Do Parents Value School Effectiveness?

by on October 10, 2017 at 12:02 pm in Data Source, Economics, Education | Permalink

That is a new NBER Working Paper by Atila Abdulkadiroglu, Parag A. Pathak, Jonathan Schellenberg, and Christopher R. Walters, on a much understudied topic.  Here are their main results:

School choice may lead to improvements in school productivity if parents’ choices reward effective schools and punish ineffective ones. This mechanism requires parents to choose schools based on causal effectiveness rather than peer characteristics. We study relationships among parent preferences, peer quality, and causal effects on outcomes for applicants to New York City’s centralized high school assignment mechanism. We use applicants’ rank-ordered choice lists to measure preferences and to construct selection-corrected estimates of treatment effects on test scores and high school graduation. We also estimate impacts on college attendance and college quality. Parents prefer schools that enroll high-achieving peers, and these schools generate larger improvements in short- and long-run student outcomes. We find no relationship between preferences and school effectiveness after controlling for peer quality.

You can read that as the parents either being super-smart about what helps their kids — better peers — or the parents being snobby per se.  Either way, they don’t seem to care so much about value-added from the side of the school.  Or is peer quality actually also the best practical-for-parents measure of value-added, when all is said and done?

So I say this is an inconclusive result from the final normative point of view, but a highly significant result in terms of cementing in some knowledge close to what I already expected was the case.

1 prior_test3 October 10, 2017 at 12:07 pm

‘a much understudied topic’

This is a joke, right?


2 Mr. Econotarian October 10, 2017 at 12:13 pm

Key findings:

“We show that a large fraction of ”bad” peers at school – as identified by students in the bottom 5% of the ability distribution – negatively and significantly affects the cognitive performance of other schoolmates. Importantly, as we show in our work, it is only the very bottom 5% students that (negatively) matter, and not ‘bad’ peers in other parts of the ability distribution (e.g. the 5-to-10% worst students).

On the other hand, we uncover little evidence that the average peer quality and the share of very ‘good’ peers – as identified by students in the top 5% of the ability distribution affect the educational outcomes of other pupils. But these findings mask a significant degree of heterogeneity along the gender dimension.

By separating our sample into boys and girls, our results also show that girls significantly benefit from interactions with very bright peers, whereas boys are negatively affected by a larger proportion of academically outstanding peers at school. We also find that the positive effect stemming from interactions with ”good” peers is more pronounced for female in the bottom part of the ability distribution. On the other hand, while not strongly significant, our results suggest that more able boys suffer from interacting with a larger fraction of outstanding schoolmates.”


3 Art Deco October 10, 2017 at 1:59 pm

It doesn’t sound like they attempted to investigate the effect of behavioral problems (v. problems in academic performance).

This particular conclusion suggests that your great grandparents’ generation knew a thing or two that’s been suppressed when they invested in schools for boys and schools for girls.


4 2nd str October 10, 2017 at 4:13 pm

Whenever you see, “while not strongly significant, our results suggest that…” you should put the paper down and spend some time reading Andrew Gelman’s blog.


5 Sam October 10, 2017 at 12:18 pm

If peer effects matter, than from a public policy point of view their ought to be an optimal distribution of high-achieving peers throughout different schools. Clustering high achieving peers into the same schools may be locally rational, but represents a form of assortative matching and “opportunity hoarding” that could worsen social mobility.


6 BDK October 10, 2017 at 12:44 pm

Or it could be that by matching high-achieving peers, you increase benefits for society overall, including by eliminating the network effects of having a bunch of high achievers interacting and working together. My intuition is that splitting up the student body of Cal Tech, Stanford, MIT, and Harvard and sending a few such students to each of the 4000 or so colleges in the US would lead to net negative effects for world wealth and well-being, including for students who otherwise would not have had the benefit of going to school with a few super high achievers. Certainly high school is a bit different, but I am not sure it is so different as to change my intuition.


7 Kevin E. October 10, 2017 at 12:56 pm

How about segregating the “bad peers” into one school? Their achievement will be low regardless, and that kids who actually want to learn can avoid them.


