Skip to main content

Where Have All the Students Gone?

Stanford Economist Thomas Dee explores the reasons for and implications of the post-pandemic enrollment dip in public schools
Empty classroom with sun shining in

When the pandemic hit, Stanford Economist Thomas Dee knew it was important to track enrollment as a means of understanding what children are experiencing. He discovered that 1.2 million students didn't enroll in public school. Dee's data reveals not only where these children went but also leaves a significant number of children unaccounted for. A closer look at this data provides some insight into aspects of academic recovery that might also be missing.

"So much of our academic recovery discourse is focused on the kids still in public schools, and particularly those older kids who are in the kinds of testing windows that draw our attention," he says. "But the enrollment data are telling us really that some of the most substantial reductions in enrollment are among younger students, who to this day haven't yet aged into testing windows and won't until we hit the fiscal cliff, when the federal resources available to school districts run out."

The implications for such significant enrollment changes, Dee tells the Harvard EdCast, range from challenges in instruction and how to distribute funding properly to possible layoffs and school closures. 


JILL ANDERSON: I'm Jill Anderson. This is the Harvard EdCast. 

Stanford economist Thomas Dee was surprised to discover that 1.2 million students never enrolled in public school during the pandemic. His new research shows many of these students went to private school or homeschool, but there is still about a third of students unaccounted for. He says what happened to all those students is important, not just for their own personal outcomes, but also has implications for public education's future. 

I wanted to understand why enrollment numbers are so key in education and why we're seeing such drastic drops. First, I asked Tom to tell me more about how the pandemic impacted enrollment numbers nationwide. 

THOMAS DEE: Yeah, really quite striking. And the reason we like to underscore this is the pandemic was such a shock to all the normal data indicators we would look to to understand how our children are doing. But it was difficult to say turn to test score data or other things when they simply weren't available, or when they were available, it was kids taking different tests under different conditions. And frankly, different kids, given the enrollment changes. 

So in the stormy sea of the pandemic, the enrollment data were really a key indicator, and I think continue to be so. A key indicator of the disruptions to learning environments that our children have been experiencing. 

JILL ANDERSON: Were you surprised when you saw these dramatic shifts happening in the data? 

THOMAS DEE: I was. And I was kind of the mind that to see the importance in that, the value of this as an indicator of what kids are experiencing. Because of some pre-pandemic work I'd done with Professor Mark Murphy at the University of Hawaii looking at enrollment changes in response to partnerships between local police and immigration and customs enforcement, there had been efforts to find a demographic impact of that, but they've been relying on survey data to try to track changes in undocumented populations. 

And so we had had the idea that looking at enrollment changes would be a better, more reliable indicator of that impact. So that was a study we did called "The Vanished Classmates Study," just prior to the pandemic. And with that as an antecedent when the pandemic hit, it was really clear to me that the enrollment data would provide important insights at a time when all our other compasses were, in a sense, broken. 

JILL ANDERSON: Right. It seemed like a lot of what you heard about during the pandemic was more about online instruction, or hybrid instruction, or how to do it better, or how to do it at all. I wasn't hearing anything about maybe kids aren't going to school, or maybe they're not going to enroll in public school. 

THOMAS DEE: I think that's right. And the other challenge we faced here was just getting the data together. It's a long and common lament in education to say that we don't quite have the data systems we'd like to have. But we really got exposed in that regard once the pandemic hit. So, for example, for us to even collect the most fundamental education data-- who is enrolled in school, what is the aggregate size-- required this kind of bespoke effort. 

So I worked in collaboration with data journalists here at Stanford, an excellent doctoral student, Elizabeth Huffaker, and a collaboration of journalists at The New York Times, and elsewhere to scour state and local sources, and to harmonize and collect the enrollment data that were initially reported out. 

JILL ANDERSON: How much do you know about the specifics of that 1.2 million kids? Do you know if it's predominantly students of color, if it's white students? Or do you just not have that information? 

THOMAS DEE: We have some, but I think far and away the most striking heterogeneity in that enrollment decline is how it was concentrated by grade. The enrollment declines were particularly sharp in kindergarten in fall 2020 and in some of the early elementary grades. And very little of that at the high school level. There's even some evidence that at the high school level, initially in the pandemic, that high school completion increased. 

And there's an older literature consistent with this, that when kids considering leaving high school don't necessarily have a job or something else drawing them towards that, they're more likely to just stay in school. So really that decline among young children was the most striking feature. 

