Bryan Caplan gives a postcard summary of a 1997 review of kid care quality: Within a broad range of safe environments, quality variations in child care have only small and temporary effects on most children’s development. With a few exceptions that can be explained by correlations between family and child-care characteristics, studies both in the United States and elsewhere fail to find any long-term effects.
KenF: My children didn’t ask to be born into this world, it is a burden I placed on them...
You placed on them the burden of living in one of the richest countries in the world in hands-down the richest era of the world. I doubt they're going to grow up to hate you for that. On the other hand, they'll probably resent you for handing down your negative attitude toward existence.
KenF, again: I believe, rightly or wrongly, that paying a lot buys me that, especially since it helps them gain entrance into elementary schools that offer a similar promise.
First off, one of the points here is to figure out whether that belief is right or wrong; if it's wrong, you can save a bunch of money to use on other things. Second, I've done the whole competitive-schools thing (highly-regarded 7-12-grade school, Ivy League university—which was harder to get into than if I'd gone to a worse high school), but at the elementary school level? That doesn't sound the slightest bit insane to you?
JH: Is there a bigger “I care about my kids” signal than a stay-at-home mom?
In the case of my family, it's a signal that child care in a city for two children costs more than their mom made. And she had a meaningful job with very family-friendly policies: she brought our second son to work with her for a while when he was a baby.
Tomasz: If the long-run effects of somewhat better or somewhat worse child care are unclear, that's an argument for reducing investment in child care and directing it toward something whose benefits are clear. (It's not a binary choice between having kids tutored by Nobel laureates on the one hand and sitting in a dark, damp cage for six hours a day on the other.)
Finally, the article in question is a literature survey, not a single study; this gives it significantly more weight. Even so, you should always keep in mind that tossing five heads in a row gets you past the p > 0.05 barrier for publishing in a social science journal.
You're right. All I have is the bayesian evdience that there occur to me several plausible criteria for choosing child care, and, aside from signalling, the study casts doubt on all but this one.
I couldn't access the study (I take it you can?) but I highly doubt it provides strong evidence for the claim you make. If you want to make a claim like this you should really explain it better.
Tomasz and Jonathon's points seem valid to me too. Many of your posts come off as being dogmatic in your support of signaling explanations. Of course I don't deny that signaling is very important, but I'm skeptical that it is as important as you seem to make it out to be (especially in areas where people are greatly incentivized by other factors, such as health care and child-rearing).
By drawing conclusions from a single study, you are implicitly rejecting any contradictory theories suggested by other studies. This comes off as being against rigor.
With all due respect you completely miss Tomasz's point. His critique has little to do with issuing disclaimers, but more with being rigorous in your research.
While the evidence and studies you present -- and the questions you pose in relation to them -- are almost always thought-provoking, your conclusions often dogmatically toe the signaling party-line based on relatively little evidence.
You ask in your post why it takes so long for people to accept the truth (or at least, to be made aware of the latest research). Perhaps the reason is because the truth, especially when discovered via empiricism, appears not overnight, but over time through replicated, falsifiable research.
It strikes me as perfectly rational for people to discount even the most rigorous studies before results are verified repeatedly. It's odd that someone who often strives to stake out compromise solutions (e.g. your global warming post), is also so willing to deviate from his priors dramatically in response to minimal evidence.
While I'm not as skeptical as Tomasz regarding the influence of signaling in many situations (in fact, I pretty much always buy your explanations, at least in part), I wish you more openly embraced these calls for more evidence. Backing into the signaling explanation, and largely dismissing calls for more rigor, only weakens the persuasiveness of your arguments and inhibits you from achieving the broader impact that us regular readers wish you would.
It sounds like the main, perhaps only, reason to pay a lot for child care is to keep your kids away from poor kids and their parents. Is there any evidence that would contradict this?
I seems you want more disclaimers, so that each time I make sure to mention all the ways the study could be misleading, and say there are other possible explanations. That would seem to me to get boring fast, which is why I'm against disclaimers.
Hmm, but these same people who want to signal their care for kids through expensive child care don’t attempt to signal this through mom (or dad) staying home. Is there a bigger “I care about my kids” signal than a stay-at-home mom?
I would guess that the stay-at-home mom signal is actually negative for most women. The high status people are upper and upper middle class families where both parents work professional or managerial jobs and childcare is outsourced to nannies, private schools, and mandarin classes (or karate, or clarinet) for kids.
The notion of stay-at-home parenting as high status is now almost exclusive to evangelical Christians, agrarians, and other members of the counterculture.