8 Albert October 10, 2017 at 1:15 pm

That’s how it works now, AFAIK. There’s always a new bottom 5% though.


9 Art Deco October 10, 2017 at 1:56 pm

You need to have a collecting pool for the incorrigibles.

The ones who just perform poorly in school but otherwise keep their nose clean you might pool, but only for economies of scale. If you have proper tracking, they’re not in the same classrooms with the better students.


10 Steve Sailer October 10, 2017 at 12:31 pm

My wife found a fine public middle school program for my who had been in private school K-5. So she talked the parents of his two best friends at the private school into transferring theirs sons with him to the public school, so he could take his peer group with him. That worked out well.


11 A Definite Beta Guy October 10, 2017 at 12:43 pm

“Or is peer quality actually also the best practical-for-parents measure of value-added, when all is said and done?” – Isn’t this the Occam’s Razor answer?


12 BDK October 10, 2017 at 12:52 pm

It seems so self evident to me that the rest of the post was striking. But I live in the world of law, finance and business, where your ability to work with other high-achieving peers, and your connections to other high achieving peers, may be more obviously important to long-term achievement than in STEM/academia, where test scores may correlate more strongly to success. In fact, if I am reading the abstract correctly, doesn’t it even acknowledge this point: “…and these schools generate larger improvements in short- and long-run student outcomes.”


13 IVV October 10, 2017 at 6:15 pm

There are still many benefits to be had in STEM/academia to having high-achieving peers. No one is going to be thinking about everything at a genius level. Knowing how to leverage each other’s strengths is good for everyone.


14 Highgamma October 10, 2017 at 12:45 pm

Or parents know that competition, if it has any effect on school quality, would have an effect with a lag, so they go for the effect with no lag, peer quality.


15 Bob October 10, 2017 at 12:52 pm

My anecdotal: My small neighborhood is the only middle class neighborhood in our designated Elementary School. I’ve opted to throw my kids into a spanish immersion school in a nearly 100% middle class schoolt. We made the decision because we didn’t want our kid to be one of 5-10 families of middle class college educated in the entire school. All the stats say that child outcome is based on parent education and income. I wanted my kids to have a peer group that would have the same advantage and as a result provide some competition to challenge them.

No way to prove it, but I’d be willing to wager that if the teacher from both school were given the same student groups the results would not very between schools.


16 wiki October 10, 2017 at 12:53 pm

The researchers should also acknowledge that they have not found any evidence that the next 5-10% of the worst peers matter, NOT that they have No effect. It may well be that parents are more sophisticated about ways in which lower tier peers might have negative effects. They might not be systematic or they might be low probability events, but they might well be significant (in real life terms, not in statistical significance) in measures researchers are not attuned to.

As the old saying goes, “Never marry for money. Hang around rich kids and marry for love.”


17 mpowell October 10, 2017 at 1:02 pm

This has the property of, well, of course that’s how it works. How could parents conceivably identify school quality independent of peer effects? And, of course, from a population level, this doesn’t justify a voucher system really. You still might ask, maybe there is a bottom 5% or 10% of schools that parents can identify as under performing? It might not show up in the overall statistics, but there might be some bad lemons out there that could be selected against.


18 Dan October 10, 2017 at 1:12 pm

I’m not exactly sure how “New York City’s centralized high school assignment mechanism” works, but is it possible that causality is in the other direction here? If parents rate a school highly, then the school will get to choose from a larger group of potential students, and so the school will wind up having a high-quality student body. In other words, high-achieving peers are a consequence of parents preferring the school.


19 mulp October 10, 2017 at 1:15 pm

In the week Thaler was recognized for bringing back the idea that conservatives tried to kill since the 70s, people are not economically rational decision makers, you publish with surprise a study that suggests parents are not rational decision makes when picking schools.

I grew up in the shadow of Eccles, Keynes, Galbraith making the point people are not rational economic decision makers.


20 Dick the Butcher October 10, 2017 at 1:45 pm

This Saturday, I got kicked in the teeth with a 10% increase in my school tax bill.