And in my more recent work I've been underscoring that we still don't appreciate the salience of that. Because so much of our academic recovery discourse is focused on the kids still in public schools, and particularly those older kids who are in the kinds of testing windows that draw our attention. 

But the enrollment data are telling us really that some of the most substantial reductions in enrollment are among younger students, who to this day haven't yet aged into testing windows and won't until we hit the fiscal cliff, when the federal resources available to school districts run out. 

JILL ANDERSON: So I want to get into where these kids went because that's really interesting. When you look at the data, 30.4% increase in homeschooling. 

THOMAS DEE: That's right. This reflects the effort of a recent collaboration that I had, again with data journalists here at Stanford and Big Local News and the Associated Press. Because we are now three full years into the pandemic. And yes, we documented that striking exodus from public schools, but still had very little idea of where those children went. 


THOMAS DEE: And what we did know was deeply enigmatic. For example, most private schools are Catholic. And the National Catholic School Association actually also reported enrollment declines over the first two full school years of the pandemic. And there was one experimental survey from the census bureau that showed in the immediate aftermath of the pandemic that homeschooling did go up. But that was fall 2020. And there was no evidence on whether that increase sustained into the '21-22 school year, when schools were back in person. 

So again, an absurd condition with regard to the data systems and the indicators available to us. But rather than curse the darkness, we decided to light a candle and get after those data at all the state websites where we could find them, and through FOIA requests, and things of that sort. And we were able to get comprehensive data for about half the schoolchildren in the US. And in particular, distinguishing between changes in private school enrollment and homeschool enrollment. 

We found that both grew-- private school enrollment by 4%, and as you note, homeschool enrollment by 30%. It was that growth in homeschool enrollment that was one of the most surprising things I saw. Because, it was a large percent increase of a smaller base, but even in absolute numbers it dominated the growth of private school enrollment and sustained into the '21-22 school year. 

We see for every newly enrolled private school student, there were nearly two in homeschooling as recent as last year. And we know so little about what those kids are experiencing. 


THOMAS DEE: Again, there's some anecdote to suggest it's the children of the affluent. There's little reason to suspect the newly homeschooled kids will totally match the pattern we saw in the pandemic. But the harsh truth is we simply don't know who these kids are, what they are experiencing, what types of learning environments. 

And we're in a climate now too where we're also observing sharp run-ups in chronic absenteeism. And this is a bit offhand, but I sometimes like to quip that there's a concern that there might be a continuum between being chronically absent, truant, and what some people are calling homeschooling. We just don't know. 

JILL ANDERSON: Is this about COVID? Or is this about other things? Because it seems like some of those states are epicenters for the culture wars that might be happening. 

THOMAS DEE: Maybe, but I suspect a lot of it was also-- in general the flight from public schools was just, given its concentration among younger students, was a reaction to remote schooling. And, in fact, we have a quasi experimental study that really identifies this in what we think is a credibly causal way, that the districts that adopted remote-only schooling saw that particular exodus of younger kids. 

Because they just-- understandably, many parents didn't want to sit a really young child in front of a computer all day and found other accommodations. What's surprising is the extent to which that was homeschooling, and the extent to which that sustained to a period when schools were back in person. 

JILL ANDERSON: A few states seem to have pretty big increases in their private school enrollments-- Massachusetts, Rhode island, Tennessee. 

THOMAS DEE: Yeah, I think there is a lot of heterogeneity by states. And I think some of that might reflect the availability of palatable private school options. Another thing that's really notable is the sharp run-up in homeschooling in Florida. 

JILL ANDERSON: Right, yeah. 

THOMAS DEE: We observed that at the beginning of the pandemic they had roughly 100,000 kids being officially recognized as homeschooled. And into fall 2021, we observed that grow by roughly 50%, roughly 50,000 additional homeschooled students. And I strongly suspect that's an undercount. 

The other thing we observe is demographic change is a major contributor to enrollment patterns. During the pandemic, there was a lot of domestic migration in response to an enduring work from home movement and response to housing costs. And Florida had a big influx of school-age children. And because many of those kids were new arrivals, and many of them might have been homeschooled, they're not necessarily being swept up in the official state counts available to us. 

JILL ANDERSON: And then, of course, the big story was the number of students that just are totally missing from school, unaccounted for, no one knows what happened to them. 

