What this covers

Emily Oster and Russ Roberts discuss school reopening during the pandemic, drawing on preliminary data Oster has collected from K-12 schools across the United States. The conversation centers on transmission risk in schools, the costs of closure, and the role of data in policy-making during deep uncertainty. Oster presents evidence that schools are not significant super-spreader sites and that school case rates track community spread rather than generating outbreaks. Roberts brings economic reasoning and pushes back on zero-risk thinking, while both speakers grapple with why school closures persisted despite accumulating evidence and what that reveals about institutional decision-making under pandemic conditions.

The exchange covers several distinct layers. On the empirical side, Oster walks through data showing case rates around 1.5 per 1,000 students per two weeks, with elementary schools appearing particularly low-risk and high schools intermediate—a gradient that puzzles her given that college reopenings clearly drove surges. International evidence from Germany, Spain, and the UK supports the no-transmission hypothesis, though Israel remains an outlier complicated by confounding variables. The conversation then shifts to costs: learning loss concentrated among poorer children, mental-health and food-security harms, and the inequity of closing schools while keeping bars open. Both speakers critique the politicization that has made data discussion partisan—Oster notes that arguing for reopening gets read as a Donald Trump signal—and Roberts articulates a philosophical position that zero-risk mitigation itself carries ruinous trade-offs. The discussion concludes with parenting, where Roberts draws on L. A. Paul and Agnes Callard to argue that parenthood is a transformative experience that cannot be fully grasped before living it, and both speakers explore the tension between helping children become themselves and pushing them toward growth.

Sharpest takeaway

Oster argues that K-12 schools—especially elementary schools—are not significant spreaders of COVID and that school COVID rates track community spread rather than originating outbreaks, so closing schools imposes large educational and social costs on children for little public-health benefit.

  • Preliminary multi-state school data show low case rates (~1.5/1000 students per two weeks) that scale with community prevalence rather than indicating in-school outbreaks
  • International evidence (Germany, Spain, UK) is consistent with schools not driving community spread, with Israel an unexplained outlier
  • School closures fall hardest on poorer children and impose mental-health, food-insecurity, and learning costs that are being neglected relative to keeping bars/restaurants open

The argument · threads4 threads · 22 claims
0.65

Pandemic policy ignored clear, stable age-gradient data to justify broad restrictions despite falling fatality rates.

1 pointcentrality 5/5
  • The age gradient of COVID risk has been clear since the earliest cruise-ship data and the age distribution of death has remained remarkably stable, yet—outside of nursing home caution—society has failed to use this stable knowledge to relax restrictions on low-risk groups, even as overall fatality rates have fallen due to better treatment across all ages.

    “the age gradient thing is obviously true. It was true--I think we knew that from the first cruise ship. It hasn't changed much, and the age distribution of death has not changed, I think a whit. It's incredibly stable.”

0.65

Free societies must weigh risk mitigation's enormous costs against benefits and prefer cultural responsibility to state coercion.

1 pointcentrality 5/5
  • A free society should not seek to mitigate risk to zero because that itself carries enormous costs—destroying businesses, employment, and children's education—and should not have government dictating private gatherings; while a neighbor's gathering imposes an externality, the cultural response of mutual care and personal caution is preferable to a state aiming for zero deaths or zero transmission.

    “We don't want to mitigate risk to zero because that also has very large costs... I want to live in something like a free society. I don't want the government deciding all the risks and having the goal of having zero deaths”

0.61

Risk perception is shaped by emotion and cognitive fallacy, not just evidence, even among data-literate people.

2 pointscentrality 4/5
  • People exhibit a fallacy of inference akin to a man counting $1 bills who stops at 47 saying 'if it's right this far it's probably right the rest of the way'—they reason 'I haven't gotten COVID yet so I'm safe' and consequently relax into riskier behavior, a manifestation of pandemic fatigue.

    “A lot of people, I feel like have said like, 'Well, I haven't gotten it yet so I think I'm safe... And you see them change their behavior.”

  • Even data-literate people maintain anxiety-driven behaviors they know are not supported by evidence—like moving from 10 feet to 12 feet from an outdoor, masked acquaintance—because the behavior is comforting rather than rational, illustrating that risk responses are partly emotional.

    “what I'm doing isn't rational, but it makes me feel good... if somebody sits within 10 feet of me, I move it to 12 feet... it makes me feel good. Even though I know it's not rational, it's just a little comforting.”

0.44

Good parenting and medical practice require understanding trade-offs and respecting both child autonomy and necessary challenge.

3 pointscentrality 2/5
  • There is no data of the conventional sort about what it is like to become a parent—only information from literature, film, and others—because parenthood is uncharted, transformative territory that cannot be fully understood until experienced, a point connected to L.A. Paul's work on transformative experiences and Agnes Callard's work on aspiration.

    “there's really no data of the kind that we normally call data about what it's like to be a parent... until you've had a baby or become a parent, it's uncharted territory.”

  • Modern parenting culture has shifted from the pre-1970s norm that parents should make children turn out like themselves (religiously, culturally, behaviorally) to a norm treating children as blank slates to explore freely; but neither extreme fits all children—good parenting requires both honoring that a child is not you and helping them push through hard things toward rewards they'll later value, raising adults rather than just doing whatever.

    “there are, there's helping your kid push through things that are hard because you know that they will like what comes out on the other end... But then there's this piece where it's like, well, actually you don't have to push through. Some things, they're just not for you.”

  • Doctors are generally not good at trade-offs—a skill economists are trained for—and tend toward over-diagnosis without weighing its downsides, so patients who understand opportunity costs can seek out physicians who share that trade-off mindset and have better-aligned conversations with them.

    “many doctors--they are not good at trade-offs. It's just not their thing. And you and I are trained to focus on trade-offs.”