I guess that comment is about parents and education. The same can be said for health care. My wife is a registered nurse (retired). Our med decisions have been more rationaler than typical people’s. She keeps yelling about my cigars.

We were heavily involved in our children’s education. Parents need to push the kids to learn. Too many parents stopped in grade school, throwing up their hands that the kid just ain’t smart. I took a different tack – the kids came home with bad grades and we addressed it. I realize that most Moms and Dads work, there it is.

In our town, the public elementary schools are excellent. The middle school/high school is poor unless your kid gets in the AP classes. We had placed all in AP classes. The last we had to go to the superintendent’s office to get him placed. For high school, we were able to get admitted (and could afford) to a competitive, highly-rigorous private high school.

Do you believe that central bankers, central planners/command economy commissars, university econ. PhDs, politicians, et al are economically-rational decision-makers? Or, that their decisions, policies, programs have been beneficial for America? Things could be worse is not a rational response.

Anyhow, your geniuses are running out of “not economically-rational decision-makers'” money.


21 Art Deco October 10, 2017 at 1:53 pm

The trouble is, for many kids academic performance is going to require superlative efforts which will tax the patience of all parties involved. Youths should be tracked and directed toward activities which make optimal use of their time and effort. I suspect what Leon Podles said half-a-generation ago is correct: we are squandering a great deal of resources on half-assed liberal education for youths for whom it is unsuitable. The British used to sort their youth into three sets, one set to academic programs, one to vocational programs, and one to basic education + life skills. This is what we should be doing.


22 John Thacker October 10, 2017 at 3:24 pm

I grew up in the shadow of Eccles, Keynes, Galbraith making the point people are not rational economic decision makers.

Yes, but they also made the point that people suddenly become rational economic decision makers when in the voting booth or serving in government, which was hardly correct either. Governments of all sorts tend to exacerbate the irrational biases of their people (and their leaders) rather than restrain them.


23 Roger Sweeny October 11, 2017 at 2:15 pm

No. No. No. Eccles, Keynes, and Galbraith ARE rational, as are any of their disciples who go to work in government and make decisions for the rest of us.

(Except when Keynes said of Hayek’s Road to Serfdom, “In my opinion it is a grand book […] Morally and philosophically I find myself in agreement with virtually the whole of it: and not only in agreement with it, but in deeply moved agreement.”)


24 A.G.McDowell October 10, 2017 at 1:27 pm

I can think of two other reasons why parents might prefer absolute performance over relative improvement:

1) Parents think that relative improvement statistics are excuses.

2) Parents don’t care about academic performance so much as student safety, and are using statistics on absolute performance as an indicator of this – if all of the other students are getting high grades, they won’t be beating up my kid.


25 Art Deco October 10, 2017 at 1:49 pm

You need to do one thing the teach-every-child twerps in school administration will not do: sequester the nasty kids and turn them over to day-detention centers run by sheriff’s deputies from the county jail service.


26 P Burgos October 10, 2017 at 2:54 pm

My impression is that most school districts throughout the country do sequester the nasty kids, and it is only a few high profile school systems that do not, such as in Minneapolis and NYC. Of course, in Minneapolis, conditions got so bad that the teachers were threatening to go on strike to protest unsafe working conditions, and in NYC, children are being murdered by other children at school. So I don’t think that the social justice warriors in school administration will be able to sustain “teach every child never suspend or discipline any brown kid” policies. The coalition of parents, teachers, and conservatives is way too powerful to ignore, and way too powerful to successfully fight.


27 Art Deco October 10, 2017 at 4:16 pm

My impression is that most school districts throughout the country do sequester the nasty kids, and it is only a few high profile school systems that do not, such as in Minneapolis and NYC.



28 P Burgos October 10, 2017 at 4:40 pm

Okay, so what are some high profile incidents outside of Minneapolis and NYC illustrating that schools don’t sequester troublemakers?

29 subdee October 15, 2017 at 5:26 pm

Most districts do have alternative high schools, or other private solutions. Which students are considered the worst is always relative, though.