I see a pattern in many of your posts:* a single study fails to find significant correlation between some measure of quality and price in X (usually health care related, but not always) very often only after correcting for effects of a few variables (like amount of fun kids are having, or doctor's experience levels)* you assume study is absolutely right and correlation after taking those extra variables into account is really essentially zero* you assume everybody can easily and reliably take the same variables into account* you assume everybody knows it already with certainty* therefore you conclude everybody prefers more expensive X for signaling reasons
Now signaling might be the case, and as you're a big fan of signaling you find it the most parsimonious explanation. However, usually:* it's very far from certain that the study measured the right thing in the right way, especially when the effects are long term and not clear cut (like pretty much anything in health care and education)* even if quality might be explained by a few underlying variables, price might still be a more easily measurable proxy for quality - underlying variables and their significance are often not easily available, or difficult to interpret* even if price and quality were completely uncorrelated, and alternative quality metrics were easily available, almost always this knowledge is very far from being universally known and believed.
Now don't get me wrong, keep posting as such studies are quite interesting. But I'm very skeptical that signaling is the whole story in most such cases.
they’d rather signal how much they care about kids by believing that kid care matters.
Hmm, but these same people who want to signal their care for kids through expensive child care don't attempt to signal this through mom (or dad) staying home. Is there a bigger "I care about my kids" signal than a stay-at-home mom?
Part of the goal of raising children is to create happy, well-adjusted adults. The evidence implies that spending on child-care won't help parents achieve this.
However, part of the goal of raising children is to create happy children. "Emotionally supportive, intellectually stimulating" environments probably do contribute to this goal.
We are in cultural period where the day-to-day business of parenting is very highly esteemed, and activity very loudly signalled. No doubt it will pass.
Although, in general, I agree with the overall point, too much $$ for marginal return. I think its a bad idea to assume pre schoolers are "little adults" --- they aren't. I.e. this, is misinformed:
Adults accept “socially unsupportive, boring work environments” all the time
Pre-Schoolers are not adults. Their brains are still rapidly developing. For infant monkeys, there are several well known experimental studies showing the detriment of poor environments on long term outcomes. (e.g. infant monkey being deprived of maternal figure, developing long term social isolation). Although pre-schoolers aren't infants either --- they are much further along and probably not going to be that dramatically altered.
I think some children could benefit from more/better childcare.
You are familiar, I assume, with the Perry Preschool experiment. Heckman has done a great deal of research on the importance of early investments in human capital. For example:
I could not access the paper you link to so I cannot assess its methodology.
I agree that people will sacrifice a lot to signal that they care, or to affiliate with high status individuals and institutions. For this reason, the correlation between quality and spending may be low. I'm not sure our measures of quality are very good.
Commenting on comments:
KenF: My children didn’t ask to be born into this world, it is a burden I placed on them...
You placed on them the burden of living in one of the richest countries in the world in hands-down the richest era of the world. I doubt they're going to grow up to hate you for that. On the other hand, they'll probably resent you for handing down your negative attitude toward existence.
KenF, again: I believe, rightly or wrongly, that paying a lot buys me that, especially since it helps them gain entrance into elementary schools that offer a similar promise.
First off, one of the points here is to figure out whether that belief is right or wrong; if it's wrong, you can save a bunch of money to use on other things. Second, I've done the whole competitive-schools thing (highly-regarded 7-12-grade school, Ivy League university—which was harder to get into than if I'd gone to a worse high school), but at the elementary school level? That doesn't sound the slightest bit insane to you?
JH: Is there a bigger “I care about my kids” signal than a stay-at-home mom?
In the case of my family, it's a signal that child care in a city for two children costs more than their mom made. And she had a meaningful job with very family-friendly policies: she brought our second son to work with her for a while when he was a baby.
Tomasz: If the long-run effects of somewhat better or somewhat worse child care are unclear, that's an argument for reducing investment in child care and directing it toward something whose benefits are clear. (It's not a binary choice between having kids tutored by Nobel laureates on the one hand and sitting in a dark, damp cage for six hours a day on the other.)
Finally, the article in question is a literature survey, not a single study; this gives it significantly more weight. Even so, you should always keep in mind that tossing five heads in a row gets you past the p > 0.05 barrier for publishing in a social science journal.
You're right. All I have is the bayesian evdience that there occur to me several plausible criteria for choosing child care, and, aside from signalling, the study casts doubt on all but this one.
I couldn't access the study (I take it you can?) but I highly doubt it provides strong evidence for the claim you make. If you want to make a claim like this you should really explain it better.
Robin,
Tomasz and Jonathon's points seem valid to me too. Many of your posts come off as being dogmatic in your support of signaling explanations. Of course I don't deny that signaling is very important, but I'm skeptical that it is as important as you seem to make it out to be (especially in areas where people are greatly incentivized by other factors, such as health care and child-rearing).
By drawing conclusions from a single study, you are implicitly rejecting any contradictory theories suggested by other studies. This comes off as being against rigor.
Jonathan, I don't see how you can read what I've said as dismissing or rejecting more rigor or more evidence.
