# Correlation is not causation! **URL:** \ **Category:** In My Humble Opinion\ **Created:** [March 28, 2014, 5:17am UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741 "2014-03-28T05:17:16Z")\ **Posts on this page:** 20\ **Page:** 2
**Author:** ![Buck\_Godot](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/buck_godot/32/6573_2.png) [@Buck\_Godot](https://boards.straightdope.com/u/Buck_Godot)\ **Post date:** [March 28, 2014, 6:37pm UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/21 "2014-03-28T18:37:56Z")
On the other hand, some people also misuse this truism. I was reading an article about the well designed study that found that all of the brains of autistic children they studied had the same abnormal prenatal brain growth. Reading the comments some anti-vaxer dismissed the study by saying “correlation isn’t causation”, and got many pats on the back from other commenters. But unless they can give some reason that parents of babies with this form of abnormal pre-natal brain growth are more likely to give their kids vaccines leading to their autism, I’m going to go out on a limb and say that in this case the correlation is pretty clearly causative. ---
**Author:** ![Inner\_Stickler](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/inner_stickler/32/318_2.png) [@Inner\_Stickler](https://boards.straightdope.com/u/Inner_Stickler)\ **Post date:** [March 28, 2014, 6:47pm UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/22 "2014-03-28T18:47:50Z")
> [@ThisUsernameIsForbidden](#): > > And what’s up with words like “won’t”? What is that supposed to be a contraction of? Will and not of course, but why isn’t is willn’t? Or, maybe it was but we changed it. Will, like so many words was once spelled a variety of ways. Will, woll, wul, wole; these were all options. When people made woll negative, they liked to spell it wonnot which contracts to won’t. We just happened to mix and match a bit when we started to standardize. ---
**Author:** ![Wednesday\_Evening](https://avatars.discourse-cdn.com/v4/letter/w/b2d939/32.png) [@Wednesday\_Evening](https://boards.straightdope.com/u/Wednesday_Evening)\ **Post date:** [March 28, 2014, 7:17pm UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/23 "2014-03-28T19:17:42Z")
> [@SlackerInc](#): > > Then there was the especially silly one mentioned in a thread here recently: that when a household has more books in it, the children living in it do better in school. Never mind that bookish parents are likely to have, you know, _books_ as well as being likely to have bookish children; there seem to be some actual serious people out there who imagine that delivering a big pile of books to every struggling student’s home will turn them around. Here is the [abstract](http://www.sciencedirect.com/science/article/pii/S0276562410000090)from the study (bolding mine): > [@](#): > > Children growing up in homes with many books get 3 years more schooling than children from bookless homes,\*\* independent of their parents’ education, occupation, and class\*\*. This is as great an advantage as having university educated rather than unschooled parents, and twice the advantage of having a professional rather than an unskilled father. It holds equally in rich nations and in poor; in the past and in the present; under Communism, capitalism, and Apartheid; and most strongly in China. Data are from representative national samples in 27 nations, with over 70,000 cases, analyzed using multi-level linear and probit models with multiple imputation of missing data. So next time you criticize a study, you may want to actually check it out first instead of making erroneous assumptions about it. ---
**Author:** ![Buck\_Godot](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/buck_godot/32/6573_2.png) [@Buck\_Godot](https://boards.straightdope.com/u/Buck_Godot)\ **Post date:** [March 28, 2014, 7:44pm UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/24 "2014-03-28T19:44:29Z")
> [@Wednesday\_Evening](#): > > Here is the [abstract](http://www.sciencedirect.com/science/article/pii/S0276562410000090)from the study (bolding mine): > > So next time you criticize a study, you may want to actually check it out first instead of making erroneous assumptions about it. Even so its not clear that it is the presence of books that made the difference rather than being raised in the sort of family that likes, and can afford books. It may be that highly educated parents without books, pursued their education for economic reasons but still did not value learning for its own sake as much as those parents who liked books. Also speaking as a statistician, I know that they may say their models “accounted for” various other co-variates, but in real life non-linear situations it is nearly impossible to fully scrub a covariate’s effect from a given set of data. In order to really test the hypothesis that it was the books you will need to take a largely bookless population, and give half of them books, and then see if they do markedly better than the other half. ---
