‘Disappointing, there’s no other way to say it’: Researcher Tom Loveless on the Legacy of Common Core

Get stories like these delivered straight to your inbox. Sign up for The 74 Newsletter

See previous 74 Interviews: Author Jal Mehta on the value of teaching, Harvard scholar David Perkins on “playing the whole game,” and Professor Nell Duke on project-based learning and standards. The full archive is here. 

Whatever happened to Common Core?

That’s the question that veteran education researcher Tom Loveless asks in the final chapter of his new book, Between the State and the Schoolhouse: Understanding the Failure of Common Core. Released this spring by Harvard Education Press, the slim volume examines the debate around the ambitious reform and the inherent limits of trying to improve education systems through regulatory means.

To the regret of its (often very vocal) detractors, nothing much seems to have happened to Common Core; even after a furious political battle in the late Obama years, most states still have some version of the controversial academic standards on the books. States attempting to replace them with new learning frameworks were often engaged in more of a branding exercise than a substantive overhaul, and once a few years had passed, politicians moved on to new skirmishes in the education culture wars.

But a decade after they were first adopted by states, little evidence exists to show that teaching or learning was significantly improved by the vast resources poured into implementing the standards. At least one study has found students in states that were early adopters of Common Core scored slightly lower on both the National Assessment of Educational Progress’s reading and math portions. If the point of spending billions of dollars to establish the mammoth set of new learning guidelines was to make sure kids became “college- and career-ready” (to use a term that was ubiquitous around 2013), not much progress seems to have been made toward that goal.

A former sixth-grade teacher, Harvard professor, and senior fellow at the Brookings Institution, Loveless has watched the development of academic standards for decades, ultimately concluding that they are an ineffective tool to improve K-12 education. As he argued to The 74’s Kevin Mahnken, regulatory reforms like Common Core are riven with utopian expectations and unlikely to change what actually goes on in classrooms.

“The problem is inherent to top-down efforts at controlling curriculum and instruction,” Loveless writes in the book. “This is not a problem that another set of standards can solve. If standards came out tomorrow, and I agreed with every single word in them, I would still give them only a slim chance of being faithfully implemented — and less than that of moving the needle on student achievement.”

This interview was edited for length and clarity.

Kevin Mahnken: Your book focuses deeply on the shortcomings of academic standards. But as a reporter, the impression I’ve developed has been that K-12 education has really been driven over the last few decades by testing and accountability reforms like No Child Left Behind. Do you think people underestimate the importance of standards — not just Common Core, but also the state based standards that preceded it?

Tom Loveless: The accountability movement of the ’90s was all based on standards. There was no state in my memory that went out and created an accountability system where the accountability was based on something other than standards. They all had tests, which were written on a grade-level basis to conform with the standards those states had adopted. So it’s hard to untangle accountability from the question of standards.

In the book, I took a much longer historical perspective. I go back over 100 years to look at standards as a regulatory tool: You’ve got upper-level officials who are trying to influence what schools do with kids in terms of what they teach. That’s been going on forever, and always with limited success. It’s hard for the top of the system to have a large impact on what happens at the bottom of the system.”

That sounds right in terms of the different levers of school reform — tests are based on standards, grad schools prepare future educators to teach to those standards, etc. So they’re at the center of things.

Right, but there’s a nuance there: Those early accountability systems were not about making sure teachers followed the standards; they made sure that teachers and schools produced scores on tests that were aligned to the standards. That’s actually a completely different thing. It was test-based accountability, and there’s a separate literature on that that’s fairly positive. If you hold schools accountable for scoring on a test, and have either rewards or sanctions, you can raise those test scores. There are three or four well-designed studies that show that.

But that’s a whole different issue from what Common Core was about. If you go back and read all the Common Core documents, those standards don’t touch the accountability question at all. And as a matter of fact, the accountability systems post-Common Core — some of the Common Core authors suspect this is why Common Core had little impact — withered away. We have very soft accountability today compared with NCLB, which kind of poisoned the waters for accountability because of the way it was designed.

Do you think the basic proposition of standards-based reform — i.e., that some students just weren’t being held to high standards — was valid? It sounds like you’re saying that rigorous academic expectations aren’t enough on their own to improve K-12 education, but are they a necessary ingredient?

Yes, some states did have standards that were too low. Some districts, some schools, some teachers had standards that were too low. But the question is, can you then force states with low standards to have high standards, and will that have a positive impact? I don’t think you can.