30 Art Deco October 10, 2017 at 1:47 pm

Before she went senile, some of Diane Ravitch reports on survey research done a generation ago suggested that given a certain baseline, parents-in-general are not dissatisfied. Professional-managerial parents are likely in select circumstances dissatisfied with content and rigor, but these are qualities distinct from pedagogic effectiveness.

One utility of private education financed by vouchers is that it injures no innocent constituencies (in addition of assisting niche constituencies). It does injure incorrigibles and their ‘parents’. It also injures those whose ideology, job opportunities, and income flows are advanced by the current system: teacher-training faculties, school apparatchiks in state departments, superintendent’s offices, and in schools themselves; marginal teachers whose career is dependent on the architecture of teacher-training and certficiation, and labor meatheads. These are people you should injure.


31 drive-by commenter October 11, 2017 at 2:30 am

>One utility of private education financed by vouchers is that it injures no innocent constituencies

Do you really believe this? Or are you really good at staying in character?


32 mike October 10, 2017 at 2:20 pm

White parents try to send their kids to the public schools that have the smallest percentage of brown students, which are usually considered the best schools.


33 Aaron Aardvark October 10, 2017 at 2:25 pm

Peer effects extend beyond education. Many of the people I hung out with in high school were not destined for Harvard (or even college for the most part – this was the late 1970s). But they were sensible teenagers who liked a bit of fun (loud music, beer, and a bit of weed) but knew where to draw the line as well as respect other people and their property.


34 Mr. Sensitive October 10, 2017 at 3:50 pm

Those days are long gone.


35 Engineer October 10, 2017 at 2:40 pm

One might conclude that the most important thing parents can do is keep their kids away from the “bad influence” kids, and that the most effective thing schools can do is expel the disrupters. The fact that “bad” schools will not do this triage is a disaster for the “could be helped” students.

After excluding the worst 5 per cent of the teachers, results seem strongly dependent on student quality, and only weakly dependent on teacher / school quality.


36 Boonton October 10, 2017 at 2:56 pm

For the purpose of modelling, the working assumption with school choice is that the school is some type of machine or treatment…students are fed into the machine and outcomes get spit out. If the machine is better, then the outcomes are better. Therefore with school choice vouchers combined with parental funds would bid up providers of better machines while harming those with crappier machines. Presumably suppliers would learn what ‘machines’ produce better outcomes and competition would improve the whole system. Sort of like how innovations by Apple, Android, Google and even side players like Amazon cause the whole cell phone market to improve in quality of features offered.

Suppose, however, schools have no impact on outcomes. Let’s say outcomes are 100% caused by parents. The impact of school choice/vouchers would have no impact on overall quality because you can’t use the vouchers to buy better parents! Instead the result of competition would be about bidding up the price of anything *but* better outcomes. So if parents happen to like their kid to go to school with more college about students than less, you’ll get clustering and competition around that but no actual improvement in anything that matters to the taxpayer…whose interest is not that any one kid get a school that’s more fun, or has a more engaging football team, but that all schools produce more output per taxpayer dollar.

I think clearly voucher advocates try to sell the first system as true….and I acknowledge the 2nd model is extreme *but* if you had to pick one extreme model as closer to the truth I suspect it would be the 2nd.


37 John Thacker October 10, 2017 at 3:21 pm

For the purposes of modeling, the “peer effects” model is different from either of the two systems that you’re positing. (Also you’re somewhat lumping together voucher systems with other sorts of choice approaches, which the original paper does not do.) The peer effects model creates some interesting results for inequality. (On a related note, another NBER paper using South Korean data showed that classrooms with more obese kids randomly assigned to the class were more likely to have other students become obese. Imagine parents also fighting to keep other people’s fat kids from being assigned to the same class as their precious kids.)

One interesting thing about these models is that people are rarely consistent between models depending on the situation. I’ve heard plenty of public school teachers and public school advocates argue in favor of the “it’s 100% caused by parents” model when it comes to opposing school choice, but the same model argues pretty heavily in favor of slashing all public school spending. If it’s 100% caused by parents, then teacher quality and school quality doesn’t matter, and so school spending (including teacher salaries) are irrelevant. There are some similar examples of inconsistency among most people.

Another interesting thing with studies of school choice programs is that a proper school choice program should show a decreasing gap between choice and not choice schools over time, as public schools improve due to competition and newer marginal private and charter schools are not as good as the original ones. Yet strangely such results are taken as opposing choice, whereas persistent gaps are the signs that either the “100% caused by parents” type theories are correct or that vouchers are not high enough to enable the worst students to actually access a good enough education, or that the number of slots is capped too low to enable the worst schools to close and students to escape.


38 Engineer October 10, 2017 at 3:40 pm

” because you can’t use the vouchers to buy better parents!”

Is this really true? I would think in a lot of cases that is exactly what you are doing, to a considerably greater extent than you are buying better facilities or teachers.

I suppose the effectiveness of this would depend on the degree to which the chosen school can also choose their students. In the strong form, you would be paying the chosen school in large part precisely to ensure better parents.


39 Bob October 10, 2017 at 11:49 pm

Ah, but those better parents might not want to be bought by you, and move to an even more expensive school. Visit St Louis, and experience white flight. You can even see it in the many private schools: Parents sort themselves not necessarily by how smart the peers are, but by all kinds of other socioeconomic queues, starting with race. You can try to get your children into a school that doesn’t match your status about as well as you can jump to the moon.

As it happens, parents are not very good at identifying actual peer achievement, or school performance. Imagine that. Next they are going to tell us that they aren’t all that good at evaluating complex tax policy proposals.


40 Boonton October 10, 2017 at 5:20 pm

I agree neither model is 100% correct but I think maybe an analogy that might be helpful is a gym with gym trainers trolling around.

In terms of getting people in shape, 80% of the result is just showing up. So you don’t need a very good trainer…even someone who just has the patience to make sure you work out on a regular basis is going to get impressive results given enough time with someone who starts out in very bad shape.

The top trainers don’t really have any other magic to offer. Given someone whose out of shape, a trainer whose worked with Olympic athletes will do almost all the stuff a regular trainer will do. The elite trainers get paid not because they have any magic to offer there but are able to get an extra 1 or 2 percent out. When dealing with professional sports, that tiny extra bit makes all the difference.

This doesn’t change the fact, though, that 80% of the result comes from the 20% ‘mediocre’ effort. To get 100 unfit people into decent shape, you are better off with more average or even below average trainers helping them than elite trainers. Putting the guy who trained 100 NFL Pros on the class isn’t going to produce that much better results.

So in terms of output I think the ‘real’ model is you need a school that’s good enough, but once you get that the value of additional output gets really expensive. Are you talking the top of the top kid who you are hoping to see score a perfect SAT and get all the Ivy League schools offering full scholarships to snag? There your school choice may make a critical difference simply because at that level even a few percent of edge is critical.

So if this is the ‘real model’ how would lots of choice play out? I think for the 10% at the bottom it might be very helpful moving from a non-functional school to a functioning one…maybe…. For the next 80% there’s no real impact in terms of actual performance….but there may be some competition for things that have nothing to do with what the taxpayer cares about (nicer looking buildings, fancy furniture, fun football teams etc.). For the top 10%, I would also suggest no change. If you, say, instituted a $10K voucher all that would happen is the elite of the elite schools would just raise their tuition by $10K per year. Why wouldn’t they? It wouldn’t change what the parents are already paying.


41 Sure October 11, 2017 at 9:37 am

Suppose we grant a model that is 100% parent effort.

I suspect vouchers or similar school choice would still have a positive effect. For instance, you can buy more time for parental investment if you can allow the parent to utilize a school that better matches their work schedule. Say the local elementary school is 20 minutes east of the house, the parent who needs to pick up the kid from extracurricular activities works 40 minutes to the west. There is a comparable elementary school 30 minutes to the west. Having choice would give the parent back 40 minutes of time every day. Additionally, they may be able to more easily manage work hours as they do not need to leave work one hour before the child is due for pickup (or more due the greater potential for traffic delays).