**Author:** ![Voyager](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/voyager/32/133_2.png) [@Voyager](https://boards.straightdope.com/u/Voyager)\ **Post date:** [March 28, 2014, 7:50pm UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/25 "2014-03-28T19:50:34Z")
> [@Wednesday\_Evening](#): > > Here is the [abstract](http://www.sciencedirect.com/science/article/pii/S0276562410000090)from the study (bolding mine): > > So next time you criticize a study, you may want to actually check it out first instead of making erroneous assumptions about it. Did they take into account the use of libraries? When I grew up we had some book, not tons, but we had a library which my mother took me to all the time. Her love of books was far more important than the actual presence of owned books - though for families who can afford books, I’d say the correlation makes sense. I also wonder about the impact of e-books. My younger daughter is in Germany, where it is hard to buy lots of books in English, let alone worry about bringing them back. She does have lots of books on her Kindle, though. I’d guess that books on Kindle should correlate also, if anyone wants to apply for a grant for a follow up study. ---
**Author:** ![Zebra](https://avatars.discourse-cdn.com/v4/letter/z/c0e974/32.png) [@Zebra](https://boards.straightdope.com/u/Zebra)\ **Post date:** [March 28, 2014, 8:25pm UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/26 "2014-03-28T20:25:01Z")
[If you want to be humerous about it.](https://www.youtube.com/watch?v=IZeWPScnolo&feature=kp) ---
**Author:** ![Wednesday\_Evening](https://avatars.discourse-cdn.com/v4/letter/w/b2d939/32.png) [@Wednesday\_Evening](https://boards.straightdope.com/u/Wednesday_Evening)\ **Post date:** [March 28, 2014, 8:37pm UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/27 "2014-03-28T20:37:53Z")
> [@Buck\_Godot](#): > > Even so its not clear that it is the presence of books that made the difference rather than being raised in the sort of family that likes, and can afford books. It may be that highly educated parents without books, pursued their education for economic reasons but still did not value learning for its own sake as much as those parents who liked books. > > Also speaking as a statistician, I know that they may say their models “accounted for” various other co-variates, but in real life non-linear situations it is nearly impossible to fully scrub a covariate’s effect from a given set of data. > > In order to really test the hypothesis that it was the books you will need to take a largely bookless population, and give half of them books, and then see if they do markedly better than the other half. > [@Voyager](#): > > Did they take into account the use of libraries? Just to be clear, I am not trying to convince everyone that the study should be taken completely at face value. I tend to be as skeptical of studies as the next Doper. I’m just trying to say that so far, the abstract itself seems to refute **SlackerInc** ’s statement. If anyone wants to download the paper, look at the analysis, and report back to the Dope, you would be doing the rest of us a service. Until then, it’s all conjecture. ---
**Author:** ![monstro](https://avatars.discourse-cdn.com/v4/letter/m/ba9def/32.png) [@monstro](https://boards.straightdope.com/u/monstro)\ **Post date:** [March 28, 2014, 10:39pm UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/28 "2014-03-28T22:39:27Z")
If critically reviewing scientific research was easy, we’d all have Ph.D’s. The truth is that it is a skill that must be learned. I had a rigorous science undergrad curriculum, and I didn’t learn it. I was well into my post-doc before I began to readily recognize the common pitfalls scientists make in drawing conclusions. “Correlation doesn’t equal causation” actually isn’t one of them, IMHO. That’s pretty low hanging fruit. It’s pretty damn hard to get published in a respectable journal if you haven’t accounted for the most obvious co-variates, like the ones the OP or any other layperson would notice. A researcher may not control for them very well (e.g., they may choose the wrong kind of multivariate analysis or model selection method). But at the very least they will acknowledge the weaknesses of their study and insert the appropriate caveats and qualifiers in the discussion. I have never come across a reputable researcher (at least in my field of study) that doesn’t do this. If they don’t do it voluntarily, one of their reviewers will force them to. No, the problem is that laypeople think the science they read about in The Atlantic Monthly is the same kind of science that scientists are debating behind the scenes. Most scientists don’t sell their work as “surprising new discovery” or “earth-shattering revelation”. That’s what the media does, because the truth is much more BORING. If you actually have a chance to sit down with a researcher and ask them what their research really means, they tend to be quite