vehicle trips were mapped onto a calibrated regional transportation network so researchers could estimate travel time using modeled speeds rather than relying only on participants to remember exactly how many minutes each journey took.
The headline result was an association, not a controlled experiment. After adjustment for demographic factors, every additional daily hour in a car was associated with 6% higher odds of obesity. Every additional kilometer walked per day was associated with a 4.8% reduction in the odds.
That does not establish that an extra hour behind the wheel will cause a particular individual to become obese. People who spend more time driving can differ from lighter drivers in commuting distance, neighborhood design, income, occupation, food environment and many other ways that a cross-sectional study cannot completely separate.
Why the electric car does not change Frank’s argument
The mechanism Frank is emphasizing in 2026 is sedentary time rather than the engine under the hood. Electrification can reduce the air-pollution burden associated with combustion vehicles, but it does not turn an hour seated in traffic into an hour of physical activity.
That is why the new commentary treats electric vehicles as an environmental improvement without presenting them as a solution to the inactivity associated with car dependence. The health argument is about how transportation systems shape everyday movement, not about claiming that electric vehicles erase none of the other harms associated with gasoline cars.
Carson makes the same policy distinction in the University of California report. Electrification and automation may transform vehicle technology, but they do not by themselves provide the walking, cycling or transit activity that comes with transportation systems designed around more than private cars.
What the 6% number does and does not tell you
The original result is easiest to misread because it concerns odds. A 6% increase in odds is not the same thing as a six-percentage-point increase in the probability of obesity, and it does not mean that a person’s weight rises by 6% for every hour driven.
It is also a population-level statistical estimate. The model compared people with different amounts of daily car time after accounting for measured demographic and built-environment variables. It cannot predict what will happen to a particular driver who adds an hour to a commute.
The broader literature is nevertheless one reason Frank and Carson returned to the finding. Their 2026 commentary notes that systematic reviews of driving time and weight outcomes had found a consistent positive direction, with 11 of 13 relevant cross-sectional and longitudinal studies reporting a significant association as of 2020.
That does not mean every later study reproduced exactly the same 6% coefficient. Different populations and different methods produce different estimates. The more defensible conclusion is that the relationship between longer car time and higher weight-related risk has repeatedly appeared, while the precise size of that association varies.
The built environment does much of the work
Frank’s larger argument has always been about the environment surrounding the trip. Sidewalks, street connections, nearby destinations, transit access and routes that feel practical on foot or bicycle help determine whether ordinary errands involve movement or another trip by car.
The same induced-demand principle discussed for roads can operate across other transportation modes. An Urban Milwaukee summary of University of Cambridge research reported that expanding cycling and transit infrastructure can generate additional demand for those forms of transportation as well.
A related 2026 Cities study co-authored by Frank examined the interaction between neighborhood walkability and smaller pedestrian features among older adults. The research focused on features such as route connectivity, streetscape conditions and crossing safety, reinforcing the idea that broad land-use patterns and the details of a street can work together.
In the UC report, Frank extended that argument to practical investments including sidewalks, trees, shade, benches and safer crossings. His policy emphasis is that relatively small changes can make walking more workable, particularly in places where residents have fewer transportation choices.
How the finding fits with the broader literature
Car time is only one part of the much larger obesity picture. Diet, genetics, sleep, occupational activity, socioeconomic conditions and many other environmental and behavioral factors matter. Neither the 2004 study nor the 2026 commentary argues that driving time operates alone.
Other research illustrates how broad that landscape is. A 2026 Frontiers in Nutrition study developed an obesity-risk prediction model for Chinese children and adolescents using variables that included physical activity, sedentary behavior and sociodemographic characteristics.
The population burden associated with high body mass index has also continued to rise in younger age groups. A 2025 analysis using Global Burden of Disease 2021 data reported substantial increases since 1990 in deaths and disability-adjusted life years attributable to high BMI among people aged 15 to 39.
Those studies provide context rather than direct proof of Frank’s car-time estimate. The evidence for the 6% figure remains the Atlanta analysis, while the case for taking transportation design seriously rests on a wider body of research connecting sedentary behavior, travel patterns and the built environment.
Where this leaves the ordinary reader
Frank’s research is not an argument that every car journey is optional, nor is it evidence that an individual driver can calculate a personal obesity outcome from the length of a commute. Many communities were built in ways that leave residents with little practical alternative to driving.
It is also not an argument against electric vehicles. The narrower point is that electrification tackles a different problem. A cleaner drivetrain can reduce pollution from the vehicle without changing the amount of time its occupant spends seated.
The strongest version of the finding is therefore more modest than a cause-and-effect slogan. In the 2004 Atlanta data, more daily car time was associated with higher odds of obesity. Later research has generally supported the direction of that relationship, and in 2026 Frank argues that swapping gasoline for electricity does not change the sedentary behavior underneath it.
Twenty-two years after the original paper appeared, the object carrying the commuter may be changing rapidly. The basic posture inside it has not.
Correction, 15 September 2026: The headline and excerpt were revised to distinguish an association in a cross-sectional Atlanta study from a causal effect and to clarify that the study did not compare electric with gasoline vehicles.