Nicholas Christakis, a medical scientist perhaps best known for his controversial claim (see also here), based on joint work with James Fowler, that obesity is contagious, writes:
The social sciences have stagnated. They offer essentially the same set of academic departments and disciplines that they have for nearly 100 years: sociology, economics, anthropology, psychology and political science. This is not only boring but also counterproductive, constraining engagement with the scientific cutting edge and stifling the creation of new and useful knowledge. . . .
I’m not suggesting that social scientists stop teaching and investigating classic topics like monopoly power, racial profiling and health inequality. But everyone knows that monopoly power is bad for markets, that people are racially biased and that illness is unequally distributed by social class. There are diminishing returns from the continuing study of many such topics. And repeatedly observing these phenomena does not help us fix them.
I have just a couple comments here. I’m no economist so I can let others discuss the bit about “monopoly power is bad for markets.” I assume that the study by economists of monopoly power is a bit more sophisticated than that!
I have studied racial profiling, and I can assure you that this work is not about the claim “that people are racially biased.” I can also assure you that, whatever it is we have learned, it’s not true that “everyone knows” it.
As Duncan Watts has written so memorably, it’s easy to say that everything is obvious (once you know the answer).
Regarding the question of illness being distributed by social class: Is it really true that “everybody knows,” for example, that Finland has higher suicide rates than Sweden, or that foreign-born Latinos have lower rates of psychiatric disorders. These findings are based on public data so everybody should know them, but in any case the goal of social science is not (just) to educate people on what should be known to them, but also to understand why. Why why why. And also to model the effects of potential interventions.
The study of the contagion of obesity is just fine. In fact, I was once part of an NIH panel where where we recommended funding some of this research. But to say that this is the real stuff, and then to dismiss studies of monopoly power, racial attitudes, and variation in disease rates–that’s just silly.
Resources are limited, and I think it’s good to have open discussion about scientific priorities. So I applaud Christakis for sticking out his neck to participate in this debate. Even though I don’t agree with his particular recommendations.
The key observation is that resources are limited. No one is arguing that all political science is crap. But if you ask me, are we putting more resources into social sciences than they deserve, I’d say yes. (I’ve no clue how such a question can be studied objectively, I’m only offering an intuitive opinion)
So also, I’d wager we are producing a lot more graduates /PhD’s in social sciences than we ought to. Not everyone becomes a professor, and the non-professorial opportunities are limited. I suspect we are sowing seeds of a lot of future heartbreak. Or excessively subsidized make-work positions.
I’ll readily admit that a lot of the questions in social science are enormously interesting; but unfortunately interest alone makes for a poor argument for resource allocation.
Rahul:
Maybe so, but I think Christakis actually pointed to some areas (monopolies, racial attitudes, and variation in disease rates) that are worth studying and are actually more important than the contagion of obesity!
Rahul:
You say that we are putting more resources into social sciences than they deserve, but you have no idea how much money actually goes “into” social sciences, nor any conception of how much they “deserve.” Excuse me if it’s hard to take your “intuition” seriously. Ironically, if you had a degree in economics or sociology, you might have a clue how such a question can be studied objectively.
Also, a degree in social science set up graduates for myriad career choices, including graduate school in business and law. This is due, at least in part, to the fact that social science can provide a quality “liberal arts” education, training students in critical thought and effective communication. What you call “a poor argument for resource allocation,” I would call “the hallmark of a well-rounded education.” And, possibly, “the difference between a university and a glorified trade school.”
You are right that there are more PhDs than tenure-track job openings, but the same can be said for many professions, especially in fields where baby boomers are not yet retiring. Since the marginal cost difference (“resource allocation”) in teaching social science graduate students is negligible, why would you care if a poli sci department produces one or two PhDs per year above market saturation? Anyway, most departments are self-correcting: if they cannot place recent graduates, they will accept fewer students … regardless of the intuition of a random internet commenter.
>>>if you had a degree in economics or sociology, you might have a clue how such a question can be studied objectively.<<<
In all seriousness, I’d love to know how. How much money “deserves” to go into the social sciences? What’s a dollar figure for the US. Alternatively, what’s the “optimal” number of Social Sciences graduates we ought to be churning out? Tell me how you’d go about, even roughly, estimating this objectively.
I’d love to shed my arbitrary intuitions and get the answer on a sound, scientific footing.
