A Dangerous Education Divide During Wildfires

Amid toxic smoke, the less educated majority ventures out far more often than college degree holders

American adults with college degrees now outlive those without by more than eight years, in part because graduates appear to be more proactive about protecting their health. For example, they smoke cigarettes at far lower rates and follow cancer screening advice more often, especially after a relative is diagnosed. Even as studies documented such disparities, the health gap widened.

In this age of relentless information everywhere, does it take a college degree to understand the risks of smoking and benefits of cancer screenings? Or do those without degrees understand perfectly well but lack the flexibility, health coverage or other resources needed to act on the information? Over decades of study, the answers are still ambiguous. Researchers rarely get the chance to watch how people make decisions when faced with real health threats.

A paper published in Proceedings of the National Academy of Sciences takes advantage of two natural events to seek explanations. Using smartphone location data, Columbia’s Clara Berestycki and UCLA Anderson’s M. Keith Chen track Californians’ decisions to shelter, or not, against ordinary rainstorms and, separately, wildfire smoke. The study focuses on California’s catastrophic 2018 wildfire season, which led to an estimated 12,000 deaths from immediate and long-term effects.

When the impetus to shelter was a mere rainstorm, the researchers found little evidence that schooling levels affected who took cover. The non-degreed majority headed indoors almost as often as graduates. But when the air became heavy with toxic smoke — an actual health threat that warrants shelter — the less-educated stayed inside at a fraction of the rate college degree holders did.  

The rain findings suggest most of those without four-year degrees had the same ability to shelter as college graduates. There’s also no knowledge gap in that scenario; everyone knows you get unpleasantly wet in the rain. So when that same less-schooled population more often exposed themselves to wildfire smoke, it probably wasn’t because they couldn’t adjust their responsibilities to stay indoors, the researchers explain. More likely, the study posits, they didn’t perceive the health threat as accurately as the college graduates. 

Messaging and the Masses

Did the subject skip a cancer screening because she didn’t grasp the stakes? Or was it because they couldn’t get transportation, or time off work, or arrange caregiving or afford the bill? 

Berestycki and Chen’s findings suggest that even when protection from a threat is virtually free, schooling still predicts who acts. That puts renewed emphasis on the possibility that official messaging, whether on wildfire risks or health care advice, is not clicking with a huge portion of the population.

Both rainfall and wildfire smoke are indiscriminate threats, the researchers point out. Rich and highly educated, young and old are as likely to encounter them as anyone else. The factors in the decision to stay at home cross both groups, too. Yes, some people have a harder time rearranging their lives to do so, but the rain results suggest it’s not a lot easier with a college degree.

Similarly, it doesn’t take a college degree to recognize wildfire smoke. But the understanding of its dangers might differ by education levels, the researchers hypothesized. They note long-standing theories, often cited in health care studies like those above, that contend college education better prepares people to evaluate and act on health information.

To investigate, Berestycki and Chen collected anonymized GPS pings from 163,000 smartphones in the San Francisco Bay Area and Sacramento during the 2018 fire season. Recall that 2018 included the so-called Camp Fire, the deadliest and most expensive wildfire in California’s history. The two focus areas were exposed to horrific air pollution but were not under evacuation orders. Also noteworthy: The season ended before the pandemic, when millions of Americans permanently abandoned work commutes for home offices.

For each device, the researchers created a daily record of movement: hours at home, at work, indoors, outdoors and distance traveled, using techniques Chen helped refine in other studies using location data. “Home,” for example, is where the phone sleeps at night. “Indoors” is assessed by overlaying each ping’s coordinates on a map of building footprints. 

The researchers didn’t know education levels for the owners of each device, so they assigned them by U.S. Census blocks. Education levels in each block, which contain about 1,500 residents each, come from American Community Survey five-year averages. That means the schooling related to some device owners was inevitably labeled incorrectly. But these random errors would likely, if anything, shrink the measured gap, not inflate it. 

For the experiments, the researchers compared how each device changed its routine, first during rainstorms, then as wildfire smoke thickened. They relied on weather reports and EPA data on smoke concentrations at each device’s home.

Seeing Is Believing

Most people went about their lives during the early light haze, the study finds. That’s hardly surprising, given the EPA’s alert system at the time didn’t advise universal sheltering until the air was thick with smoke. Local Twitter discussions about the wildfires were fairly muted at this low level of smoke also, the authors note.

Medical research since suggests that wildfire smoke at levels well below dense smog harms even healthy people. As a result, in 2024, the EPA lowered the levels that triggered its early warnings. But there’s ongoing debate over whether those rather complicated adjustments to alert timing would move people soon enough to make a difference in their health.

When wildfire smoke did get alarmingly thick, people across the study population interrupted their schedules to take shelter, according to the data. In extreme smoke, time at home increased by 40 minutes on average, and time indoors by 26 minutes, compared with low air pollution days. Time at work decreased by 19 minutes, while overall mobility dropped 3% to 5%.

Education stood out as the only demographic that registered a significant differential response. In the higher educated neighborhoods, the retreat to home was about five times what it was among the less schooled ones. After controlling for education, other factors like income, race and age seemed to make little difference in sheltering decisions. 

That said, the richest and most educated were more likely to transition from “sheltering at home” to “temporarily relocating” when the smoke got dense. But money alone didn’t make people more or less likely to take shelter. 

Not an Either/Or Problem

The question of why college graduates live longer has perplexed economists for decades. One long-standing, but still debated, theory suggests that education better equips people to understand and evaluate the health information that everyone gets. With less education, it is known, people tend to make worse decisions around their own health

That’s exactly what’s going on in the responses to wildfire smoke, Chen suggests in an email exchange. Everyone could see the smoke and hear the alarms. The college graduates appeared more likely to understand that they really should act on them. He sees wildfire smoke as the newer example of this quite old allocative efficiency hypothesis. 

There’s no question that America’s inequitable system of health care makes it hard or even impossible for a lot of people to get care. But the Berestycki and Chen findings suggest that a knowledge gap — something unrelated to complicated insurance requirements or out of pocket costs — may be limiting optimal decision-making around protecting one’s health, alongside those obstacles. It’s a potentially more fixable problem.