[P]ublic schools appeared to be attaining higher levels of mathematics performance than demographically comparable private and charter schools—and math is thought to be a better indicator of what is taught by schools than, say, reading, which is often more influenced directly and indirectly by experiences in the home. These patterns flew in the face of both the common wisdom and the research consensus on the effectiveness of public and private schools. Immediately, we checked to see what had happened in the analysis, whether “public” and “private” had been “reverse-coded” or some other such error was involved. But after further investigation and more targeted analyses, the results held up. And they held up (or were “robust” in the technical jargon) even when we used different models and variables in the analyses. We eventually posted a technical paper on a respected website and published a short article, which received some attention. And then, like any good researchers, we applied for funding to study this issue in more depth using the most recent, comprehensive databases. The results across datasets are consistent and robust—indicating that these patterns are substantial and stable, regardless of changes in the details of the analyses.

These results indicate that, despite reformers’ adulation of the autonomy enjoyed by private and charter schools, this factor may in fact be the reason these schools are underperforming. That is, contrary to the dominant thinking on this issue, the data show that the more regulated public school sector embraces more innovative and effective professional practices, while independent schools often use their greater autonomy to avoid such reforms, leading to curricular stagnation.