non-committal, almost overly so. For someone like me who relies on scientists to help shape regulatory policy, it can actually be frustrating sometimes. You can blame the media if you want, but in my opinion, that also misses the point. It’s not the media’s job to critique the scientific studies they’re reporting on. I wouldn’t trust a science writer, even a competent one, to do a good job of critically reviewing a study anyway. No, the job of the media is to draw attention to exciting research. If the reader is truly interested, they should go to the original white paper and dive into the weeds themselves. And if they are still critical of the findings, then they should contact the primary author (contact info is always provided) and criticize away. But what they shouldn’t do is just automatically assume the research was poorly conducted based on a brief magazine article. There’s healthy skepticism. Then there’s just anti-intellectualism masquerading as skepticism. ---
**Author:** ![Voyager](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/voyager/32/133_2.png) [@Voyager](https://boards.straightdope.com/u/Voyager)\ **Post date:** [March 28, 2014, 10:50pm UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/29 "2014-03-28T22:50:49Z")
> [@Wednesday\_Evening](#): > > Just to be clear, I am not trying to convince everyone that the study should be taken completely at face value. I tend to be as skeptical of studies as the next Doper. I’m just trying to say that so far, the abstract itself seems to refute **SlackerInc** ’s statement. If anyone wants to download the paper, look at the analysis, and report back to the Dope, you would be doing the rest of us a service. Until then, it’s all conjecture. Okay, I was under the impression that you had read the paper. My question was purely out of curiosity. And because I’m very good at coming up with ideas for further research. When I used to sit in at a grad seminar at Stanford the grad students would quiver in their boots as I threw out ideas for five more years of research But looking at library use would be an interesting thing to do. ---
**Author:** ![Voyager](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/voyager/32/133_2.png) [@Voyager](https://boards.straightdope.com/u/Voyager)\ **Post date:** [March 28, 2014, 10:53pm UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/30 "2014-03-28T22:53:35Z")
> [@monstro](#): > > If critically reviewing scientific research was easy, we’d all have Ph.D’s. > > The truth is that it is a skill that must be learned. I had a rigorous science undergrad curriculum, and I didn’t learn it. I was well into my post-doc before I began to readily recognize the common pitfalls scientists make in drawing conclusions. If you’ve seen as many paper reviews as I have, you’ll know that it is a skill that not even many people with Ph.Ds have. ---
**Author:** ![monstro](https://avatars.discourse-cdn.com/v4/letter/m/ba9def/32.png) [@monstro](https://boards.straightdope.com/u/monstro)\ **Post date:** [March 28, 2014, 11:07pm UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/31 "2014-03-28T23:07:50Z")
> [@Voyager](#): > > If you’ve seen as many paper reviews as I have, you’ll know that it is a skill that not even many people with Ph.Ds have. Just to be clear, I don’t even think I’m that good at it myself. I usually have to read a paper multiple times before I even feel comfortable summarizing the researcher’s findings, let alone critiquing their methods or conclusions. And I’m not saying that one must have a Ph.D to be able to tear into a paper. But there is a right way to go about it. People should definitely ask questions like the OP’s. But if they want those questions answered, they’ve got to go directly to the source and not confuse a fluffy magazine piece for actual science, like the OP seems to be doing. ---
**Author:** ![zut](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/zut/32/2875_2.png) [@zut](https://boards.straightdope.com/u/zut)\ **Post date:** [March 28, 2014, 11:57pm UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/32 "2014-03-28T23:57:50Z")
> [@Wednesday\_Evening](#): > > Just to be clear, I am not trying to convince everyone that the study should be taken completely at face value. I tend to be as skeptical of studies as the next Doper. I’m just trying to say that so far, the abstract itself seems to refute **SlackerInc** ’s statement. If anyone wants to download the paper, look at the analysis, and report back to the Dope, you would be doing the rest of us a service. Until then, it’s all conjecture. Jesus Christ. Talk about being invisible. ---
