Hacker News new | past | comments | ask | show | jobs | submit login
How Does Your Kindergarten Classroom Affect Your Earnings? (docs.google.com)
78 points by luu on Dec 29, 2016 | hide | past | favorite | 41 comments



> This article has shown that the classroom environments that raise test scores also improve long-term outcomes. Students who were randomly assigned to higher quality classrooms in grades K–3 earn more, are more likely to attend college, save more for retirement, and live in better neighborhoods. Yet the same students do not do much better on standardized tests in later grades.

> Yet the same students do not do much better on standardized tests in later grades. Researchers who had examined only the impacts of STAR on test scores would have incorrectly concluded that early childhood education does not have long-lasting impacts.

This seems to pretty definitively end the argument I'm having with my wife in her favor; my thought was that early school education probably doesn't matter too much, since they're learning very basic things that we can teach them at home anyway.


If you're going to be actively involved in your child's education, it may not. This was a randomized test: it's not clear that means good early education is additive versus binary. For instance, it may be that for a significant number of students, the only education they receive is at school. And so for them a great teacher sets them up with a great foundation. But if you already set that great foundation up at home, it's not clear that the teacher will significantly affect improve upon it. Just that for enough students for it to show a significant correlation, the great early educator mattered.


I wish they could have controlled for parental involvement, but that's probably very hard to even measure.


Right, speaking for myself, my parents were REALLY involved (they were both teachers themselves), so even though I attended a lackluster public high school in rural Pennsylvania, I ended up doing pretty well, going onto a top 20 university and becoming a software engineer in Silicon Valley. I also loved learning and computers though, so I learned a lot on my own just by tinkering in my spare time for fun.


Hm, I'm sure there are good proxies, like the number/length of parent-initiated phone calls/emails to the school/teachers.


These are good proxies for parental involvement with school, but as Mark Twain famously suggested, school != education.


Which is clever, but all that matters for validity of the metric here, is whether a parent with a child in school, would be willing to make all these calls to the school, but otherwise be an uninvolved parent. It doesn't seem likely IMHO.


Helicopter parents can be highly involved with school but not effective in supporting their child's education. I know far too many parents who do their children's homework for them, as an example.

There are also many parents actively engaged in their child's education who don't feel the need to be in constant contact with a teacher.


The issue is whether "being involved with the child's education" has predictive power over long-term outcomes that explains away the influence of the school itself. While it may weaken the correlation for the different involvement types to be more vs less helpful, I was only speaking to the issue of whether your could measure involvement at all.


And my statement was that involvement with school ignores other types of involvement in education.


Randomized assignment to the classrooms is controlling for parental involvement, so long as the quality of the classroom is not correlated with parental involvement.


The grandparent comment explains how it still doesn't tell whether the parental involvement matters or not.

The experiment results show only that on average the kids perform better in better classrooms. But it's still possible that kids with better parental involvement have no benefit of it, and the statistical difference happens due to better results of just the kids with worse involvement.

In other words. Yes, the article controls for the fact that better parents can send kids to better kindergartens. It doesn't control for the fact that kids of better parents benefit from the better kindergartens.


It's possible that adding an interaction term for classroom quality and involved parent would show that there's zero marginal benefit to a good classroom if you're an involved parent. That result seems orthogonal to their research goal, which was estimating the mean impact of a quality classroom for the "average" child.

The phrasing, "they should control for" has a specific statistical meaning, criticizing the model as having an omitted variable problem -- that they left out a regressor that is correlated with both the dependant variable and at least one of the other regressors. That is a different criticism than suggesting an interaction term.


Uhh, that's not entirely clear. There's two causal explanations for why students assigned to higher quality classrooms do better: either good environments cause beneficial outcomes, or bad environments cause negative outcomes. Like, the important factor may be something along the lines of "don't send your kid to what's best describe as a dystopian hellscape".


Since they are defining quality classrooms after the fact, based on how well students did, the third causal explanation might be students who do well in K are the kinds of students who will succeed in life regardless. Though the experience level of teacher and size of class appear to have an impact, many other studies show a poor relation to class size and outcomes.


Yes, the big missing options were no kindergarten, and kindergarten without formal teaching. There was a recent article in the Economist that said that almost all of countries with the highest PISA scores don't start formal teaching until children are aged around 7.


What about the glaring third option, "kids who go to nice kindergartens are probably well off and hence well positioned to succeed in life"?


Yeah, that's definitely a thing as well. I wasn't terribly precise in how I worded things. I was getting at something like "given that going to a nice kindergarten causes better life outcomes than going to a shitty kindergarten, there's two mechanisms that explain that causal link."


That's not a causal explanation, in this context. ThrustVectoring specifically limited the comment to causal explanations: why do better kindergarten classes cause better life outcomes.


Common cause is a causal explanation.


What's the difference between the two? Either way, a better environment leads to a better outcome.


They're the same if you're wondering which school to send your child, but different if you're deciding whether to send your child to school.


Oh, I see what you mean.


Super hard to prove either way. The best I could do to rationalize why I believe early ed matters is to see a 'good' classroom versus a 'bad' classroom.

Seeing good teachers who are able to instill critical thinking into kids who were learning language and math for the first time had me stunned (in a good way).


>definitively end the argument

Based on one study that has never been replicated ?

