I've lived in TN since 1989. I raised two boys here. My wife is a retired teacher and a current elementary school volunteer in a rural town. I love TN, but please stop talking about "the Tennessee miracle" in public education. The public school system here is bad. It was bad in Nashville and it's bad in this rural county. God bless everyone who knows this and is working hard to make it better. But, for the record, I don't want all of Matt's readers around the country to think that Tennessee actually has a good public school system. It does not, apart from a few affluent areas, which is what you'd expect.
In just about every area of life people like and hope to have choices: where to live, what to wear, what to eat, what doctor to use, etc. Why is it so surprising that people also want choice as to who teaches their children? You don't need to wish teachers' unions harm to want the best for your children.
I also see it as perfectly reasonable for the government to contribute any amount up to the cost of a public school education to private tuition at qualified schools (at least in areas where population density could support multiple competing schools). It's not a request for a "subsidy" or a special privilege -- just a request for the freedom to accomplish the public goal of educating children in in the way their parents think best.
You seem utterly fixated on the politics of it, to the point where you don't appear to be able to listen to simple justice without imputing enmity.
> Matthew Yglesias: Schools are getting worse in most red states <https://www.slowboring.com/p/schools-are-getting-worse-in-most>: 'The basic issue here, as we know from the higher education market, is that schools don’t really compete on the basis of being highly effective at teaching students. The “best” colleges in America are the ones that attract the strongest applicants. And most students just want to go to the “best” schools. Nobody even attempts to measure whether the teaching and learning practices at Harvard and Princeton are effective; it’s not considered relevant to the undertaking...
> Affirmative Action, Mismatch, and Economic Mobility after California’s Proposition 209. 2022. Quarterly Journal of Economics 137(1): 115–160.
> Affirmative Action and Its Race-Neutral Alternatives. 2023. Journal of Public Economics 220: 104839.
> Will Studying Economics Make You Rich? A Regression Discontinuity Analysis of the Returns to College Major, with Aashish Mehta. 2022. American Economic Journal: Applied Economics 14(2): 1-22 (Lead Article).
> Metrics that Matter: Counting What’s Really Important to College Students, with Mukul Kumar, Aashish Mehta, Christopher Muellerleile, and Christopher Newfield. 2023. Baltimore: Johns Hopkins University Press.
> On the Meritocratic Allocation of Higher Education
> College Major Restrictions and Student Stratification, with Aashish Mehta. NBER Working Paper 33269.
> Changes in the College Mobility Pipeline Since 1900, with Sarah Quincy. NBER Working Paper 33797. Reject and Resubmit, Quarterly Journal of Economics.
There's at least one non-Southern red state that has bucked the trend of declining reading scores: Indiana. On the most recent NAEP test, in 2024, it jumped to sixth in the nation for fourth- AND eighth-grade reading from its previous positions of 19th and 17th, respectively.
But I don't think that had anything to do with accountability. Standardized reading tests, which are divorced from any content kids might learn in school, are extremely blunt instruments for measuring progress -- and in fact are counterproductive, because they incentivize teachers to focus on the reading comprehension skills that the tests purportedly measure. In fact, what largely determines performance on those tests (and success at higher grade levels) is background knowledge and vocabulary, which most schools aren't focusing on building.
Yes, NCLB accountability is associated with an increase in MATH scores, but reading scores have essentially been stagnant or declining since 1998.
The results of any education reforms, policies, testing ideas, theories about how to teach this or that, etc. are sadly subject to something that educators can only minimally control: student motivation. Years ago a close friend of mine taught in the elementary schools of a poor, urban city where most students were not white. I’ve never forgotten when he told me that the kids entering first grade were bright, curious, engaged and successful, but by the fourth grade they had figured out what US culture and society were really like, and whatever they did in school was probably useless, and gave up trying.
Matt, as always I appreciate your interest in and analysis of this issue despite the national media having no interest in schooling unless culture wars are involved.
One thing we need to consider: as all SB fans know, if you look at economic graphs in the pandemic period there are wild distortions in employment, average wages, GDP, etc. Those have mostly recovered and resumed basically historically normal trajectories.
The disruption to EDUCATION cased by Covid is going to linger in the stats for a long time, because the youngest kids to have an extreme educational disruption won’t be graduating until the 2030s. And it will not, at any point, be easy to disaggregate the “what we’re doing wrong” statistics from the “that age cohort really got screwed” statistics. So when we analyze policy, we should consider that pre-2020 statistics are probably still a better place to start evaluating them, because it will be hard to tell what works with the Covid cohort with so many confounding variables.