**Author:** ![SlackerInc](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/slackerinc/32/275_2.png) [@SlackerInc](https://boards.straightdope.com/u/SlackerInc)\ **Post date:** [March 29, 2014, 12:01am UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/33 "2014-03-29T00:01:51Z")
> [@Wednesday\_Evening](#): > > So next time you criticize a study, you may want to actually check it out first instead of making erroneous assumptions about it. How about next time you criticise a comment, you actually read it closely? I did not refer to education, occupation, or class. I said bookish parents tend to have bookish children. > [@Voyager](#): > > It seems to me that a lot of this kind of reaction is from those who don’t understand research at all, and are sure that the pointy headed academics don’t have enough common sense to understand that correlation is not causation, unlike the good old Americans in the media. Say what?!? I targeted the media primarily, and only mentioned academics themselves as an afterthought. > [@](#): > > Sometimes the mechanism might be hard to find - doesn’t mean it isn’t there. But it also doesn’t mean those who make policy, and those of us who vote for the people who make policy (or appoint policymakers), should not expect to have extraordinary claims proven to us beyond reasonable doubt. (And BTW: “inhaling smoke deep into your lungs all day is bad for your health” is not an extraordinary claim; “getting involved in your kids’ education will just make them worse off” **is.)** Keep in mind that this is, however inconvenient it can often be, a representative democracy–not an oligarchy run like the old Chinese bureaucracy, by members of the academy. Academics (at least those oriented around public policy rather than private industry) depend on the public (at least the reasonably well-informed subset of the public who follow these issues but do not themselves have doctorates) for funding in many cases, and in any event must convince the public and policymakers to make changes in policy, not just a few members of their own ranks. > [@zut](#): > > So the logical string goes like: owning books is _correlated_ with promoting a “scholarly culture” in the home, which _causes_ increased educational achievement. One could perhaps quibble with that logic, but it’s not at all ridiculous, and the authors give a well-reasoned and well-cited background argument (at least it looks like that to me as a layperson). Moreover, it leads to a useful conclusion that might not have been apparent without doing the study: parents who are undereducated and poor (which are normally indicators of lower educational attainment by their children) can still hope to stimulate their child educationally through fostering a family scholarly culture. That takes more than just dumping a pile of books into the home, but it’s something that, with willing parents seems obtainable. Sure, if you (a) think genetics plays no role (something Steven Pinker has pointed out is an all-too-common view among social scientists, while there is no one on his side of the debate that similarly discounts the importance of environment); and (b) you think being “willing” can erase a parent’s lifetime (or even generations of ancestors) of having not accumulated social capital. Count me as extremely dubious. > [@Buck\_Godot](#): > > But unless they can give some reason that parents of babies with this form of abnormal pre-natal brain growth are more likely to give their kids vaccines leading to their autism, I’m going to go out on a limb and say that in this case the correlation is pretty clearly causative. True! Same holds with things like nutrition and smoking. The cases I’m highlighting are those in which (I’d argue) the most logical, commonsense explanation would be different from what is being promoted. > [@Buck\_Godot](#): > > In order to really test the hypothesis that it was the books you will need to take a largely bookless population, and give half of them books, and then see if they do markedly better than the other half. Right. And one wonders as well whether it can really be a random assortment, like that Frank Sinatra’s character bizarrely owned in the cinematic masterpiece _The Manchurian Candidate._ > [@monstro](#): > > No, the problem is that laypeople think the science they read about in The Atlantic Monthly is the same kind of science that scientists are debating behind the scenes. > > […]You can blame the media if you want, but in my opinion, that also misses the point. It’s not the media’s job to critique the scientific studies they’re reporting on. I wouldn’t trust a science writer, even a competent one, to do a good job of critically reviewing a study anyway. No, the job of the media is to draw attention to exciting research. If the reader is truly interested, they should go to the original white paper and dive into the weeds themselves. And if they are still critical of the findings, then they should contact the primary author (contact info is always provided) and criticize away. This is ridiculous–you are the one missing the point here. If we are talking about the frontiers of cosmology or the taxonomy of Indonesian beetles, then sure. But this is social science research into educational policy, whose aim is to inform and influence that policy. It’s not just to quietly develop a scientific picture of the world “behind the scenes”. Magazines like _The Atlantic_ are not as erudite as academic journals hardly anyone reads. But within the body politic, they \*are \*elite. The hoi polloi may read celebrity gossip magazines or maybe even TIME; _The Atlantic_ is a bit more of a select, highbrow audience. And those are the people who ultimately are going to make the greatest impact on whether, and how, educational policy is going to change. To expect such people to go and “dive into the weeds” for every research result that comes down the pike, while they continue to work at their own occupations and raise their families, is just beyond reason. It is reminiscent of the insanely complex Jim Crow-era literacy tests, which guaranteed the right of blacks to vote in theory but scuttled them in practice. And even leaving the political policymaking aspect aside: this writer is flat out telling parents, in an influential publication (as well as on podcasts and so on), “don’t help your kids with homework or get involved in their schools: you’ll just make everything worse”. So even if individual parents just take this advice in their childrearing, it could destroy many children’s lives. But I’m not supposed to be critical of the media?!? > [@](#): > > There’s healthy skepticism. Then there’s just anti-intellectualism masquerading as skepticism. Are you calling me an anti-intellectual? Hah, that’s a new one! I’ve been accused of being an elitist, a moniker I don’t entirely disavow; but never an anti-intellectual. ---
**Author:** ![monstro](https://avatars.discourse-cdn.com/v4/letter/m/ba9def/32.png) [@monstro](https://boards.straightdope.com/u/monstro)\ **Post date:** [March 29, 2014, 1:01am UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/34 "2014-03-29T01:01:31Z")
> [@SlackerInc](#): > > This is ridiculous–you are the one missing the point here. If we are talking about the frontiers of cosmology or the taxonomy of Indonesian beetles, then sure. But this is social science research into educational policy, whose aim is to inform and influence that policy. It’s not just to quietly develop a scientific picture of the world “behind the scenes”. Magazines like _The Atlantic_ are not as erudite as academic journals hardly anyone reads. But within the body politic, they \*are \*elite. The hoi polloi may read celebrity gossip magazines or maybe even TIME; _The Atlantic_ is a bit more of a select, highbrow audience. And those are the people who ultimately are going to make the greatest impact on whether, and how, educational policy is going to change. Policy makers are not dictated by _The Atlantic_. They either have a staff of professionals whose job it is to know the research, or they hire consultants who do. At least in government, almost every policy can be traced back to technical document(s), which are drafted by scientific and technical committees, who are comprised of researchers or those who are heavily steeped in research. Those technical documents do not cite publications like _The Atlantic_. They cite journal articles. What individual people choose to do with the information they pick up from the environment is not the same as institutional policy. > [@](#): > > To expect such people to go and “dive into the weeds” for every research result that comes down the pike, while they continue to work at their own occupations and raise their families, is just beyond reason. It is reminiscent of the insanely complex Jim Crow-era literacy tests, which guaranteed the right of blacks to vote in theory but scuttled them in practice. Um…that comparison is just so over-the-top crazy that I’m really not taking you seriously right now. Most people reading a magazine article aren’t going to remember it an hour later. Maybe they’ll mention it in passing at the water cooler the next day if something triggers their memory. And that’s it. For people who are actually interested in the content, whether for professional reasons or because of a personal passion, they would do well to follow-up on their own. They will probably go to the source or to a publication that treats the study more critically. As they should. > [@](#): > > And even leaving the political policymaking aspect aside: this writer is flat out telling parents, in an influential publication (as well as on podcasts and so on), “don’t help your kids with homework or get involved in their schools: you’ll just make everything worse”. So even if individual parents just take this advice in their childrearing, it could destroy many children’s lives. But I’m not supposed to be critical of the media?!? I never said you can’t be critical of the media. I haven’t read the article or listened to the podcasts, but if she is making these kinds of statements, without the appropriate qualifiers or hedging, then I agree with you that she’s being irresponsible. I just think the media has a different standard than science does. I’m not going to expect a science reporter to ask hard-hitting questions UNLESS the scientific discourse has reached such a high-pitch to warrant that kind of attention. > [@](#): > > Are you calling me an anti-intellectual? Hah, that’s a new one! I’ve been accused of being an elitist, a moniker I don’t entirely disavow; but never an anti-intellectual. If you want to infer from my post that I’m calling you an anti-intellectual, go right ahead. But I didn’t call you anything. I don’t know you or your posts well enough to call you anything. Yet. ---