> small (S) with 13-17 pupils

Would a class size of < 10 pupils mean even more success in life ? Would your 1:1 attention at home mean even better outcomes ?


> Would your 1:1 attention at home mean even better outcomes ?

Almost certainly; I'm not just gonna stop interacting with my child if she gets into a better school :P


Fancy seeing you here Pavel!



Also, middle-/upper-oriented people are more likely peer-pressured to placing in the "best" kindergartens, whereas lower-oriented people may skip it (much like skipping parental-school involvement). And, I suspect both nature and nurture aspects are self-amplifying: smarter/more competitive kids may make more money and attempt to seek similar mates to have improved offspring and try really hard to inculcate successful attitudes. Plus, having more money helps reinforce getting into selective kindergartens and making more childhood friends of future powerful people, leading to better social access to higher quality mates (better genes), more money and hopefully better lifestyle-choices.


Right, I'd really like to know about the "random" process that assigned students to classrooms. I know that when I was in elementary school the assignments were theoretically random, but if parents really wanted to influence them, they could. Since this study gets a lot of its results from clustering, they might just be finding that certain classes were perceived to be better (right or wrong) and so involved parents influenced the school to get their kids in it. Now it really is going to be "better" by the metric of those kids' later success.


I am wary that these nature versus nurture studies will always find some way in which purchasing or investing in additional or more expensive schooling is key. That is until such time as genetic engineering become consumer available, then I imagine a slew of studies proving that people with x heritable trait succeed. Of course, I'm a father with a 1 and a half year old already in Montesory school because despite my wariness that it's all just marketing pretending at science, I can't not want to give me daughter every advantage I can. They've got me over a barrel!


Not sure why we even have money to spend on these kinds of research. These trials are always random and inconclusive due to the nature of sample sizes.


The ultimate purpose is to improve education until no body ever again spews this bullshit again.

The sample size here was 22,000. Is that enough?

(the answer, by the way, is "It depends on the effect size you want to measure". I can do a study with n=2 that is pretty convincing. Question: "are these pills actually cyanide?". Study design: "I give one pill each to two healthy undergrads, after they tried to appear smart by saying 'correlation does not prove causation'". Observation: "The ethics board should definitely not have approved this!" Discussion: "Background death rate from non-violent causes for 22 year old males in the US is 0.01096/100000/day. Pill is poisonous with p<0.000000000000012")


You're going to have a hell of a time getting participates to fill out follow up surveys for a long term analysis.


Get them to 'donate their body to science' and you can keep observing them forever, or until the cryo machines stop working.


Well the sample size should probably be mentioned as 22k / # students per class (~700 classes?), since the students results are purported to be correlated. The statistics is meagre as well. Regressing percentages, counts and dollars all non-normalised is bad form at the least.

That said, I like the study. True random design takes guts and effort and gives data to be useful for years. We're still discussing RAND... On an important subject as well.

Lot more to be found [1].

Thing that wonders me how state vs national differences are corrected for. On the one hand they shouldn't matter, since it is a random design. But the differences with national averages were huge. And far larger than the measured impact of the intervention.

[1] https://www.povertyactionlab.org/evaluation/effects-kinderga...


You can't easily get money for a large study without some suggestive data from a small study. If you complain about the inconclusivity of research as a reason not to spend money on it, you're likely to keep getting worse or no research, at which point you'll be making policy on the basis of hunches rather than evidence.


They are cheaper to do than the robust ones.

At some level, these do provide data to look at. Right?


Ok. I will explain why these kinds of studies create more confusion. (Also, ask your nearest teacher if you want a different perspective)

From Adam Grant's Raising a Moral Child: http://mobile.nytimes.com/2014/04/12/opinion/sunday/raising-...

"A couple of weeks later, when faced with more opportunities to give and share, the children were much more generous after their character had been praised than after their actions had been. Praising their character helped them internalize it as part of their identities. The children learned who they were from observing their own actions: I am a helpful person. This dovetails with new research led by the psychologist Christopher J. Bryan, who finds that for moral behaviors, nouns work better than verbs. To get 3- to 6-year-olds to help with a task, rather than inviting them “to help,” it was 22 to 29 percent more effective to encourage them to “be a helper.” Cheating was cut in half when instead of, “Please don’t cheat,” participants were told, “Please don’t be a cheater.” When our actions become a reflection of our character, we lean more heavily toward the moral and generous choices. Over time it can become part of us."

If you are familiar with Carol Dweck’s work on praising action instead of intelligence, what does that tell you relative to Adam Grant?

I respect both Adam Grant and Carol Dweck, but what do common people get from reading both their work that is well backed up by research? Confusion


Dweck's work is about praising correct/good actions, not action in general. She doesn't advocate praising effort that has zero/no chance of success.

The key conclusion of her work is consistent with the passage you quoted. Instead of praising a child for trying in systematic ways, praise the child for being one of those 'organised' kids who tries in systematic ways rather than [insert ineffective way that kid could work hard with low likelihood of success]


I would think that is more problems from the reporting of these studies. There is certainly an argument that making the studies allows the shoddy reporting.

However, my hope would be we could do both. Run as many cheap studies as we can. Just honestly report on the limited applicability. And keep adding to a corpus of data. Patterns may emerge.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: