Hacker News Reader

Comment by delecti | Hacker News Reader

Does the research take into account that charter schools are opt-in, and thus likely to be populated by children with parents more heavily involved in their upbringing? Maybe I missed it being mentioned in the article, but it seems like that'd be a much bigger factor.

If that wasn't controlled for, then it's less that the schools themselves are necessarily any better (though that could also be the case) and more that students with already better chances of doing better are being "self"-selected into a different population.

View on HN

Does it also take into account that charter schools have a tendency to boot out any low-performing or "behavioral problem" students after a certain date (usually right after state funding levels for schools are set for the next year)?

It mentions some of the randomized experiments using lottery-admissions, which seem to be consistent with the overall thesis.

The problem is that with charter school advocates, 'seems to be consistent' usually means not consistent.

In other words, did the researchers completely forget about the single most obvious confounding variable? The answer is no.

Of course it was taken into account. Why do you think researchers would miss sth so obvious?

If you click links in the article, you will see they compared to kids who failed in the lottery. And how they address other methodological concerns.

Not winning the lottery is not the same as not even participating in the lottery.

One could argue that simply participating in the lotter shows that those kids had active parents, and therefore are self-selecting as the OP stated.

"One could argue that simply participating in the lotter shows that those kids had active parents, and therefore are self-selecting as the OP stated."

Exactly the point: comparing students with equal motivation and parental involvement, who are separated only by luck, is a convenient way to control for the self-selection bias.

Your objection shouldn't be so lightly dismissed. It's possible that merely being placed in an environment populated with other students whose parents are supportive is enough to account for the gains. The students who failed to win the lottery were placed back into a general population whose influence (direct and indirect) could be the reason for their poor performance relative to the charter school kids.

I didn't reply to the others because I didn't want to get into a whole fight about it. But the OP's original message about selection-bias is still correct.

The article's original claim is that charter schools does work for disadvantage and black students.

But that is NOT what the data supports.

The data supports that it works for disadvantaged and black students who have active parents. That conclusion raises more questions than it answers.

I think your theory makes a lot of sense. Thank you for sharing that.

Your logic doesn't work. Of those people who simply participated in the lottery, the ones who won it and went to charter school did better than the ones who weren't selected and didn't. This means the groups being compared - the winners and losers - both "had active parents"; this factor does not constitute a difference between the two groups.

You're making an assumption that itself needs to be evaluated. My city (Albany, NY) was selected as an early experiment in Charter Schools in the state, as it's the state capital, and the public schools have a bad reputation.

In the media, the message is always about quality of education and high standards -- because the audience is the political class who funds the schools. But the outreach to parents (found in bus shelters, pamphlets, hip hop radio, etc) is a message that usually is around childcare and uniforms. Longer school hours + longer school year == less daycare expenses and arrangements.

In terms of outcomes, it's all over the map. Several charters have closed for poor performance, and some do well.

I'm sorry, but your comment wasn't clear. What assumption am I making? I mean, I'm sure I made several, but which do you think needs to be evaluated/supported?

>populated by children with parents more heavily involved in their upbringing

Is there evidence to back up this statement?

Simple logic would dictate that parents who opt in for better schools for their kids are more involved than those who don't opt in for better schools..

It's getting a bit out of hand where every single statement online is now getting a [Citation Needed] afterwards. There isn't a randomized controlled study for every single aspect of the universe, we'll have to use intuition occasionally.

I'm told it's called sea-lioning. http://wondermark.com/1k62/

Statements of preference are different than assertions about the world. The comic is funny, but it does have an antiphilosophical spirit, especially if you think it relates to the present conversation.

I'm sorry. I care. (sorry (sorry)) ((((((sorry))))))

That's pretty good, I may start using that.

I now have a favourite comic :) Thanks!

>It's getting a bit out of hand where every single statement online is now getting a [Citation Needed] afterwards

I wasn't asking if there was evidence just to be annoying, I was asking because the statement was exactly opposite of my expectation and I wanted to know if there was evidence to back it up.

My personal experience with charter schools is that it's where many parents "dump" kids that were either kicked or failed out of public school. This comes from many conversations I've had with a high-school teacher I've become good friends at that switched from teaching public school to charter school ~5 years ago. I admit my perspective is very narrow but was hoping there was actual data to support the chart-school-parents-more-involved statement.

I see two assumptions I made with that.

First, that parents more heavily involved are more likely to opt-in to charter schools (or similarly though not equivalently, that parents who choose to opt-in are likely to have been more heavily involved). I consider it to be largely self-evident that parents who are not involved are very unlikely to opt-in, and while the converse isn't necessarily true, I think that fact alone would significantly sway the overall correlation of the "non opt-in" population.

And second, that being more heavily involved leads to better outcomes for children, which I feel is self-evident, and is also highly supported by a quick google search for 'involved parents better students'.

Reply on news.ycombinator.com