**Author:** ![Voyager](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/voyager/32/133_2.png) [@Voyager](https://boards.straightdope.com/u/Voyager)\ **Post date:** [March 29, 2014, 1:11am UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/35 "2014-03-29T01:11:27Z")
> [@SlackerInc](#): > > Say what?!? I targeted the media primarily, and only mentioned academics themselves as an afterthought. But from what I’ve seen, the people who react to these kind of studies act like the media is accurately reflecting the science - and it was those people who I was referring to. You’ve seen it - the “big time scientist spends money proving something we normal people already know.” Or, like in this case, some think the scientists are too dumb to realize that correlation is not causation. The media only has space to cover the big story, not the provisos and quid pro quos. > [@](#): > > But it also doesn’t mean those who make policy, and those of us who vote for the people who make policy (or appoint policymakers), should not expect to have extraordinary claims proven to us beyond reasonable doubt. (And BTW: “inhaling smoke deep into your lungs all day is bad for your health” is not an extraordinary claim; “getting involved in your kids’ education will just make them worse off” **is.)** Keep in mind that this is, however inconvenient it can often be, a representative democracy–not an oligarchy run like the old Chinese bureaucracy, by members of the academy. Academics (at least those oriented around public policy rather than private industry) depend on the public (at least the reasonably well-informed subset of the public who follow these issues but do not themselves have doctorates) for funding in many cases, and in any event must convince the public and policymakers to make changes in policy, not just a few members of their own ranks. Politicians - either out of ideology or the desire to grandstand and show the voters that they are just like them or out of ignorance are prime culprits - Proxmire being the most famous example. > [@](#): > > This is ridiculous–you are the one missing the point here. If we are talking about the frontiers of cosmology or the taxonomy of Indonesian beetles, then sure. But this is social science research into educational policy, whose aim is to inform and influence that policy. It’s not just to quietly develop a scientific picture of the world “behind the scenes”. Magazines like _The Atlantic_ are not as erudite as academic journals hardly anyone reads. But within the body politic, they \*are \*elite. The hoi polloi may read celebrity gossip magazines or maybe even TIME; _The Atlantic_ is a bit more of a select, highbrow audience. And those are the people who ultimately are going to make the greatest impact on whether, and how, educational policy is going to change. Science writers have a very tough job of digging into a vast variety of fields, trying to understand who is on which side of the controversy, and understanding it well enough to simplify for the masses without over simplifying it. When I was involved in selling a very complex piece of engineering software I got interviewed by people from the trade press, and I was quite impressed by the broad if not very deep knowledge they had of a diversity of subject matter. So even the Atlantic is going to get it wrong or not perfect. We their readers have to give them and the people they interview a break. ---
**Author:** ![SlackerInc](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/slackerinc/32/275_2.png) [@SlackerInc](https://boards.straightdope.com/u/SlackerInc)\ **Post date:** [March 29, 2014, 2:58am UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/36 "2014-03-29T02:58:33Z")
> [@monstro](#): > > Policy makers are not dictated by _The Atlantic_. They either have a staff of professionals whose job it is to know the research, or they hire consultants who do. At least in government, almost every policy can be traced back to technical document(s), which are drafted by scientific and technical committees, who are comprised of researchers or those who are heavily steeped in research. Again, we don’t have that type of technocratic government. Look at what’s happening with Common Core. The general public, or at least the moderately well informed subset thereof, needs to buy in. But in any event, the choice of whether to get more or less involved with one’s children’s schools rests with indivudual parents, not with policymakers. > [@](#): > > Most people reading a magazine article aren’t going to remember it an hour later. Maybe they’ll mention it in passing at the water