We see this all the time with the children of teachers, if mom works an hour away at a public school it makes sense for her to have her children educated at the school where she teaches instead of having to deal with after school care with the local public school. While that is an obvious case, there are plenty of parents whose work and domestic geography is similarly ill suited for the traditional assignment methodology that implicitly assumed that Mrs. Cleaver would be home waiting when school let out.

Likewise, a parent may become more invested in their child’s education if the school chosen for the child has some hook that engages the parent more. For example, a school with a strong athletic program may encourage dad to spend more time watching the child play and might even encourage dad to help the child with school work to remain eligible for play. Conversely a school focused on computers may help mom become more invested with her programming experience. Choosing schools with emphases that match parent emphases may well buy better parenting.

Ultimately the real question is, do we trust parents to make good decisions for their kids. There are certainly cases where parents’ jobs dictate that the simple assignment to the nearest school is inferior to free choice. There are likely cases where parental investment would increase with school curricula choice. There are also cases where parents would choose poorly. On balance is choice going to lead more effective parenting or worse? I lean towards the former.


42 Boonton October 11, 2017 at 1:29 pm

True but why is that a taxpayer concern? If the only benefit of vouchers is that parents might pull off easier commutes to work, that sounds great but how about the taxpayers who have no children? Why do they pay more (possibly having to work more hours or work at a higher paying job with a worse commute) in order to buy soccer mom and dad a better commute?


43 Sure October 11, 2017 at 10:52 pm

Likely because parents would be willing to engage in some sort of Coasian bargain to achieve these benefits. Say the student would cost X dollars to instruct in their home district. The home district should be able come up with some offer where the student takes Y dollars with them to the new district and the home district pockets X-Y. This in turn can either increase per capita students in the home district or reduce home district taxation.

Parents can then decide, how much it is worth to save time commuting or to have junior get a first class experience in whatever the parents deem appropriate.

As a further bonus, there will likely be some parents who will earn more because the logistics of being a single parent who has to make it home when school, soccer, Headstart, or what have you lets out makes it impossible to work even a little bit late (which can make it harder to get promoted). This should then translate into more income taxation and more money to spend or invest in the home district.

Regardless, it is already done in my neck of the woods for the children of public school teachers. Affording this same flexibility to other parents seems only fair. Certainly avoiding a deadknock loss should be pretty easy tax policy for everyone.


44 edgar October 10, 2017 at 3:19 pm

Do Parents Value School Effectiveness? Some may more than others. How many hits do school test score reporting web pages get? If only 5% of parents display behavior deemed to satisfy the criteria, what percent of other parents free ride off of that effort? It is the way of macro-level “scholars” to treat parents as fungible so as preserve the extant hierarchies and decision-making authorities that support them. God forbid anybody think about allowing parents any autonomy. Only the aggregate matters and any gains that autonomy may accrue to some minority of parents must be suppressed and ruthlessly denigrated.


45 Boonton October 11, 2017 at 9:26 am

Why do test scores equal school effectiveness? Imagine if you are looking to buy a house and your real estate broker refuses to take you out looking at them, refuses to even show you pictures…instead he shows you a spreadsheet of house listings with an associated ‘nice score’?

Funny observation, I’ve been going to lots of different doctors offices in the last year. I’m noticing every office seems to have a plaque or trophy that says “America’s Best Doctors” or “New Jersey Best doctors”. Either I’ve just happened to have stumbled upon several doctors who are among the ten best…..or a bunch of people have figured out a great marketing strategy is tell doctors they were voted on top of a list of elite doctors and for a reasonable payment they can display that award in a nice plaque for the patients to see.

No doubt we can create a company that will find some great metric for any private school to hawk on their admissions page.


46 edgar October 11, 2017 at 11:38 pm

Effectiveness, eschmectiveness. “We” don’t need to do bupkus. In a more laissez faire educational ecosystem, like that in communist China, individuals are free to choose based upon whatever their little hearts desire without an elite class of pontificators receiving government funding to sit in judgment.


47 Eric377 October 10, 2017 at 3:25 pm

Peer “quality control” seems to work. Not perfectly, but well enough that property values are hugely impacted by school district lines. I am pretty sure that the generous amounts spent per pupil in my current school district are overkill from the point of view of student achievement, but that spending level helps keep property values and tax bills at levels which it is a very, very hard decision for a family to live here just “because”.