A key things to watch is whether kids who started school in 2022 or later are doing better or worse or the same as their older siblings. Stats are starting to trickle in as those kids move through elementary school.
Vouchers on average draw less money per student than the regular public school system and in many instances are yielding better results. It’s not a simple equation, but for multiple reasons it’s true.
Louisiana is cite as having vastly improved. New Orleans replaced a miserable public system after Katrina with a commitment to charters which isn’t a ringing endorsement for public systems.
In my opinion, no child left behind wasn’t opposed by the unions because low achieving schools could suffer mild consequences. It was because it was an opening wedge that could result in accountability and consequences for low achieving teachers. While socialization is a key teaching objective, there is no reason in an orderly environment that it precludes measurable gains in reading and arithmetic.
This really dodges the tough questions. Sure, we want schools to be better. Great! But how do you do that? And how do you measure it? Education "reformers" seem to think that it involves schools "competing" for students via charters. That's pretty clearly a failure; charters seem to be good at skimming poor students whose parents are committed to their kids' education out of public schools, and then finding pretexts to kick out those who might perform poorly to deliver the illusion of success (a good friend of mine was a teacher when Eva Moskowitz's Success schools; what she described was 90% of the model).
But equally important, are test scores a good way to measure outcomes? You could probably deliver some results by teaching kids to take tests. But how good a proxy is test taking for... much of anything really?
Reducing the gap vs improving overall performance is a continuing dilemma for progrogressives obsessed with differences between identity groups. Too often we see these as zero sum issues. It's silly to think that one group can't get better unless another group losses something. I've seen real educational gains among low income students ignored because high performing students also made gains and the gaps persisted. Likewise reducing abject poverty is an absolute good even if it does not automatically reduce income inequality.
An acquaintance who is a public school principal once said to me, “The parents complain but the kids age out and we’re still here.”
If I am unhappy with the food or service at a local restaurant, I don’t circulate petitions or attend board meetings; I simply eat somewhere else. This system works pretty well - I usually have no trouble finding a place that I like at a reasonable price. And many of the restaurants I didn’t like have gone out of business.
Now isn’t a child’s education at least as important as finding the best chicken sandwich? Why shouldn’t education be subject to the most powerful tool of accountability - choice?
You might answer that some parents might make the wrong choices - burgers and fries instead of a healthy salad. But saying you don’t trust parents to make the best choices for their kids is like saying you don’t trust people to make the right choices at the ballot box - you don’t trust freedom. And that’s a whole different conversation that goes beyond education.
So how about we start looking for an international cause to an international issue. Maybe it is the phones and social media. Or to put it another way, maybe the billions of dollars invested in R&D by multiple trillion-dollar companies, with psychiatrists and economists, have succeeded in taking large amounts of attention away from school work and socialization, and/or shortening attention span. Neither the end of NCLB, nor dropping the SATs, nor whatever DEI is being blamed for this week is what caused a decline in Germany or Poland or Iceland.
Democrats are not going to be trusted on education until they throw the teachers unions under the bus. The strongest indicator of school closures and child masking was strength of teachers union, when the rest of the developed world had schools open (without spending hundreds of billions). This follows on them getting involved in some of the most extreme culture war stuff. Then, Randi Weingarten writes op-Ed’s about how she pushed for school opening, when she clearly did not. It was so self transparently self-interested, and not particularly well hidden.
While we see underperforming schools all over the country, I do wonder how much of this underperformance can be connected to bad pedagogy that is pushed from few prestigious institutions?
Bad literacy pedagogy (whole world/Lucy Caulkins method) appears to have been pushed by Columbia teachers school, and sold to districts throughout the country.
One of my concerns is that if we punish poorly preforming schools with lower funding, isn't it more likely that they will just continue to get worse and enter into a death spiral?
I wonder if we could structure this by having the state or feds give direct income bonuses to superintendents/principals and teachers that is based on school performance. That way the schools won't atrophy if they don't get the bonuses, but there is a clear incentive to improve performance. Although this would also greatly increase the incentive for cheating.
"Although this would also greatly increase the incentive for cheating."
My immediate reaction when I started reading your proposal was to think of the multiple documented instances of schools cheating on performance metrics even where there was no direct financial benefit to the administrators or teachers. If people actually directly get paid for student performance, I think you're going to end up with cheating scandals that collectively make the Minnesota Medicaid fraud look like peanuts.
One of the things that I think did not work well in NCLB was the measurement of school quality, based largely on meeting certain achievement thresholds. Alas, achievement is largely determined not by the quality of the school but by the characteristics of the student body. The result was that the schools subject to closure were those educating the most challenging students, even if doing a semi-decent job of it. The result was unhelpful churn and turmoil for those neediest students, and nobody else. Ideally, schools would be measured by intake achievement versus output achievement. If that’s too hard, there at least needed to be realistic SES adjustments.
