Perception is a complicated foundation upon which to build public policy. In a recent Gallup poll, 71% of respondents declared that the American healthcare system was either in “crisis” or had “major problems.” In the same poll, when asked to rate the quality of one’s own healthcare, 76% of Americans declared those experiences to be either “excellent” or “good.” Despite decades of political maelstrom over healthcare reform, these two measures are unchanged since the late 1990s.
Humans are inconsistent in their perceptions of even the simplest of political phenomena, and that unreliability is exacerbated when presented with the extraordinary complexities of public systems. That’s why, when making judgments about how to improve, tweak, or outright reform the healthcare system, we rely on factors that have nothing to do with perception, but involve results: quality of care, health outcomes, outcomes per cost unit, and other measures that aspire to impose modest guardrails on a difficult-to-standardize endeavor. While healthcare battles rage on, the field has moved past the idea that the practice of medical care is too subjective to be measured. Despite arguments that healthcare is too much of an “art” to invite scientific measurement, the public and policymakers insist on measuring outcomes to justify the multi-trillion-dollar annual expenditure that characterizes our and ethically correct public commitment to public health.
For over a century, our country has made a similar moral commitment to financing and mandating public education. The debate over how to measure that commitment remains slippery. That’s why it’s baffling that prominent commentators continue to question the use of objective data in education. Here’s Jack Schneider, an education historian, writing in The Atlantic about how the public perceives education:
Asked to assign letter grades to their children’s schools, the vast majority of parents—generally around 70 percent—issue As and Bs … When asked to rate the nation’s schools, however, respondents are far less sanguine. Reflecting on public schools in general, a similar share of respondents—roughly 70 percent—confer a C or D … The biggest factor shaping the perception divide, however, may be data … What evidence exists to suggest that data is the problem? One piece comes from the Kappan poll. In 2002, the year NCLB was signed into law, 60 percent of respondents gave the nation’s public schools a C or a D grade. Thirteen years later, that figure was up to 69 percent. Yet school performance did not change markedly during that time. In fact, scores on the National Assessment of Educational Progress—a nationally representative assessment of student learning—generally held constant or rose slightly during that period.
Schneider’s reasoning is circular; he argues that the public has wrongheaded perceptions about public education because of a series of tests, and that we know their perceptions to be wrong because of … a different series of tests. What Schneider eventually gets around to arguing is that we should trust the perceptions of the people closest to schools, rather than objective external evaluations. He also argues that “out-of-school factors like family and neighborhood account for roughly 60 percent of the variance in student test scores; teachers, by contrast—the largest in-school influence—account for only about 10 percent.”
To accept Schneider’s argument is to acquiesce to two undesirable ideas. The first is that schools don’t matter that much when accounting for social outcomes, which is the conclusion that his last point suggests. Here again, healthcare offers a useful comparison. Healthcare research demonstrates that the “social determinants” of health – in other words, everything outside of the healthcare system – can account for up to 90% of the variance on critical healthcare outcomes, while medical care itself accounts for only 10-30% of outcomes. In other words, teachers alone account for the same magnitude of educational outcomes as the entire healthcare system appears to contribute to health outcomes. To discount test scores as a useful metric would be like dismissing health outcomes as a useful measure of physician performance.
The second wrongheaded conclusion Schneider derives from the “perceptions obviate test scores” argument is that the current outcomes of the education system are mostly tolerable. Schneider, in his final paragraphs, rejects an interventionist approach to education policy, which sounds great, if you happen to be a family who enjoys high quality public schooling. I went to a great suburban public school system in New Jersey for the entirety of my own primary and secondary education. That experience colored my perception of schools for decades. Even if most other Americans shared that positive experience, that “perception” does nothing for the African-American child on Chicago’s south side whose school is closing after decades of neglect. Similarly, my suburban-white-kid perspective has zero effect on improving a school that serves mostly Latino students from immigrant families in East Los Angeles. Public education policy ought to exist to protect the most vulnerable people in the system. The perceptions of the majority shouldn’t prevent us from hearing from the people who are getting screwed by reality.
Schneider is right about some things. He’s right that how we measure schools ought to be more nuanced, and that our current testing regimes are far from perfect. But the last thing education policy needs is to revert to the idea that our feelings should supersede objectivity. That’s the sort of world in which fake news gets traction, and sincerely articulated untruths can never be debunked. It’s a world in which families and children without access to power continue to suffer in failing public institutions, because their wealthier peers with degrees from selective colleges are “mostly satisfied” with their own public schools.