48 John Thacker October 10, 2017 at 3:26 pm

Or is peer quality actually also the best practical-for-parents measure of value-added, when all is said and done?

Peer quality is certainly the most available to parents measure of value-added. There are all sorts of rankings (GreatSchools, etc) that publicize the easily available average test scores, and use that as shorthand. Better measures of value-added are not readily available. Certainly a good first step would be to try to calculate them and provide them to parents.


49 Phil S. October 10, 2017 at 8:01 pm

Indeed, value-added measures would be useful for parents, and are hard to come by.

I’m a parent of 3 kids – 2 in HS, 1 in MS. A few years back, I was concerned that the school environment (public suburban) was not ideal for the middle of the 3 kids. Tried to research this, but the academic literature is very focused on the educator/administrator side (what works/doesn’t for schools, teachers, etc.), not so much on the parent side (what qualities to look for in a school for your kid(s)).

Test scores are one measure, and I think they have some value, despite likely being vastly more influenced by what kinds of students enroll in a school, versus what the school actually teaches. But, as others have said, they’re limited at best.

In our area, the main decision point for a parent is public versus private. (You can SORT of choose between publics, but for various reasons, its an option few parents in our area choose). Private school data is even more suspect, IMO, than public. At least with publics in our area, there is a fair amount of standardized testing, the results of which can be found on a state portal.

In the end, we have kept all of our kids in publics. The results haven’t been perfect, but have been pretty good, IMO.


50 Massimo Heitor October 10, 2017 at 7:24 pm

Tyler Cowen actually misinterprets the study.

The benefits of student peer quality are already known. The problem is, from a public policy perspective, you can’t give every student better peers. If you redistrict or move students to raise the peer quality of some students, you lower the peer quality of other students.

This study was hoping to find parents selecting for school value beyond the obvious mostly zero sum peer quality effects. The study didn’t find that. If it had, that would have made a stronger case for increased school choice policy.

If school choice just allows parents to compete for better peer quality for the students, that makes educational outcomes more dependent on the actions of the parents; kids with great parents get even better educational outcomes and kids with mediocre parents get worse educational outcomes, and the education/culture gap grows. This is the opposite of what most policy types want.


51 asdf October 11, 2017 at 12:35 pm

The current method for sorting for peer effects, vastly overpaying for real estate based on the school district, is massively inefficient. Letting people select peer groups directly, rather then indirectly, would be a massive net improvement for society.

There are also people caught in situations where they can’t get a good peer group. Say the better families in a bad urban district but who aren’t rich enough to go private. Or students in the top 1% of ability where even ordinary AP tracking at a typical public school is too slow for them.


52 Massimo Heitor October 11, 2017 at 12:48 pm

Roughly 50% of students get above average peers and 50% get below average peers. It’s odd that you chose the scenario of “better families in a bad urban district but who aren’t rich enough to go private”. All families in bad districts have bad peer groups. And at least the kids with better families have better families. The kids with bad families in the bad districts get the worst odds.


53 Boonton October 11, 2017 at 2:27 pm

The peer problem is that it’s a zero sum game. People here think about choice only in terms of their choice without realizing everyone else would get the same choice. If you have the choice to move your kid to a school with peers who are higher than average, those parents who already have higher than average kids aren’t going to be thrilled that you’re trying to put your lower performing kid next to theirs. As all the sorting plays out you’re not left with any improvement. Either all the top kids gather in some schools leaving the bottom in others OR schools end up with a mix of kids because the game of chasing around the good kids never settles into equilibrium.

In contrast if all kids end up in a single large school, at least high performing kids can both be a group within that larger population AND other kids might join them.


54 C October 11, 2017 at 10:09 am

This jibes with what I’m going through right now – my daughter is moving on to middle school so I’m shopping around.