Matt—this is exactly the kind of rigorous analysis that keeps me subscribing. You're right that the "Southern Surge" can't be extrapolated to "red state governance works." Four states isn't a pattern; it's an exception that demands explanation.
But I'd push back on one implicit assumption: that we're watching essentially random variation in state competence. You note that these states "are trying" while others aren't—but that begs the question: why do some states try?
I'm working on a book that offers a framework for this. The core insight: tensions are neutral; identity determines transformation. Every state faces the same tension (declining student achievement). What differs is how state-level actors conceptualize themselves in relation to that tension.
Mississippi, Louisiana, Alabama, Tennessee adopted what I'd call "architect identity"—they saw themselves as agents capable of fixing the problem. This enabled them to implement comprehensive reforms (science of reading, universal screening, accountability, retention with supports) and sustain them over 10+ years.
Texas and Florida adopted complacent identity—they'd won earlier battles and stopped innovating. Their mental structures ("we figured this out in 2003") prevented them from seeing that the game had changed.
Union-dominated states adopted victim identity—external forces (poverty, inequality, funding levels) cause the problem, so solutions must come from above. This mental structure literally filters out the Mississippi model. They can't see it because their frameworks don't include "we can fix this ourselves with evidence-based instruction."
Your Virginia test case perfectly illustrates this. Spanberger has a choice: adopt architect identity ("we will implement accountability and get results") or defer to mental structures that make accountability invisible. The former is harder but works. The latter is politically easier but guarantees continued decline.
The deeper insight from your data: accountability isn't just a policy preference—it requires a particular identity stance. States that see themselves as agents who measure and adjust will improve. States that see themselves as victims of forces beyond their control will not.
Same tension. Different identities. Different outcomes.
I've lived in TN since 1989. I raised two boys here. My wife is a retired teacher and a current elementary school volunteer in a rural town. I love TN, but please stop talking about "the Tennessee miracle" in public education. The public school system here is bad. It was bad in Nashville and it's bad in this rural county. God bless everyone who knows this and is working hard to make it better. But, for the record, I don't want all of Matt's readers around the country to think that Tennessee actually has a good public school system. It does not, apart from a few affluent areas, which is what you'd expect.
In just about every area of life people like and hope to have choices: where to live, what to wear, what to eat, what doctor to use, etc. Why is it so surprising that people also want choice as to who teaches their children? You don't need to wish teachers' unions harm to want the best for your children.
I also see it as perfectly reasonable for the government to contribute any amount up to the cost of a public school education to private tuition at qualified schools (at least in areas where population density could support multiple competing schools). It's not a request for a "subsidy" or a special privilege -- just a request for the freedom to accomplish the public goal of educating children in in the way their parents think best.
You seem utterly fixated on the politics of it, to the point where you don't appear to be able to listen to simple justice without imputing enmity.
> Matthew Yglesias: Schools are getting worse in most red states <https://www.slowboring.com/p/schools-are-getting-worse-in-most>: 'The basic issue here, as we know from the higher education market, is that schools don’t really compete on the basis of being highly effective at teaching students. The “best” colleges in America are the ones that attract the strongest applicants. And most students just want to go to the “best” schools. Nobody even attempts to measure whether the teaching and learning practices at Harvard and Princeton are effective; it’s not considered relevant to the undertaking...
Ahem. My former student Zach Bleemer:
> <https://zacharybleemer.com/CV,%20Zach%20Bleemer.pdf>
> Affirmative Action, Mismatch, and Economic Mobility after California’s Proposition 209. 2022. Quarterly Journal of Economics 137(1): 115–160.
> Affirmative Action and Its Race-Neutral Alternatives. 2023. Journal of Public Economics 220: 104839.
> Will Studying Economics Make You Rich? A Regression Discontinuity Analysis of the Returns to College Major, with Aashish Mehta. 2022. American Economic Journal: Applied Economics 14(2): 1-22 (Lead Article).
> Metrics that Matter: Counting What’s Really Important to College Students, with Mukul Kumar, Aashish Mehta, Christopher Muellerleile, and Christopher Newfield. 2023. Baltimore: Johns Hopkins University Press.
> On the Meritocratic Allocation of Higher Education
> College Major Restrictions and Student Stratification, with Aashish Mehta. NBER Working Paper 33269.
> Changes in the College Mobility Pipeline Since 1900, with Sarah Quincy. NBER Working Paper 33797. Reject and Resubmit, Quarterly Journal of Economics.
