The Houston Chronicle Cites Iffy Data To Make the Case for Mask Mandates in Schools

When you are already convinced a policy makes sense, any evidence will do.


"Are mask mandates helping Houston school districts slow the spread of COVID-19?" asks the headline over a Houston Chronicle story published today. The data presented in the article do not answer that question, although it is clear that the authors—Hannah Dellinger, Julian Gill, and Alejandro Serrano—really, really want the answer to be yes. In that respect, they resemble most mainstream news reporters, who take it as a given that "universal masking" in K–12 schools plays an important role in reducing virus transmission, then cast about for evidence to support that assumption.

"Houston-area school districts that require students and staff to wear face masks generally have experienced fewer COVID-19 cases," Dellinger, Gill, and Serrano report. "The numbers, which offer a snapshot in time of active cases, appear to reinforce what doctors have been saying for months: requiring or strongly encouraging masks can help reduce the spread of the virus, more so when paired with other strategies, such as vaccines and social distancing."

That is a pretty slippery summary. School districts with mask mandates "generally have experienced fewer COVID-19 cases," except when they haven't. Dellinger and her colleagues think the numbers "appear to reinforce" something, but it's not clear what that thing is. Is "requiring" masks necessary, or is "strongly encouraging" them enough? Were these "doctors" talking specifically about masking in K–12 schools, or were they referring to the general question of whether face masks "can help reduce the spread of the virus"? And if mask mandates are usually "paired with other strategies," how can the effect of this particular precaution be isolated?

Dellinger et al. focus on "active" COVID-19 infections reported by school districts last Thursday. What does "active" mean? That varies from one school district to another. The Chronicle notes that "cases in some districts are considered active for 10 days after a positive test, the length of the recommended isolation period, regardless of an individual's actual condition."

Did these cases result from in-school transmission? Not necessarily, which undermines any attempt to infer a causal relationship between a district's mask policy and its infection rate. Dellinger and her colleagues note that "not all communities have the same rate of transmission outside of classrooms." If school districts without mask mandates tend to be located in communities with higher general infection rates, that could help account for differences that might otherwise be attributed to a lack of face coverings in school.

Similarly, vaccination rates vary widely across the Houston area. In Fort Bend County, for example, the Texas Department of State Health Services reports that 74.5 percent of residents 12 or older are fully vaccinated. In Waller County, by contrast, the rate is 42.8 percent. The rate for Harris County, where Houston is located, is 63.9 percent.

These county-level data obscure differences within counties. Harris County, for example, has 4.7 million residents and covers 1,777 square miles, encompassing communities with different demographics and population densities, ranging from Houston to small towns and unincorporated areas. To complicate matters further, some school districts include parts of different counties with different vaccination rates.

Even based on county-level data, it is clear that failing to take vaccination rates into account could skew any attempt to measure the impact of mask mandates, especially if school districts that don't require masks tend to have lower vaccination rates, which seems plausible given the weird politics of COVID-19 precautions. "Some districts without mask mandates but located in counties where more individuals are fully inoculated against COVID-19 also have logged fewer COVID-19 cases per capita," Dellinger et al. write. "Districts without mandates in counties where fewer individuals have been vaccinated experienced more cases per capita."

Waller Independent School District (ISD), which does not require masks, had the highest incidence of active cases in the Chronicle's analysis: 10.89 per 1,000 students. It also happens to be located in the county with the lowest vaccination rate. Humble ISD and Sheldon ISD, which likewise do not require masks, had the lowest rate of active cases: 2.63 per 1,000 students. Like Houston ISD, they are both located in Harris County, but their active-case rate was more than 40 percent lower than Houston ISD's (4.44 per 1,000), even though Houston ISD requires masks. Another Harris County school district with a mask mandate, Spring ISD, had 5.76 active cases per 1,000 students, more than twice the rate in mask-optional Humble ISD and Sheldon ISD.

All told, just six of the school districts in the Chronicle's analysis were defying Gov. Greg Abbott's ban on mask mandates in public schools, and one of them, Magnolia ISD, required masks only for elementary school students. Their active-case rates ranged from 2.73 to 5.76 per 1,000 students. The remaining 20 districts did not have mask mandates, and their active-case rates covered a much broader range, from 2.63 per 1,000 at one extreme to 10.89 per 1,000 at the other.