THOMAS DEE: This was one of the enigmas that really motivated this recent collaboration with the Associated Press, because of what I was observing, first here in California. California was a big loser of public school enrollment. We observed over the first two school years of the pandemic well over a quarter of a million student reduction in public schools, roughly 271,000 students. 


THOMAS DEE: But when we looked to the official state data on private schools and homeschoolings, less than 10% of that loss was picked up as increase enrollment there. Now we did see big population change. California did lose population during the pandemic. But that was less than 100,000 school-age children, according to the census estimates. 

So it raised a question of where are these kids? In California, it was roughly 150,000 students missing from public school enrollment changes, couldn't be accounted for by population change or non-public school enrollment. And we were also wondering the extent to which that occurred in other states. And we did observe it in a number of other states. 

There are at least, to my mind, three explanations for what might be going on here. One is that kids are truant, and we haven't picked up on that. The other is some of these kids actually are being homeschooled, but it's simply unregistered. So they're not showing up in the official counts. And all of these have developmental implications, implications for ed policy. 

And the third possible explanation-- and none of these are mutually exclusive also-- the third possible explanation is kids are skipping kindergarten. And we actually see some indirect evidence for this. Because when we look across all the states where we can see the extent to which kids are missing, the extent to which public school enrollment losses can't be explained by these other factors, it's really concentrated in the subset of states where kindergarten is not required. 

And, again, this just underscores my broader concern with our current academic recovery activities under the pandemic. It's what I've called in a recent op-ed "a streetlight effect," where there's this natural bias we have to look where the data are shining, to look to the kids who are represented in the data. But that streetlight effect can bias us away from other salient phenomena and problems where the data aren't shining so brightly. 

And I think that's the case for younger learners, and for kids who are missing from public schools right now. They're experiencing substantial disruptions to their learning environments, but they're not in the schools where other metrics are being taken. Or if they are, they're not yet in those tested grades. Yet, they're at this developmentally consequential stage and experiencing substantial disruption. 

So I'm not saying that we shouldn't pay attention as we do to things like high-impact tutoring and all the other thoughtful recovery activities people are undertaking, but I think we need to yes-end that with attention to younger students. And the available survey data from school business officials suggests that isn't happening. 

Like there was a widely-discussed survey of school district officials from last fall that showed that spending on early childhood, spending on outreach to families in the community were the least popular ways in which ESSER fundings, federal support, was being spent to address learning loss. 

JILL ANDERSON: So for those kids who didn't enroll in kindergarten, we would potentially see them at this point because now they would be in second grade, right? 

THOMAS DEE: Yeah, and we haven't really seen that return yet. One of the things I was really following after our initial study of that sharp fall 2020 drop in enrollment is what happened in fall 2021. And basically, those kids did not return. Overall, public school enrollment had a very modest, but an additional decline. 

But I was curious to see whether first grade enrollment would spike, which would imply kids had simply skipped kindergarten in states where they could. Or would we see kindergarten enrollment spike, as many kids effectively red-shirted. And what we saw is first grade enrollment continued to decline. 

And there was a modest bounce back in kindergarten enrollment, but nothing like the pre-pandemic levels. So there wasn't really a recovery there or any evidence of a dramatic red-shirting phenomenon. 

And I suspect what was going on, if we think back to summer of 2021, there was still a lot of uncertainty about what public schools might do in the future as we're cycling through additional variants and questions about the availability of vaccines for young kids. I think with that kind of uncertainty, many parents looked to safe harbor. And if they had found accommodations for their kids, they were likely to abide with that as circumstances evolved on the ground. 

JILL ANDERSON: And this is a huge challenge because you have no way of finding these kids if they're just essentially missing in a way. 

THOMAS DEE: It definitely raises challenges and questions about what is the proper social role of oversight here? Particularly, for example, with all these newly homeschooled kids. And you had mentioned this divisive political moment we're in. I can't help but think of how that exacerbates this challenge. You can imagine someone saying, hey, maybe we should look into what kinds of learning opportunities are available to all these kids who are being homeschooled, and an opportunistic politician trying to seek advantage through demagoguing that issue. So that's a concern. 

But I think it also-- if kids are skipping kindergarten, that scale which is indirectly suggested by what we're seeing, I think it raises challenges for those first grade teachers, who are going to be confronting kids who are rather different than what they saw before the pandemic, kids who are experiencing for the first time the more formal structure of a classroom. 