There's at least one non-Southern red state that has bucked the trend of declining reading scores: Indiana. On the most recent NAEP test, in 2024, it jumped to sixth in the nation for fourth- AND eighth-grade reading from its previous positions of 19th and 17th, respectively.
But I don't think that had anything to do with accountability. Standardized reading tests, which are divorced from any content kids might learn in school, are extremely blunt instruments for measuring progress -- and in fact are counterproductive, because they incentivize teachers to focus on the reading comprehension skills that the tests purportedly measure. In fact, what largely determines performance on those tests (and success at higher grade levels) is background knowledge and vocabulary, which most schools aren't focusing on building.
Yes, NCLB accountability is associated with an increase in MATH scores, but reading scores have essentially been stagnant or declining since 1998.
The results of any education reforms, policies, testing ideas, theories about how to teach this or that, etc. are sadly subject to something that educators can only minimally control: student motivation. Years ago a close friend of mine taught in the elementary schools of a poor, urban city where most students were not white. I’ve never forgotten when he told me that the kids entering first grade were bright, curious, engaged and successful, but by the fourth grade they had figured out what US culture and society were really like, and whatever they did in school was probably useless, and gave up trying.
Matt, as always I appreciate your interest in and analysis of this issue despite the national media having no interest in schooling unless culture wars are involved.
One thing we need to consider: as all SB fans know, if you look at economic graphs in the pandemic period there are wild distortions in employment, average wages, GDP, etc. Those have mostly recovered and resumed basically historically normal trajectories.
The disruption to EDUCATION cased by Covid is going to linger in the stats for a long time, because the youngest kids to have an extreme educational disruption won’t be graduating until the 2030s. And it will not, at any point, be easy to disaggregate the “what we’re doing wrong” statistics from the “that age cohort really got screwed” statistics. So when we analyze policy, we should consider that pre-2020 statistics are probably still a better place to start evaluating them, because it will be hard to tell what works with the Covid cohort with so many confounding variables.
A key things to watch is whether kids who started school in 2022 or later are doing better or worse or the same as their older siblings. Stats are starting to trickle in as those kids move through elementary school.
Vouchers on average draw less money per student than the regular public school system and in many instances are yielding better results. It’s not a simple equation, but for multiple reasons it’s true.
Louisiana is cite as having vastly improved. New Orleans replaced a miserable public system after Katrina with a commitment to charters which isn’t a ringing endorsement for public systems.
In my opinion, no child left behind wasn’t opposed by the unions because low achieving schools could suffer mild consequences. It was because it was an opening wedge that could result in accountability and consequences for low achieving teachers. While socialization is a key teaching objective, there is no reason in an orderly environment that it precludes measurable gains in reading and arithmetic.
This really dodges the tough questions. Sure, we want schools to be better. Great! But how do you do that? And how do you measure it? Education "reformers" seem to think that it involves schools "competing" for students via charters. That's pretty clearly a failure; charters seem to be good at skimming poor students whose parents are committed to their kids' education out of public schools, and then finding pretexts to kick out those who might perform poorly to deliver the illusion of success (a good friend of mine was a teacher when Eva Moskowitz's Success schools; what she described was 90% of the model).
But equally important, are test scores a good way to measure outcomes? You could probably deliver some results by teaching kids to take tests. But how good a proxy is test taking for... much of anything really?
Reducing the gap vs improving overall performance is a continuing dilemma for progrogressives obsessed with differences between identity groups. Too often we see these as zero sum issues. It's silly to think that one group can't get better unless another group losses something. I've seen real educational gains among low income students ignored because high performing students also made gains and the gaps persisted. Likewise reducing abject poverty is an absolute good even if it does not automatically reduce income inequality.
An acquaintance who is a public school principal once said to me, “The parents complain but the kids age out and we’re still here.”
If I am unhappy with the food or service at a local restaurant, I don’t circulate petitions or attend board meetings; I simply eat somewhere else. This system works pretty well - I usually have no trouble finding a place that I like at a reasonable price. And many of the restaurants I didn’t like have gone out of business.
Now isn’t a child’s education at least as important as finding the best chicken sandwich? Why shouldn’t education be subject to the most powerful tool of accountability - choice?
You might answer that some parents might make the wrong choices - burgers and fries instead of a healthy salad. But saying you don’t trust parents to make the best choices for their kids is like saying you don’t trust people to make the right choices at the ballot box - you don’t trust freedom. And that’s a whole different conversation that goes beyond education.