Assuming that mask policies affect these numbers, there are clearly other variables at play, since many school districts managed to keep infection rates low without requiring masks. "It is difficult to determine how mask mandates alone are affecting transmission of the virus in the region," Dellinger et al. concede, citing Katelyn Jetelina, an epidemiologist at UTHealth School of Public Health in Dallas. "Schools with face covering requirements may be implementing other mitigation strategies, which clouds the data, she said. And other factors contribute to infection rates, such as community vaccination rates."

Jetelina notes that "when you look at these policy questions and try to analyze them, you have to take into account all of these extraneous factors." She adds that "it's just going to take a while for the data to play out to find out if there's a true causal link between mask mandates and case rates."

The latest guidance from the Centers for Disease Control and Prevention (CDC) recommends that schools require all students to wear masks, regardless of their age or vaccination status. But by and large, the studies that the CDC cited to support that recommendation did not even compare schools with mandates to schools without them, let alone "take into account all of these extraneous factors." When it issued its latest guidance for schools, the CDC's best attempt at a more rigorous analysis was a large study of Georgia schools published in May, which found no statistically significant evidence that requiring students to wear masks reduced infection rates, even before vaccines were widely available.

In a preprint study posted the same month, Brown University economist Emily Oster and four other researchers analyzed COVID-19 data from Florida, New York, and Massachusetts for the 2020–21 school year. "We do not find any correlations with mask mandates," they reported. But they noted that "all rates [were] lower in the spring, after teacher vaccination [was] underway."

In a report published today, the CDC looks at "school-associated COVID-19 outbreaks"—defined as two or more confirmed cases among students or staff members within a 14-day period—in Arizona. Of the 999 Pima County and Maricopa County schools included in the analysis, 21 percent imposed mask requirements at the beginning of the 2020–21 school year, 31 percent had a "late mask requirement" (imposed a median of 15 days after the school year began), and 48 percent had no mask requirement. A total of 191 outbreaks were reported between July 15 and August 31, 2021; 113 (59.2 percent) happened in schools with no mask mandates, 62 (32.5 percent) happened in schools with late mask mandates, and 16 (8.4 percent) happened in schools with early mask mandates.

"After adjusting for potential described confounders," the CDC says, "the odds of a school-associated COVID-19 outbreak in schools without a mask requirement were 3.5 times higher than those in schools with an early mask requirement." The chance of an outbreak was only slightly higher in schools without mask mandates than in schools that imposed mask mandates after the school year began, although the CDC notes that late mandates might have been imposed in response to outbreaks.

The CDC's adjusted analysis considered "school county, enrollment size, grade levels present, Title I status [a measure of neighborhood income levels], and [the] 7-day COVID-19 case rate in the school's zip code during the week school commenced." It did not take into account vaccination rates, because "vaccination coverage for staff members and students was not available at the school level." Nor did the CDC consider other precautions, such as ventilation improvements, that schools with mask mandates may have been more likely to adopt, although it notes that "CDC K–12 school guidance recommends multiple prevention strategies."

Rashid Mosavin, dean of Texas Southern University's College of Pharmacy and Health Sciences, warns the Chronicle that "it is not that easy or straightforward to interpret some of these numbers because there are just so, so many moving pieces." He adds that "what we do know for a fact" is that "the only way out of this pandemic is vaccination."

Given the shaky empirical basis for the belief that K–12 mask mandates have an important impact on infection rates, school officials who support that policy seem unjustifiably confident. At a September 9 meeting, the Chronicle notes, Houston ISD Superintendent Millard House II proudly informed the district's Board of Trustees that just 0.6 percent of students had tested positive for COVID-19 since the beginning of the school year. "Without, of course, the mask mandate," he said, "I don't think this would have happened." The Chronicle's analysis, which found that Houston ISD had a higher active-case rate than 10 school districts without "universal masking," does not seem to support that conclusion.

Terri Burchfield is deputy superintendent for support services at Texas City ISD, which requires masks but was not included in the Chronicle's analysis because it did not report data on active cases. "I really think that conversations just have to be about the data, you know, and take a little bit of the emotion out of it," Burchfield told the Chronicle. "We are doing it for the best interest of the kids."

I have no doubt that Burchfield is completely sincere. But whether she and her colleagues are in fact "doing it for the best interest of the kids" depends on whether the public health payoff from forcing them to wear masks all day justifies the substantial burdens that policy imposes, taking into account the rarity of life-threatening COVID-19 symptoms in children and teenagers. If her position really is "about the data," she should be able to cite more evidence that the benefits of mask requirements outweigh the costs.