Mainers dramatically reduced their movement around the state over the past month as the growing severity of the coronavirus pandemic influenced social behavior and government policies.
State traffic data and publicly available smartphone metadata show mobility in Maine decreased steadily over the past five weeks as company work-from-home policies transitioned into school and business closures, and finally a statewide stay-at-home order from Gov. Janet Mills.
But changes in mobility have wide regional variations, according to traffic data from the Maine Department of Transportation and smartphone tracking data from New York-based “human mobility data company” Unacast. Movement in urban and more densely populated parts of southern Maine has plummeted, while declines in some rural areas has been less pronounced.
Local livelihoods, average driving distances, the timing of business closures and community lockdowns, and the geographic spread of coronavirus all contributed to those variations, experts said.
There are also questions about the accuracy of smartphone data in rural parts of the state, they said.
“That data has to be taken with a huge grain of salt,”said Kathryn Ballingall, a researcher focused on transportation data for the University of Maine’s Margaret Chase Policy Center. “There may be methodological issues with this that are unintentionally biased against rural areas.”
Traffic at major intersections in and around Portland on Fridays, typically among the week’s busiest travel days, declined by more than 60 percent between the first weeks of March and April, according to regular monitoring reports from the Maine Department of Transportation.
Average mobility, defined as miles traveled, declined by around 55 percent statewide, according to Unacast.
Nonessential travel – meaning not for groceries or other necessities – fell by as much as 75 percent, according to the firm, which maintains a COVID-19 online scorecard based on anonymous smartphone and other publicly available data.
Unacast assigns a letter grade to states and counties based on their reduced mobility. It gave Maine an overall B-minus, meaning there has been a decrease in mobility of between 55 percent and 70 percent.
Cumberland County, home of Maine’s largest city and about one-fifth of the state’s population, earned an A-minus grade from the company, thanks to a 70 percent decrease in nonessential travel. York, Kennebec and Penobscot counties all earned B grades and limited nonessential travel by the same degree, according to Unacast.
But Oxford County, in Western Maine, earned a D grade overall and an F, or failing grade, in nonessential travel, registering less than a 55 percent decrease. Aroostook, Lincoln, and Hancock counties also earned failing grades.
A company spokeswoman did not directly answer a question about whether it had a large enough data set to accurately quantify mobility and movement in rural areas.
“Unacast has not specifically examined peculiarities in different types of regions and is using the same methodology to measure change in distance traveled across the scoreboard,” Tanya Merisier said.
Declines in traffic volume also varied regionally. Friday traffic at I-295 exits in Portland, Freeport, Brunswick and South Portland dropped by more than 60 percent over the course of March.
Those declines differ greatly from State Route 117 in Turner, where traffic fell by 33 percent over the month, and Route 2 in Wilton, where the decline was under 40 percent.
Just because some parts of the state have limited movement less than others doesn’t mean they are ignoring social distancing guidelines, said Maine State Economist Amanda Rector.
“If people are perceiving more cases close to them, they may be more likely to reduce their travel and stay at home,” Rector said.
The state’s first cases of COVID-19, the respiratory illness caused by coronavirus, were recorded in southern Maine. Two-thirds of the 537 confirmed cases as of Wednesday, and almost all deaths attributable to the virus, occurred in Cumberland and York counties.
Rural counties, on the other hand, have few confirmed cases, although some hospitals have not reported cases to the state, and not enough testing is being done to capture the true scope of the problem. As of Wednesday, Oxford County had only 12 confirmed cases, while there were three in Hancock County and two in Aroostook County.
Major population centers in southern Maine, including Portland, implemented curfews and limited business hours in mid-March, prior to Mills’ closure of nonessential businesses two weeks ago and stay-at-home order March 31, Rector added.
“This may have caused earlier or larger decreases in mobility compared to more rural parts of the state,” she said.
Overall, mobility and nonessential trips dropped off swiftly in the week since Mills’ stay-at-home order was issued, and Maine’s overall mobility-reduction score has since improved by two full letter grades, according to Unacast. Traffic volumes at intersections in Vassalboro, Trenton, Mattawamkeag, Winthrop and other towns dropped quickly after nonessential businesses closed.
But there are other important distinctions that can help explain continued mobility. People in rural Maine generally have longer commute times and travel farther to essential shops and services and may not have as many delivery options, Rector said.
It may also be the case that jobs easily converted to remote work are clustered in southern and coastal Maine, whereas rural parts of the state have a higher proportion of jobs that require workers to be on site, or they lack the broadband infrastructure to support remote work.
It also may be that Mainers living in more densely populated parts of the state had more opportunities to reduce nonessential trips, said Yuseung Kim, a professor at the Muskie School of Public Service at University of Southern Maine.
Limiting trips to shops, restaurants, coffee shops and other activities would show as a large reduction in nonessential trips, but people in rural areas had fewer nonessential places to visit to begin with, he said.
“There is a much smaller number of the nonessential destinations in the rural counties, so there’s little room to reduce the nonessential trips there,” Kim said.
Both Kim and Ballingall, from the Margaret Chase center, have concerns that Unacast’s data might have too small a sample size or other flaws that skew the data. It may be that its definition of nonessential makes sense in big cities, but is less effective in outlining rural travel patterns.
There also may be issues of low smartphone ownership, spotty reception and other issues. Unacast lacks data about nonessential travel from Franklin, Washington and Piscataquis counties.
“Is it correct to say that Aroostook County is doing a really bad job at social distancing while Cumberland County is (doing a good job)?” Ballingall said. “Perhaps, this data set is not accurately pinpointing what is considered essential or nonessential.”
Send questions/comments to the editors.
Comments are no longer available on this story