Teachers have come under fire across the country in the last few years, as a small group of hyper-conservative advocates and state policymakers have worked to ban books depicting diverse characters, and limit the topics educators can teach to students, all in the name of promoting “parents’ rights” and reigning in what they said was “liberal indoctrination” by educators. That has created an environment in which teachers are, very understandably, censoring themselves, even when their state or district doesn’t have regulations prohibiting the topics they can discuss.

Then the crusade went national: the Trump administration attacked equity initiatives and limited the topics of books in Defense Department-run schools. Earlier this fall, that attack on accurate history spurred even further to the national level, with the Department of Education creating a new definition of “patriotic education” that would sweep many of the ugliest — and most important — parts of the American story under the rug.

Besides the outstanding question of whether these values really are “patriotic” by any definition other than the one set by the Trump administration, the research the story relies on is flawed in several ways

Now comes an article in Education Week with the somewhat jarring headline, “Teachers Value ‘Patriotic’ Education More than Most Americans.” Rather than shed light on the important topics and diverse books that teachers are missing, the article — and the research it highlights — muddles the message of what teachers, parents, and Americans writ large think students should be learning.

Based on separate surveys of teachers, parents, and American adults — which were sponsored by EdChoice, an organization that promotes vouchers — researchers said K-12 teachers were more likely than parents or the public at large to say it’s important to teach that “The United States is fundamentally a good country,”  and about the Constitution’s core values.

Besides the outstanding question of whether these values really are “patriotic” by any definition other than the one set by the Trump administration, the research the story relies on is flawed in several ways.

First, it’s not nationally representative. The polling was conducted online, and data were weighted to approximate a target sample of teachers based on gender, age, ethnicity, race, years of experience, region, and several other undefined characteristics. This would be an acceptable methodology if the participants were chosen via random sampling. However, the polling firm did not select teachers randomly. Since the participants were not selected randomly, and there is limited information about the selection process, readers should not make any broad generalizations from the data.

Notably, the survey did not attempt to balance respondents’ political ideology, and the teachers sampled do not match that of the general population. Specifically, the demographic data released with the poll results indicate that 44% of respondents are MAGA supporters and 49% are MAGA opposers. But nationally, just one-third of registered voters identified themselves as MAGA supporters. The underlying sample of teachers likely does not reflect the true population of teachers. Based on available documentation, the poll methodology does not adjust for that bias.

Finally, as with all sample surveys, there is a margin of error — within which we cannot be sure that differences between groups are true differences.

The teacher survey has a margin of error of +/- 3.72 percentage points, the parent poll has a margin of error of +/- 3 percentage points and the survey for all adults has a margin of error of +/- 2.3 percentage points. As a result, differences in responses between teachers and parents/all respondents that are less than 4 percentage points are not notable differences; Toto be statistically significant, differences would need to be even larger. The only finding highlighted in the EdWeek article with a large enough difference is that parents were more likely to say it’s important to teach that it’s good to question policies and actions of the U.S. government.

Looking at the question that is most objectively related to “patriotic education,” based on the results from the three polls, we can only conclude that there is no difference in opinion between teachers, parents, and adults (who participated in these polls) on whether it’s important for schools to teach students to be patriotic and loyal toward the United States. So, EdChoice’s survey doesn’t offer any clear conclusion at all.

What is clear in this moment is that students deserve lucid, accurate lessons about the history of our country and its values, and their teachers are grappling with how they can make that happen. These are the kind of decisions that should happen after meaningful conversations between families and educators, not under a cloud of political rhetoric and flawed data.