There has been an international decline since about 2013 -- not all countries, but the overall trend is downward based on PISA tests. See e.g., https://www.reuters.com/world/mathematics-reading-skills-unprecedented-decline-teenagers-oecd-survey-2023-12-05/#:~:text=Summary,Reuters%20Graphics. (e.g., "Germany, Iceland, the Netherlands, Norway and Poland saw particularly sharp drops in mathematics scores, the OECD said.") They note the declines predate COVID.
So how about we start looking for an international cause to an international issue. Maybe it is the phones and social media. Or to put it another way, maybe the billions of dollars invested in R&D by multiple trillion-dollar companies, with psychiatrists and economists, have succeeded in taking large amounts of attention away from school work and socialization, and/or shortening attention span. Neither the end of NCLB, nor dropping the SATs, nor whatever DEI is being blamed for this week is what caused a decline in Germany or Poland or Iceland.
Democrats are not going to be trusted on education until they throw the teachers unions under the bus. The strongest indicator of school closures and child masking was strength of teachers union, when the rest of the developed world had schools open (without spending hundreds of billions). This follows on them getting involved in some of the most extreme culture war stuff. Then, Randi Weingarten writes op-Ed’s about how she pushed for school opening, when she clearly did not. It was so self transparently self-interested, and not particularly well hidden.
While we see underperforming schools all over the country, I do wonder how much of this underperformance can be connected to bad pedagogy that is pushed from few prestigious institutions?
Bad literacy pedagogy (whole world/Lucy Caulkins method) appears to have been pushed by Columbia teachers school, and sold to districts throughout the country.
One of my concerns is that if we punish poorly preforming schools with lower funding, isn't it more likely that they will just continue to get worse and enter into a death spiral?
I wonder if we could structure this by having the state or feds give direct income bonuses to superintendents/principals and teachers that is based on school performance. That way the schools won't atrophy if they don't get the bonuses, but there is a clear incentive to improve performance. Although this would also greatly increase the incentive for cheating.
"Although this would also greatly increase the incentive for cheating."
My immediate reaction when I started reading your proposal was to think of the multiple documented instances of schools cheating on performance metrics even where there was no direct financial benefit to the administrators or teachers. If people actually directly get paid for student performance, I think you're going to end up with cheating scandals that collectively make the Minnesota Medicaid fraud look like peanuts.
One of the things that I think did not work well in NCLB was the measurement of school quality, based largely on meeting certain achievement thresholds. Alas, achievement is largely determined not by the quality of the school but by the characteristics of the student body. The result was that the schools subject to closure were those educating the most challenging students, even if doing a semi-decent job of it. The result was unhelpful churn and turmoil for those neediest students, and nobody else. Ideally, schools would be measured by intake achievement versus output achievement. If that’s too hard, there at least needed to be realistic SES adjustments.
“…achievement is largely determined not by the quality of the school but by the characteristics of the student body”
The student body does not set school policies such as, for example, social promotion.
That is correct, but it is not inconsistent with my comment
Can't we just do the DODEA thing of giving teachers $80,000 starting wages and then teachers unions become irrelevant?
How much better if we simply outlawed unions.
Matt—this is exactly the kind of rigorous analysis that keeps me subscribing. You're right that the "Southern Surge" can't be extrapolated to "red state governance works." Four states isn't a pattern; it's an exception that demands explanation.
But I'd push back on one implicit assumption: that we're watching essentially random variation in state competence. You note that these states "are trying" while others aren't—but that begs the question: why do some states try?
I'm working on a book that offers a framework for this. The core insight: tensions are neutral; identity determines transformation. Every state faces the same tension (declining student achievement). What differs is how state-level actors conceptualize themselves in relation to that tension.
Mississippi, Louisiana, Alabama, Tennessee adopted what I'd call "architect identity"—they saw themselves as agents capable of fixing the problem. This enabled them to implement comprehensive reforms (science of reading, universal screening, accountability, retention with supports) and sustain them over 10+ years.
Texas and Florida adopted complacent identity—they'd won earlier battles and stopped innovating. Their mental structures ("we figured this out in 2003") prevented them from seeing that the game had changed.
Union-dominated states adopted victim identity—external forces (poverty, inequality, funding levels) cause the problem, so solutions must come from above. This mental structure literally filters out the Mississippi model. They can't see it because their frameworks don't include "we can fix this ourselves with evidence-based instruction."
Your Virginia test case perfectly illustrates this. Spanberger has a choice: adopt architect identity ("we will implement accountability and get results") or defer to mental structures that make accountability invisible. The former is harder but works. The latter is politically easier but guarantees continued decline.
The deeper insight from your data: accountability isn't just a policy preference—it requires a particular identity stance. States that see themselves as agents who measure and adjust will improve. States that see themselves as victims of forces beyond their control will not.
Same tension. Different identities. Different outcomes.