There was a lot of research in the ’90s and the ’00s: Mississippi or some other state had terrible standards, and lots of kids were scoring proficient, but on NAEP, they never even got close [to proficiency]. So obviously the state has much lower standards than what you’d want. But the people in Mississippi read the newspaper; they know their NAEP scores. And where’s the political pressure from the state, from the bottom up, to fix that? Now, in a lot of places, there was that pressure. But if it’s not there, can you come in from some supra-state level and force higher standards onto a state that they implement with fidelity, and eventually believe in? Because if they don’t, you’re probably not going to get very much.

Now take that same argument and just swap out the actors: Can a state come in and do the same thing to a reluctant district? Can a district come in and do it with a reluctant school? See, I don’t even think a school principal can do that in his own building with a teacher who has low standards. So the idea that we’re going to have this broad-scaled, top-down implementation of standards in a way that improves learning — that’s the thing I’m skeptical of. It’s just never worked, and it didn’t work with Common Core. So the whole approach is flawed.

The most recent evidence I’ve seen about the impact of Common Core on academic achievement comes from Joshua Bleiberg’s study in AERA Open, which found a pretty modest boost to NAEP math scores. Is that typical of the research findings thus far, and do we have reason to think that the reform’s effects could grow with time?

I consider the Bleiberg effect, a positive effect of about .1 standard deviations, to be the upper bound of what the different studies show. The C-SAIL study, which I spend more time with in the book, shows a .1 [standard deviation] decrease, which is kind of the lower bound, and all of my own studies fall in between those two boundaries. The probable real effect of Common Core — although I’m not that confident in any of these studies, including my own — is probably somewhere within that range. And that is disappointing, there’s no other way to say it. Especially over many, many years of implementation, all the money that was spent on it, all the teacher development, and the debate that got so bitter and nutty. What a distraction to get us so fired up over one-tenth of a standard deviation. It’s just miniscule.

There’s one thing in the study that gets at the question: “What if we just stuck with this thing? Maybe there are great things that are going to happen just over the horizon.” If you read Bleiberg’s analysis, most of the effect kicks in after the first two years. It’s not going up; if anything, it’s petering out. The C-SAIL study found that the effect was not only negative, but that it was getting more negative over time. So even though those two studies have different signs in front of the effect — one’s positive, one’s negative — they really kind of find the same thing: The most positive impact was very early in the process of implementing Common Core. To me, that makes total sense because all the professional development, the initial billions of dollars, was all spent in the first few years to get this thing off the ground. I don’t know any study of professional development that says, ‘Oh, wait a decade, and then really good things kick in.’ It just doesn’t work that way.

You mentioned that you’re not totally sure about the findings in these studies, including your own. What are the challenges in measuring effects from reforms like Common Core?

In my work, I don’t even make a causal claim because there are too many impediments to do that.

Both Bleiberg and the C-SAIL study used an interrupted time series design. In order to do that, you need to have a very clear break period: Here’s when this thing didn’t exist, and then on this day, it existed. There are studies that use that design very effectively — for instance, a Josh Angrist study of [the effects of] lowering the age at which people can buy alcohol, which was a big issue in the ’70s. A lot of states lowered their legal drinking age from 21 to 18, and those laws went into effect at midnight on January 1. So suddenly, the bars were filled with 18-, 19-, and 20-year-olds, whereas before, they couldn’t get in. There was a clear cut-point in the state’s actions that could be measured in terms of pre and post. Academic standards just don’t work that way.

A few different researchers studying Common Core, including myself, ended up going about it in the same way. Virtually all the states in the country adopted Common Core, and you had to sort them: one group that really did Common Core, another group that sort of adopted it and did a half-baked approach, and then the five states that just rejected Common Core. Those were the three groups whose NAEP scores I tried to measure over time. Pretty much all of my analysis showed the same thing, which was very little effect.

Another problem was that the natural comparison group is the five states that rejected Common Core from the beginning: Texas, Virginia, Nebraska, Alaska, and Minnesota in math — they kept their existing math standards but adopted the ELA standards. But each of those states, if you go back and read the standards they did adopt, they’re not terribly different from Common Core. And it’s not as if Common Core was revolutionary; it wasn’t the first set of standards that said, ‘You know, we should teach kids fractions!’ I would argue that Common Core has 80-90 percent overlap with the previous standards that a lot of states had.