The merits of studying “standard” social sciences (poly sci, econ, etc.) aside, I’m just having a hard time believing that social sciences as a discipline hasn’t changed. For example take this quote from the linked article: “It is time to create new social science departments that reflect the breadth and complexity of the problems we face as well as the novelty of 21st-century science. These would include departments of biosocial science, network science, neuroeconomics, behavioral genetics and computational social science. Eventually, these departments would themselves be dismantled or transmuted as science continues to advance.”
I’m going to be a freshmen in college soon studying “markets and social systems engineering” also called (the naming convention confuses me) “networks and engineered social systems”. There’s a lot of computer science, economics, and network sociology in the “core”. In fact, I might be tempted to say it sounds a lot like… “computational social science”.
Now, I understand that one undergraduate major says little about the state of social science academe as a whole. I also understand that the papers I read thereof are disproportionately selected towards topics that interest me – namely computational social science. That said, there is certainly no dearth of interest and study on networks in the social sciences.
Take the 2010 textbook by Stanford’s Matt Jackson: Social and Economic Networks (http://www.amazon.com/Social-Economic-Networks-Matthew-Jackson/dp/0691148201) – which was the basis of a very popular online course in computational methods for network science.
Take a lot of the microfinance empirics from India which study the diffusion of financial information through graph theory.
I’m 18 and I can name the “hot topics” of this “new” social science without sounding like a heterodox quack. And the work is being done by 50 year olds. Just imagine the stuff that this generation of grad students will produce.
I’m not an expert on the “traditional” social sciences and can’t comment on their usefulness, but isn’t it a bit passé to say that the social sciences haven’t changed? You don’t have to look particularly far to see the new computational stuff.
I don’t buy it… I’ve read and graded undergrad papers, and either 1) you will forget how to write a complete sentence (let alone stringing several together into a coherent argument) the minute you start college, or 2) you are not an 18 year old pre-freshmen. Obviously, I’m ruling out that you are just one of those students who is going to make their TAs and professors very happy by being smart and clear and concise.
Penn? Those jerks don’t hire smokers anymore, and Philly is terrible. How about one of our fine public universities out here in California? We have better weather, nicer public transportation, and they would probably let you play with the big lasers at Livermore (ok, maybe not, but still).
I’m having a hard time understanding what you said. But I am definitely 18 and I sincerely hope I won’t forget how to write a “complete sentence” over the next few months.
It was just a snarky and long-winded way of saying that you write and think incredibly well for someone so young. And then I was wondering if you were referring to the program at Penn, and suggesting some silly reasons that California might be preferable.
Sorry…. sometimes I forget that my humor (¿nonsense?) isn’t always intelligible over the internets.
Alright, in retrospect that makes sense and I sound dense, thanks! :)
I’ve never been to Philly or California. I can’t comment. But as far as the weather is concerned, I spent 7 years in Iowa and I don’t know if the winters can get worse than that!
I was referring to Penn, and for what it’s worth that does seem unique in the total package, but in my search I found quite a few other institutions offering things along the lines of what Nicholas Christakis was suggesting – so I doubt it’s an outlier, and is probably more (along with similar things elsewhere) a harbinger of what is to come.
Another important point is that a lot of these, novel as they are, are still founded on traditional techniques. For example standard rationality is being updated into networks through modification of things like restricted Bayesian learning on networks (bounded for tractability). And this sort of stuff isn’t all that new, dates back to DeGroot 1974 ish I want to say. Makes it a lot easier to test against empirics but if you look under the hood it has a distinctively neoclassical econ flavor.
I know a lot less about econometrics, but it’s hard to say there hasn’t been some concrete soul-searching since the ’80s when Leamer published “taking the con out of econometrics”.
I mean I can understand the frustration a social scientist might have with his discipline. But everything he wants seems to be well “under way”, or at least there’s the foundation for further research, you know?
Ashok:
Regarding your last paragraph, I think the key might be the bit from Christakis’s article, when he refers to “citizens, politicians and university donors.” In the Ted-talk world in which Christakis lives, I’m sure he gets a lot of positive feedback about his ideas, and he gets lots of approving nods when he refers to the outmoded 20th-century university, etc. This kind of thing also probably sounds good on NIH applications. NIH panels are mostly not reviewed by social scientists, and I’d guess that the medical scientists on NIH panels like hearing stories about how the social sciences are so backwards. I’m guessing that, in the medical-research world, social science is a soft target, and perhaps Christakis hasn’t really thought so hard about what it really means to study monopolies or racial attitudes or variation in disease rates from a social-science perspective. As a start, he might consider reading my (edited) book, A Quantitative Tour of the Social Sciences, which gives a sense of how scholars in different social sciences think about research.
