(James Glanz, Benedict Carey and Matthew Conlen for The New York Times)

As schools grapple with how to reopen, new estimates show that large parts of the country would likely see infected students if classrooms opened now.

Millions of families face an excruciating choice this fall: Should their children attend if local schools reopen their classrooms, and risk being exposed to the coronavirus? Or should they stay home and lose out on in-person instruction?

No single factor can settle such a fraught decision. But new estimates provide a rough gauge of the risk that students and educators could encounter at school in each county in the United States.

The estimates, from researchers at the University of Texas at Austin, range from sobering to surprisingly reassuring, depending on the area and the size of the school.

Based on current infection rates, more than 80 percent of Americans live in a county where at least one infected person would be expected to show up to a school of 500 students and staff in the first week, if school started today.

In the highest-risk areas — including Miami, Fort Lauderdale, Nashville and Las Vegas — at least five students or staff would be expected to show up infected with the virus at a school of 500 people.

The high numbers reflect the rapid spread of the virus in those areas, where more than 1 in 70 people are estimated to be currently infected.

At the same time, smaller, isolated groups of students face a much lower risk. Some schools are considering narrowing classes down to small “pods,” with students who mainly come in contact with their teacher and each other. While the chance of having an infected person at the school would stay the same, the risk of exposure within those pods would be much lower.

If they remain isolated from the rest of the school — a tall order — 10-person pods in every part of the country would be unlikely to include an infected person in that first week. 

Education experts and disease researchers said information that reflects local conditions could be critical in shaping decisions by parents, teachers, administrators and political leaders.

“It’s meant to guide schools so they can anticipate when it might be safe, or easier, to open and bring kids in,” said Lauren Ancel Meyers, an epidemiologist at the University of Texas at Austin who led the research team.

The projections are rough guidelines based on the estimated prevalence of the virus in each county, which is drawn from a New York Times database of cases, and estimates that five people may be infected for each known case. Those estimates reflect current levels of infection around the country and are likely to change, improving or worsening in individual communities over the next weeks and months.

The estimates assume that children are as likely to carry and transmit the virus as adults —  “a large assumption, given the unknowns about children,” said Spencer Fox, a member of the research team.

“This is meant to be a rough guide, a first step,” Dr. Fox said.

Some preliminary studies have suggested that children are infected less often, or that young ones do not transmit the disease as readily, which could reduce the risk, said Carl T. Bergstrom, a professor of biology at the University of Washington. But those questions remain unresolved, he said.

Still, the information “really helps put things into context for parents,” Dr. Bergstrom said. “Anything that could help you do that both helps you make better decisions and offers a level of comfort and assurance.”

Many districts will start the school year remotely. Those that do open buildings will hedge the risks by taking various measures, such as requiring masks and social distancing, holding classes outside when possible or bringing students to school on alternating schedules.

Plans announced by some of the nation’s largest school systems already show the range of choices in play. Districts in San Diego and Los Angeles, citing the risk of crowded classrooms, said they would operate online in the fall, as will the vast majority of schools in California under guidelines issued by the state. New York City, though, is planning a partial reopening, allowing classroom attendance one to three times a week.

But decisions on remote learning come with their own concerns, said Greg J. Duncan, an education professor at the University of California, Irvine. Studies have shown that younger children and those in lower-income districts do not learn as well online as they do in person. For lower-income children, that gap in learning can persist, he said.

Wealthy families, which have more resources and workarounds, will be far more risk-averse than others, Dr. Duncan said.

“One infection is too many” will likely be the refrain of wealthier families, he said. “Any slight chance that their child is going to be infected is probably going to get them to jump to a decision more quickly than lower-income families.”

Although the risk varies by school size, in the hardest-hit areas of the country, even small schools face significant risks.

In eight states, most people live in counties where even a school of only 100 people would probably see an infected person in the first week if school started today, the estimates say: Louisiana, Alabama, Mississippi, Florida, Nevada, Tennessee, Arizona and Georgia.

The list is even longer for schools of 500 people: The vast majority of people in 19 states, including California, Texas and Illinois, live in counties where at least one infected person would likely show up to school in the first week if in-person classes were held. Many of those areas have elected to hold classes online for now.

Many parents are consumed with the question of returning to school, and there is hunger for solid guidance, said Annette Campbell Anderson, deputy director of the Johns Hopkins Center for Safe and Healthy Schools.

“They want to see the data to make them feel that they have a model that they can trust,” Dr. Anderson said. “And we need it. We need this kind of data.” 

Notes: Estimates approximate the proportion of the population that is infectious based on the number who were infected during the preceding seven days, from data ending July 28. The calculations assume that students and teachers come in to school at least once a week and won’t come in if they are symptomatic. The estimates are likely to change as the epidemic gets better or worse in certain areas. The estimates assume that five people are infected for each known case. The estimated range of infected people shown on the map is based on scenarios with between three and 10 people infected for each known case.