JILL ANDERSON: Someone listening to this might think, this is a good thing, right? Smaller schools, smaller classes. 

THOMAS DEE: There is great evidence that early exposure to a small classes is quite good. But we're also looking at kids who've been in some isolation and missed early learning opportunities. And there's also a pretty extensive literature that suggests those early learning opportunities are vital. It's like an extraordinarily propitious age for kids. And so that's a moment that's being lost. 

But, yes, there will be smaller class sizes. But I think the other moving part of this that many of us are anticipating is that when the federal support runs out, many school districts that haven't seen enrollment come back are going to be under strong financial pressure to close under-enrolled schools. And that's something that past experience tells us is really painful for communities, a really difficult process. 

And another extraordinary developmental risk for kids, who are attending closed schools. And it's really a function of where are they sent. In the past, we often see disadvantaged, often minoritized kids, are more likely to have their schools close and to be sent to underperforming schools. 

JILL ANDERSON: How long do you think it will be before we have a better understanding about the implications of this enrollment shift? 

THOMAS DEE: Well, I think it'll start in probably two or three years, as assessment data begin to roll in for the kids who experience these early learning disruptions. My hope is-- my encouragement to educators would be to try to field early assessments to identify where they might be able to target effective support to these kids. 

But in terms of the national conversation, I think as state assessments in third and fourth grade begin coming in from spring '24-25, we'll begin to see some evidence, at least for those kids who have returned to public schools. But we also have this big lacuna around those who simply aren't there, who are in private schools, and who may continue with homeschooling through that period. 

JILL ANDERSON: So you're going to continue, I'm imagining, tracking this and trying to see what happens, right? 

THOMAS DEE: We are, and I think also I want to monitor what might be happening with the enrollment pressure to potentially close schools. Because I think that's going to be such a pain point for many communities in the coming years. 

JILL ANDERSON: I didn't even ask you this, about states where you saw fully remote instruction versus states that remained open. 

THOMAS DEE: We were actually able to measure that at the district level. 


THOMAS DEE: And that was part of this study we have forthcoming at the American Educational Research Journal, that identifies that the decision to stay remote in fall of 2020 definitely contributed to that enrollment loss. We estimate that-- we see that roughly a little over half of districts went remote then. And that explained about a quarter of that 1.1 enrollment loss we observed in fall of 2020. 

But I want to be careful in interpreting that normatively. Because some people like to put a culture war frame on that, and said, see, we should have stayed open. I view that, instead, as one part of a larger conversation we have to have about how we view the decisions we made in the past and the decisions we might make in the future. 

So, for example, California and Florida provide an interesting contrast. California, like Massachusetts, tended to keep its schools closed, and it experienced much more dramatic enrollment loss. Florida stayed open. It didn't experience quite the same enrollment loss. But if you look at the excess mortality statistics, Florida had-- even adjusting for population size, much higher mortality rate during COVID than California did. 

So these are complex, difficult things. And people sometimes press me, what should we have done? Given what we've learned, what would have been a better path forward to balance very real public health concerns during a pandemic with the very substantial developmental consequences that keeping schools closed can have for kids? 

And, I guess, what I kind of gravitate towards is something that's a bit more of a hybrid solution. We know, for example, that the very youngest kids were not as susceptible to health risks during the pandemic, while older kids were, and that the younger kids experienced perhaps more dramatic learning consequences. 

So I think a really nimble and smart school district should gravitate toward something that's a bit more of a hybrid solution, where maybe they create more remote learning opportunities for older students, possibly with touch-points, where they come to campus every week or every month for some social engagement, and some type of counseling, and engagement with their teachers. And within the expanded physical space that makes available, younger kids attend in-person school more frequently.
I know that's a difficult logistical thing to do and creates staffing and curricular challenges. But it also, I think, meets the evidence about public health risks and developmental risks, and balances them in a way that seems appropriate. I know that's not very satisfying to people who want a one-size-fits-all panacea solution, and preferably one that accords with their political priors. But that strikes me as a better way to anticipate a compelling solution to the future pandemics I hope we never see. 

JILL ANDERSON: Thomas Dee is an economist and a professor at Stanford Graduate School of Education. I'm Jill Anderson. This is the Harvard EdCast produced by the Harvard Graduate School of Education. Thanks for listening. 


An education podcast that keeps the focus simple: what makes a difference for learners, educators, parents, and communities

Related Articles