Yep. As I said I thought that kind of yesteryear criticism was passé, in the Ted-world as well as academia, but I’m part of neither so I wouldn’t know.
The sad and ironic thing is I think Christakis – rightly – wants social scientists to adopt certain methods and rigor from the natural sciences. He says:
“[Social scientists should devote their energy to fields which], not coincidentally, lie at the intersection of the natural and social sciences. Behavioral economics, for example, has used psychology to radically reshape classical economics.”
I don’t know how academics work, but it seems this sort of op-ed works more to further the wedge and strengthen the silos than really promote the sort of cross-fertilization that we all want.
I also like to think academics study what (they think) is important. Surely traditional social sciences are better than the kind of contrived interdisciplinarism that’s the flavor of the month (decade?) in many popular publications? It might be too free market of me, but surely if there is an clearly evident benefit to importing a certain method from the natural sciences one eager-to-rise social scientist would have done it already? I mean I’m not saying academia is perfect, but this doesn’t stand up to the standard critique of journal bias (like DSGEs in economics, or something).
This might all be something you’ve talked about in A quantitative tour – I should read.
Let me rephrase my above comment slightly. I wrote, “In the Ted-talk world in which Christakis lives.” But this is not fair. Christakis lives in many worlds, and to identify him solely with Ted would be about as fair as dismissing me as “a blogger” or as “that guy who sometimes gets quoted about the polls.” I suspect that Christakis’s experiences with high-profile events such as Ted are relevant to his perceptions of the views of politicians and university donors. But I assume that his larger views about science are formed much more by his own experiences as an active multidisciplinary researcher.
“Another important point is that a lot of these, novel as they are, are still founded on traditional techniques.”
I think this gets at something important, certainly more important than what names people want to give their departments. I may study in a department called “economics”, but there are epidemiologists who look at almost exactly the same questions as I do. The difference tends to be in the set of tools we employ to look at the problem (and the language/jargon we use to describe what we’re doing).
But this also gives me a little bit of pause regarding multi-disciplinary degrees (maybe not at the undergrad level, but at the graduate/professional level). A social science gives you a set of tools for looking at the world. It seems to take people about 5 years to get reasonably good at those tools. So if students are now expected to learn tools from a bunch of disciplines, I can’t imagine they’ll get very deep into any of them. I basically worry that we will be producing researchers who don’t understand deeply enough the complex quantitative tools they are using. I’m five years into studying causal inference and quasi-experimental methods, and I still have ~1,500tons of things to learn. If it weren’t for my colleagues in Nutrition and Epi, I would’ve already tried to publish some really wrong-headed research (my methods were good, my biology was terrible). I rely on them because I can’t learn everything myself.
I think it is incredibly important for researchers to push into new areas of research. I think multi-disciplinary teams (I’m obviously on one) are great – everyone brings different viewpoints and tools. But I sure don’t think that someone could learn everything about biology, applied microeconometrics, epidemiology, and public policy in one degree and so be qualified to do “biosocial science”. I think they are more likely to learn a little bit of each, and then make tons of mistakes.
In the end, I really think people learn to be very good at one thing (maybe a couple for the really, really gifted people), and then they apply that thing, with the help of others, in new and interesting directions. Would it help if there were more inter-departmental collaboration? Sure. Does re-organizing the Institution accomplish that? Maybe, but maybe it just arbitrarily imposes cross-pollination in popular (as in “pop”) ways (Andrew’s “TED” comment).
…reading back over this, maybe I’m just thinking like a conservative today. So I’ll conclude with: social sciences are in fact changing and adapting and cross-pollinating, as you point out, and I fail to see how setting up a bunch of cool-sounding, inter-disciplinary programs is really going to push that forward, and not just water it down.
I think some of the tools could and should be consolidated. There’s a lot of isolated, domain-specific methodology that really shouldn’t be. For example, a lot of different things like regression discontinuity, instrumental variables, even multilevel modeling, can probably be presented in a graphical model framework. ANOVA in the traditional sense can probably be folded into regression.
