We messed up. Last month I celebrated and widely announced that the Washington Post ranked five of our schools in the top 10. Those rankings were incorrect because IDEA Public Schools provided Jay Mathews and the Post with faulty data and information.

With the corrected data, our school rankings dropped: it appears as though our five schools in the top 10 were in the top 25. I want to personally and publicly apologize to Jay and the Washington Post; they trusted the veracity of the information that we provided—and we let them down.

Ultimately, it was Jay’s transparency that helped bring this mistake to light: he was very clear about what criteria was used to calculate the rankings. Carol Burris, executive director of the Network for Public Education Foundation, did her own analysis of IDEA’s numbers and told Jay she did not think they were accurate given the number of AP tests taken by each IDEA student. Jay asked us to reexamine our data and we fully cooperated to help get to the bottom of this issue.

I’m sorry for taking attention away from the schools that legitimately made the top 10—there were some very deserving students, teachers, and principals who were not recognized as a result of our error. I also want to apologize to our many supporters, funders, and cheerleaders who posted, tweeted, highlighted, and spread the word about our rankings.

We made an honest mistake; we did not purposely lie in an attempt to inflate our rankings. How did this happen? In short, we were asked to provide the number of Advanced Placement tests our students took in 2016. Instead, we ran the 2016 report that listed the number of tests our students had taken over multiple years. The result is that it looked like our students had taken a lot more exams in one year than they actually had. We provided back-up and documentation for these numbers, but that just added insult to injury—the flurry of documents and data we provided simply made our errors all the more credible looking. We have determined that the data we sent Jay for last year’s rankings was accurate.

There were multiple safeguards IDEA had in place that were supposed to have caught this mistake, but they all failed—which is why the fault rests with me as the founder and CEO: I ought to have taken much more care to test and ensure the quality of these systems, and I didn’t. I should have known our data well enough to have caught this mistake, but I didn’t.

I know that our rankings are still very high—but that’s not the point; the point is that we put out information and data that wasn’t correct, and we feel morally obligated to set the record straight and apologize.