cooler the next day if something triggers their memory. And that’s it. For people who are actually interested in the content, whether for professional reasons or because of a personal passion, they would do well to follow-up on their own. I am going to go out on a limb and guess you are not a parent? Most parents I have ever known treat recommendations like this much more seriously than that. Again, we are not talking about pure science that is aimed, for laypeople, at eliciting a “huh, interesting, a new subatomic particle”. We are talking about the most fraught subject I can think of: How Not to Screw Up Your Kids. > [@](#): > > \*f she is making these kinds of statements, without the appropriate qualifiers or hedging, then I agree with you that she’s being irresponsible. A point of agreement, hey. > [@](#): > > I don’t know you or your posts well enough to call you anything. Yet. Weird how people keep saying things like this, implying I’m a newbie. I have been commenting on the SDMB for many years (even going back to the AOL days, when I was **AlanKngsly** ). ---
**Author:** ![monstro](https://avatars.discourse-cdn.com/v4/letter/m/ba9def/32.png) [@monstro](https://boards.straightdope.com/u/monstro)\ **Post date:** [March 29, 2014, 3:37am UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/37 "2014-03-29T03:37:33Z")
> [@SlackerInc](#): > > Again, we don’t have that type of technocratic government. In theory, at least, yes we do. Not every reg is backed by science, but policy makers at least pay lip service to it because it makes whatever they dictate easier for the stakeholders to swallow. And the way they pay lip service to it is by forming independent technical advisory committees/panels, who are tasked with bringing science to institutional practices. Stakeholders (the folks who are most affected by arbitrary and capricious policies) demand this kind of rigor and are quite willing to sue if they think it’s not happening. I’m not saying government never has hare-brained schemes. But department/agencies do have institutionalized requirements that force beaurocrats to interact with egg-heads, who will then hopefully dissuade them of any crazy notions they may gotten from NPR or their bathroom magazines. > [@](#): > > Weird how people keep saying things like this, implying I’m a newbie. I have been commenting on the SDMB for many years (even going back to the AOL days, when I was **AlanKngsly** ). Sorry, I don’t have a good memory for names. With a few exceptions. all of you people are the same to me. 🙂 ---
**Author:** ![Hector\_St\_Clare](https://avatars.discourse-cdn.com/v4/letter/h/b5ac83/32.png) [@Hector\_St\_Clare](https://boards.straightdope.com/u/Hector_St_Clare)\ **Post date:** [March 29, 2014, 4:01am UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/38 "2014-03-29T04:01:42Z")
> [@Voyager](#): > > health effects of smoking were still controversial, correlation does not imply causation was used all the time by the tobacco companies. Sometimes the mechanism might be hard to find - doesn’t mean it isn’t there. On that note, it’s an irony of history that one of the big proponents of the ‘tobacco does not cause cancer’ thesis was R. A. Fisher- the guy who discovered a good deal of modern genetics, statistics, behavioural ecology and evolutionary biology, and who is one of the most influential scientists of the last century. Fisher was also an avid smoker, and apparently was extremely indignant at the idea his hobby was bad for you. Sometimes even incredibly smart people can make mistakes. ---
**Author:** ![SlackerInc](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/slackerinc/32/275_2.png) [@SlackerInc](https://boards.straightdope.com/u/SlackerInc)\ **Post date:** [March 31, 2014, 10:07am UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/39 "2014-03-31T10:07:54Z")
[Here’s another one,](http://www.npr.org/2014/03/21/292225798/does-diversity-on-research-team-improve-quality-of-science?ft=1&f=3) from ten days ago. This time we hear from the horse’s mouth: the Harvard researcher who published the study. In this case, the explanation offered strikes me as idealistic wishful thinking, almost the opposite of what I suspect is the true reason for the pattern discovered. Basically, the hypothesis is that scientific papers written by a multicultural group of researchers are “better” (because more cited) than those written by a more monocultural group: > [@](#): > > VEDANTAM: So if you count the number of citations, it gives you a measure of how effective and powerful papers have become. I spoke with Richard Freeman. He’s an economist at Harvard University. He decided to use the citation method, along with his co-author Wei Huang, to measure how the quality of scientific papers has changed. > > He looked at the ethnic identity of the authors of 1.5 million papers written between 1985 and 2008 and he compared the number of citations that papers received when they were written by diverse teams compared to when all the authors came from