What I want in a school is:
1. kids who aren’t knuckleheads
2. parents who aren’t knuckleheads
(I’m still not sure which of the ordering of two – it should probably be reversed)

way further down in importance but not in order
3. competent teachers
4. good infrastructure i.e. building not falling apart and gee-gaws likes computers, 3D printers, interactive whiteboards

The problems I’m seeing is that 4. is trivially easy to spot. 3. you can sort of get a feeling for if you spend a bit of time investigating. But 1 and 2 are open only to anecdote.

So far all I can figure out is that private schools are probably better in it – They tend not to hesitate in asking troublemakers to leave.


55 Boonton October 11, 2017 at 1:53 pm

Just as an amusing mental exercise, pretend for a moment your goal was to find the worse possible school for your daughter. Among the schools you have to reasonably choose from in your area (I’m not talking about sending her to a madrassa in Pakistan, for example), what would you choose if your aim was to find the worst possible outcome for her.

Now after doing that suppose someone offered you a deal….you either must send her to that school OR a rich benefactor would be willing to send her to what you think is the best school in the area BUT you would be required to limit the time you spend with her each week, not help her with homework nor attend school functions with her.


56 Roger Sweeny October 11, 2017 at 11:35 am

Good schools do not make good students.
Good students make good schools.

That goes for everything from Harvard to P.D. 99.


57 Massimo Heitor October 11, 2017 at 12:50 pm

The study said that. “Parents prefer schools that enroll high-achieving peers, and these schools generate larger improvements in short- and long-run student outcomes.”


58 Roger Sweeny October 11, 2017 at 2:05 pm

I agree. I just thought I said it better 🙂


59 Boonton October 11, 2017 at 2:14 pm

Actually wouldn’t it make more sense then to give parents small vouchers that instead of being used by the school could be used by them to pay *other parents* who send their kids to their kids’ school?


60 Roger Sweeny October 11, 2017 at 2:21 pm

That would probably lead to fairly effective tracking, which is thoughtcrime in today’s education.


61 Boonton October 11, 2017 at 3:37 pm

Let’s say everyone has a voucher that’s good for some marginal sum, say $200 a year. The voucher can only be given to another kid where it is split…50% for the parents and 50% into a college fund for the kid. Parents are free to give the vouchers to anyone they wish provided they are in the same school.

How would this impact behavior? Well if parents want their kids around ‘good kids’, this would be a great reward for parents who can signal they are raising clean-cut, good kids. Maybe those parents will publish their kids’ test scores to further demonstrate that, maybe not. Doesn’t seem to matter because you end up with a market dynamic going on.

One interesting aspect would be that simple ‘choice’ wouldn’t change the actual supply of ‘good kids’ to try to get your kid to rub up against but if parents and kids had a reward for being good (or desirable for other parents to send their kids around) you could actually increase the # of positive peer groups rather than just set off a zero sum scramble for a fixed supply of them.


62 Butler T. Reynolds October 11, 2017 at 2:32 pm

Parents don’t have good information regarding government schools, though the internet might be helping a little. The best a parent can do is look on websites that publish test scores and the % of kids on free lunch.

However, too much of a “good thing” can also cause problems. I live near a hyper-desirable school zone. Similar houses are significantly more expensive than in neighboring zones. There are stories of parents moving out of the more sought-after schools because of the extreme competitive nature of the schools — both academically and socially.

A lack of markets in education produce murky signals and flawed choices.


63 Boonton October 11, 2017 at 3:40 pm

Not following this. If the prices around a ‘good school’ are going up that implies a market reaction. Are you saying the market has been fooled by a spattering of website statistics?

What about the stock market? The ‘information’ the stock market has about companies are very high level reports that companies only have to issue every quarter and press releases that are often written by the companies themselves for propaganda purposes. Yet the stock market is able in the long run to evaluate profitable companies versus poor ones.


64 Jeff October 13, 2017 at 9:39 pm

I raised three successful children and now have a bunch of delightful grandchildren. Peer groups are by far the most important influence on children once they’re out of elementary school, so it is important to try to make sure the kids they hang around with the most are not awful. I did this by encouraging them to do band or other extracurricular activities. Two of them did band, the third was a cheerleader on a competitive squad. The bad-influence kids were nowhere to be found in either of these activities.


Leave a Comment

Previous post:

Next post: