Having spent most of my career commissioning, editing, or writing about public policy research, I understand its inherent limitations. Still, I cling to a belief, however naïve it may be, that careful study of complex problems can produce at least some clear answers that most policymakers will accept.

In education, for example, there is a wide range of strongly held views about how to improve teacher quality, student learning, and school outcomes. On most issues — funding levels; choice and competition; academic standards; technology; school design; and how to train, evaluate, and compensate teachers — you can find plausible arguments, backed up with data, on all sides.

There are exceptions. One is the common practice of paying schoolteachers more if they possess or acquire a graduate degree. Its justification sounds plausible. If undergraduate training in academic subjects or educational practice confers value, surely additional training in graduate school would confer more value.

It’s not true, though. The relationship between graduate study and teacher effectiveness is one of the most frequently studied issues in education policy. You’ll find more than a hundred studies in peer-reviewed scholarly journals. The vast majority (more than 80%) find no connection between graduate degrees and effective teaching. And even in the few studies that do find a link, the effects are modest and typically limited to degrees in math or science, not the ones in education or non-STEM subjects that constitute the vast majority of graduate degrees earned by educators.

Nevertheless, most school systems in America provide pay bumps for all graduate degrees. Most North Carolina teachers with such degrees receive higher pay, as well — but only because they’ve been grandfathered. Nearly a decade ago, state lawmakers in Raleigh did something that, as far as I know, no other state legislature in modern times has done. Guided by the evidence, they eliminated pay bumps for new teachers with graduate degrees or existing teachers who go back to school to get them.

North Carolina still differentiates pay on other grounds. The General Assembly retained pay bumps for teachers obtaining national board certification, for example. Lawmakers also authorized schools to pay teachers for demonstrated performance and created pilot programs for advanced teaching roles — paying educators more when they lead teams of other teachers, for example. These practices have some empirical support. Paying for graduate degrees really doesn’t.

A success story for evidence-based policymaking? So far, yes.

Unfortunately, the story isn’t over. When the North Carolina House of Representatives released its budget-adjustment bill a couple of weeks ago, it contained a provision to reinstate a 10% salary boost for graduate degrees. While the initial cost is only $8 million, the education-policy group BEST NC estimates that full implementation would add about $280 million in annual expenditure. Previous House budgets have also called for restoring the pay bump. The Senate doesn’t appear to agree. Thank goodness.

For every such victory, alas, there are many defeats. Also a couple of weeks ago, the Charlotte City Council voted to spend $650 million to help renovate Bank of America Stadium, the home of the Carolina Panthers. Here’s another clear consensus in empirical research: government funding for professional sports does not confer net benefits on taxpayers. It simply forces them to subsidize billionaire owners, superfans, and other special-interest groups.

In a summary for the Federal Reserve Bank of St. Louis, economist Adam Zaretsky wrote that advocates of taxpayer subsidy routinely overestimate the benefits and minimize the opportunity costs.

“Regardless of whether the unit of analysis is a local neighborhood, a city, or an entire metropolitan area,” he explained, “the economic benefits of sports facilities are de minimus” and far lower than alternative uses of tax dollars. This is even true for hotel and restaurant taxes — the costs of which are borne partially or mostly (respectively) by locals, not visitors, and ought to be spent on true public services, not sports teams.

I keep clinging to my belief that evidence matters — clinging by my fingertips, that is.

John Hood is a John Locke Foundation board member. His latest books, Mountain Folk and Forest Folk, combine epic fantasy with early American history.