Robin,
With all due respect you completely miss Tomasz's point. His critique has little to do with issuing disclaimers, but more with being rigorous in your research.
While the evidence and studies you present -- and the questions you pose in relation to them -- are almost always thought-provoking, your conclusions often dogmatically toe the signaling party-line based on relatively little evidence.
You ask in your post why it takes so long for people to accept the truth (or at least, to be made aware of the latest research). Perhaps the reason is because the truth, especially when discovered via empiricism, appears not overnight, but over time through replicated, falsifiable research.
It strikes me as perfectly rational for people to discount even the most rigorous studies before results are verified repeatedly. It's odd that someone who often strives to stake out compromise solutions (e.g. your global warming post), is also so willing to deviate from his priors dramatically in response to minimal evidence.
While I'm not as skeptical as Tomasz regarding the influence of signaling in many situations (in fact, I pretty much always buy your explanations, at least in part), I wish you more openly embraced these calls for more evidence. Backing into the signaling explanation, and largely dismissing calls for more rigor, only weakens the persuasiveness of your arguments and inhibits you from achieving the broader impact that us regular readers wish you would.
It sounds like the main, perhaps only, reason to pay a lot for child care is to keep your kids away from poor kids and their parents. Is there any evidence that would contradict this?
I seems you want more disclaimers, so that each time I make sure to mention all the ways the study could be misleading, and say there are other possible explanations. That would seem to me to get boring fast, which is why I'm against disclaimers.
Hmm, but these same people who want to signal their care for kids through expensive child care don’t attempt to signal this through mom (or dad) staying home. Is there a bigger “I care about my kids” signal than a stay-at-home mom?
I would guess that the stay-at-home mom signal is actually negative for most women. The high status people are upper and upper middle class families where both parents work professional or managerial jobs and childcare is outsourced to nannies, private schools, and mandarin classes (or karate, or clarinet) for kids.
The notion of stay-at-home parenting as high status is now almost exclusive to evangelical Christians, agrarians, and other members of the counterculture.
I see a pattern in many of your posts:* a single study fails to find significant correlation between some measure of quality and price in X (usually health care related, but not always) very often only after correcting for effects of a few variables (like amount of fun kids are having, or doctor's experience levels)* you assume study is absolutely right and correlation after taking those extra variables into account is really essentially zero* you assume everybody can easily and reliably take the same variables into account* you assume everybody knows it already with certainty* therefore you conclude everybody prefers more expensive X for signaling reasons
Now signaling might be the case, and as you're a big fan of signaling you find it the most parsimonious explanation. However, usually:* it's very far from certain that the study measured the right thing in the right way, especially when the effects are long term and not clear cut (like pretty much anything in health care and education)* even if quality might be explained by a few underlying variables, price might still be a more easily measurable proxy for quality - underlying variables and their significance are often not easily available, or difficult to interpret* even if price and quality were completely uncorrelated, and alternative quality metrics were easily available, almost always this knowledge is very far from being universally known and believed.
Now don't get me wrong, keep posting as such studies are quite interesting. But I'm very skeptical that signaling is the whole story in most such cases.
they’d rather signal how much they care about kids by believing that kid care matters.
Hmm, but these same people who want to signal their care for kids through expensive child care don't attempt to signal this through mom (or dad) staying home. Is there a bigger "I care about my kids" signal than a stay-at-home mom?
The second link is still broken; the extra "e" came out of the word "archives" in the URL.
Part of the goal of raising children is to create happy, well-adjusted adults. The evidence implies that spending on child-care won't help parents achieve this.
However, part of the goal of raising children is to create happy children. "Emotionally supportive, intellectually stimulating" environments probably do contribute to this goal.
We are in cultural period where the day-to-day business of parenting is very highly esteemed, and activity very loudly signalled. No doubt it will pass.
Although, in general, I agree with the overall point, too much $$ for marginal return. I think its a bad idea to assume pre schoolers are "little adults" --- they aren't. I.e. this, is misinformed:
Adults accept “socially unsupportive, boring work environments” all the time
Pre-Schoolers are not adults. Their brains are still rapidly developing. For infant monkeys, there are several well known experimental studies showing the detriment of poor environments on long term outcomes. (e.g. infant monkey being deprived of maternal figure, developing long term social isolation). Although pre-schoolers aren't infants either --- they are much further along and probably not going to be that dramatically altered.
I think some children could benefit from more/better childcare.
You are familiar, I assume, with the Perry Preschool experiment. Heckman has done a great deal of research on the importance of early investments in human capital. For example:
http://papers.ssrn.com/sol3...http://www.sciencemag.org/c...
I could not access the paper you link to so I cannot assess its methodology.
I agree that people will sacrifice a lot to signal that they care, or to affiliate with high status individuals and institutions. For this reason, the correlation between quality and spending may be low. I'm not sure our measures of quality are very good.