So that invites the notion of just what the change was. Of course, the Common Core people would say, ‘It led to better curriculum, better instruction, better tests,’ and again, there’s no evidence of that. Anyway, that’s just a taste of some of the methodological constraints on measuring this.

Is the main problem here that states and districts didn’t implement Common Core well? Or is it just asking too much of academic standards to expect them to really improve teaching and learning? It seems like Morgan Polikoff, a professor at the University of Southern California who also wrote a recent book about standards for Harvard, feels that it’s the latter. 

I just did an interview with Morgan on [American Enterprise Institute scholar] Nat Malkus’s podcast. He started with that question: Is it a problem of implementation, or a problem with the theory of action? Morgan and I both said that the theory itself is flawed. We can’t engineer our way to better standards.

Again, standards are a regulatory tool, and we’re not going to be able to simply regulate better K-12 learning. It’s not going to work that way. Just to give an example — and this isn’t necessarily bad or good implementation, it’s just what happens — when you ask some teachers or district people what the main tenets of Common Core mathematics are, they’ll say, ‘Well, kids need to be working in groups.’ And then they’ll list a bunch of other things that have nothing to do with Common Core, which does not mandate that kids work in groups. It doesn’t even talk about that! It was like NCLB in that if you asked people what it meant, you’d get different answers in different places.

Not long ago, I wrote an article about the press coverage of Common Core and its implementation. Within a couple of years of Common Core’s adoption, you’d have journalists attending these workshops where professional development was being given. And in a particular math workshop, the developer was saying all the stuff I just mentioned: “You need to put your kids in groups, you need to be using manipulatives, you need to deemphasize procedures and rote learning, you need to emphasize conceptual understanding.” Now, Common Core does shoulder some guilt on the conceptual understanding thing, but it doesn’t say you should deemphasize anything.

The point is that, everywhere across the country, we have educators who have belief systems of their own. And if they believe in putting kids in groups, or believe in what we used to call ‘progressive education,’ or student-centered instructional practices, they’re going to interpret any policy coming down the line to promote those things; they’re going to read the documents through that lens. It’s not a heartfelt effort to distort, and these people aren’t sinister. It’s just how they read things. So you’re going to get actual implementation that’s different from what’s on paper, like the old children’s game of telephone where things sound different at the end of the line. That’s not corrupt intent, it’s that you have so many people sifting through these things as they make their way down the system.

It sounds like if you want to really change instruction through academic standards, you’d have to be so prescriptive just to avoid people doing something totally unrelated to what you want. 

And besides that, standards tend to be utopian. They tend to be aspirational, wishful thinking, and Common Core is a clear example of that. Common Core used this phrase, “college- and career-ready,” and then mapped standards back from the twelfth grade. But nobody yet has defined “career-ready” in such a way that doesn’t really just mean “college-ready.” At least, I haven’t seen any good definition of career readiness come out of these standards movements. So you can just delete the word “career,” and essentially what these standards are saying is, “Everybody, 100 percent, will be ready for college by the end of high school.” That’s very much like NCLB’s 100 percent proficiency goal. So what did the test makers, both PARCC and Smarter Balanced, do? They adopted NAEP proficiency as their standard.

The last batch of data I saw from the states that still use Smarter Balanced showed that 32 percent of eleventh-graders pass in math, and 68 percent fall below the threshold indicating readiness. If Common Core were working at all — and if we should have faith in this test to measure a goal that we could actually achieve — we’d be doing better than 32 percent. I mean, are you going to deny a diploma to two-thirds of the kids because they fail math? Politically, it’s a non-starter.

A lot of the reformers point to high-achieving countries like Singapore and South Korea, but if you map international assessments like TIMSS and PISA onto NAEP proficiency, it shows that at least 25 percent of their populations would fail. And these are the highest-achieving countries on the planet. So the goals are ones that no society has ever attained, and it’s not going to happen.

You’ve also written previously about the fact that NAEP proficiency levels might just be set too high. The NCES commissioner basically said as much in a recent meeting of the National Assessments Governing Board.

It turns our national test, which should be something that gives real information, into a kind of disinformation. It makes it like one of those late-night cable ads: “You can look like this if you just buy cans of this stuff and drink it five times a day!” It doesn’t work, it winds up undermining the validity, and I think Common Core suffers from it all. If you look at the outcomes by the end of high school, they’re much more than ambitious; they’re unrealistic.

Get stories like these delivered straight to your inbox. Sign up for The 74 Newsletter