a similar background. Here’s Freeman. > > RICHARD FREEMAN: We found that if you write a paper largely with people of your own group, it’s likely the paper gets less citations than if you write it with a broader group of people. You tend to do better if you stretch. > > GREENE: So he’s dealing with more than a million papers he’s looked at. And, I mean, the link seems pretty clear. You have a more diverse group of researchers and a better paper. > > VEDANTAM: That’s right, David. What’s interesting is that this seems to affect all groups. So papers written exclusively by Anglo authors don’t do as well as papers that have Anglo authors and, let’s say, Chinese authors. But that’s not because Chinese authors always make the papers better. If you look at papers that are written exclusively by Chinese authors, those papers tend also not to be as good as papers that are written by diverse teams. As I was listening to this, I was already musing to myself “or is it because people are more likely to cite a paper that has an author of their own ethnicity–therefore, if you ‘cover your bases’ you’ve got a little hook for everyone?” And then almost comically, another point made later in the story would, one would think, seem to really underline this possibility–yet it never seemed to dawn on anyone, sheesh: > [@](#): > > VEDANTAM: Freeman thinks that what’s going on is that ethnic diversity on these papers is really serving as a marker for something else that’s really important in science. Here he is again. > > FREEMAN: Ethnic diversity is an indication of ideas’ diversity. People who are more alike are likely to think more alike and one of the things that gives a kick to science is that you get people with somewhat different views. > > GREENE: So avoiding group think is essentially good for science is what he’s saying here. > > VEDANTAM: That’s exactly what he’s saying, David. > > GREENE: And are there other types of diversity where this seems to apply? I mean, I can imagine geographic diversity, other things like that. > > VEDANTAM: You know, Freeman actually has looked at the question of geographic diversity and he’s found exactly the same thing - papers are going to be more highly cited if the teams come from different parts of the United States. So if you have a paper co-authored by people in Boston and Texas and California, that’s going to be better than if all the authors were in Boston or all the authors were in Texas. > > GREENE: And it sounds like one important thing to note here is that it’s not a matter of diversity broadly in the world of science; it’s a matter of looking at individual teams and making sure they’re different perspectives, different points of view, different backgrounds. > > VEDANTAM: That’s exactly right, David. In fact, Freeman said that he began to study when he noticed that when you go to different scientific labs today you see a lot of clustering along ethnic lines. So you see Anglos working with Anglos, Russians working with Russians, Chinese graduate students working with Chinese professors. What Freeman’s work seems to suggest is that this might make you feel comfortable but it’s probably not going to lead to your best work. So he noted that people tend to like to work in labs with people of their own ethnicity, but it never occurred to them that they might like to cite \*authors _of their own ethnicity_, irrespective of how “good” the papers are? Good grief. \*ETA: Or, for that matter, of their own regional affiliation. ---
**Author:** ![Smeghead](https://avatars.discourse-cdn.com/v4/letter/s/f1d935/32.png) [@Smeghead](https://boards.straightdope.com/u/Smeghead)\ **Post date:** [March 31, 2014, 11:48am UTC](https://boards.straightdope.com/t/correlation-is-not-causation/684741/40 "2014-03-31T11:48:45Z")
I don’t know if they looked at geographical diversity, or just, for instance, different ethnicities within the same university, but another factor is that truly interesting, important papers are likely to require a wider range of expertise than more routine papers. So if someone in the UK is collaborating with someone in China, the odds are good that they’re looking at something in a new way. If it’s more routine, it’s more likely that some guy down the hall can do the needed job. Also, just in general, there’s likely to be a correlation between the importance of a paper and simply the total number of authors in general. A paper that required ten authors is more likely - on average (obviously there are zillions of exceptions) - to be more interesting than one with two or three authors. At the very least, it’s going to have more data in it. Obviously, the more authors there are, the more chance there is to pick up an ethnically diverse mix. The report in question may have considered and accounted for these effects. I haven’t read it. [Previous page](https://boards.straightdope.com/t/correlation-is-not-causation/684741.md?page=1) [Next page](https://boards.straightdope.com/t/correlation-is-not-causation/684741.md?page=3)