A lot of time is wasted on suboptimal methodology. Look at how much training is wasted on p-values in domains where testing is not appropriate – see Sander Greenland’s papers. (and people still don’t use testing appropriately). Multiple hypotheses testing in biology should move away from multiple comparisons methods and towards shrinkage approaches in statistics/social science, etc.
Just as a quick follow-up, a quick google suggests that Chistakis does not seem to know anything about the behavioural sciences [1]
Since I was an undergrad in Psych, the Perception and Cognition people in the Psych Departments seem to have morphed into part of the Cognitive Sciences departments and presumably the behavioural biologists and behavoural gentics psych types are in the new Institutes for Behaviour Genetics that I pickup in a 2 second google, or the various schools or institues mainly concerned with behavioural economics
It kinda reminds me of one of my favourite introductory sentences in a article
As late as the end of the 19th century, even a visionary like Jules Verne could not imagine a city with more than a million inhabitants.
Jonas Rabinovitch and Josef Leitman
Scientific American March 1996
Hint: Poplation of London UK was ~1.1 million in the early 1810’s
1. I don’t consider psychology or such things as part of the ‘social sciences’ but that may be because of historical funding rules for research in Canad
I understood your comments immediately – revealing that 1) I am older and cynical, and 2) that Ashok more than likely is young (by comparison). Additionally, your comments were ironical not snarky.
Have any of you read his obesity and networks study? I took a quick look at the analysis, and it seems to suffer from the mismeasured peer effects problems that have been studied in economics. Basically, it looks to me like he doesn’t properly control for confounding variables that effects groups of friends. If my friends and I all like to eat lunch at the same unhealthy restaurant each day, then we might get all get fat together; if the restaurant closes, then we might get skinny again, without network effects ever playing a role.
His argument seems to be that certain types of friendships matter more than others, which lets him identify the causal direction. But those are probably exactly the types of friendships which most closely proxy for the peer groups that are affected by the omitted variables! If economics, at least, you would need an instrumental variable for this to be remotely convincing.
Or did I miss something? Did you find the paper convincing?
Also, I’d like to point out his suggestion that evolutionary psychology is a new, exciting field with insights to bring to the social sciences…
Ah, sorry for the typos!
I had looked at the obesity study when the NYTimes first covered it, and no, it is not convincing AT ALL, for the reasons you mention. In fact those authors did a series of studies with that same design, and all of it was very poor quality research. Obesity may very well be contagious, but from Christakis’ study, we don’t learn much about that question.
I read the article as a testament to the value of cross-disciplinary research rather than a serious attack on traditional social science per se. In general, I think Christakis has the right idea, but I wish he would have mentioned a few other explicitly cross-disciplinary departments like Cal Tech Social Science, OPIM at Duke, and (my home) Social and Decision Sciences at Carnegie Mellon. Rather than stagnating, I think the social sciences are actually becoming much MORE interdisciplinary. The founding of new journals like the forthcoming Review of Behavioral Economics, and the formation of the US version of the “Nudge Unit” suggest to me that interdisciplinary work (both academic and applied) is becoming much more mainstream. In my own opinion, Christakis rightly supports collaboration, but comes off as a bit too alarmist.
As an economist, I can say that Christakis’s statements about economics are clueless. This leads me to conclude he is probably clueless in general about everything outside his own narrow field of expertise. I would dismiss him as yet another pretentious medical-natural sciences blowhard.
It is sad, because I think there is a great potential idea in having professors be “free” rather than part of a department, and sponsored by research centers or “new departments”. It is an exciting idea, though fraught with practical obstacles.
I did note that he complains about social science students not doing lab work. Well, in psych I did lab work with rats, various perception experiments, nearly froze my postior off in a car in winter observing driver behavior in shopping mall and managed to talk my way into Bell Canada to do some environtal observations.
Others in my class spent weeks in the local psyciatric hospital (as students not patients).
This was a long time ago but unless things have changed very drastically he seem equally clueless about psychology.
Great comments on a terrible article. I understand that the NYT limits how well once can expound on an argument, but there was barely anything salvageable from that piece. The most frustrating thing is that his comments will probably be eaten up by people unfamiliar or already hostile to the social sciences.
I haven’t read the original article yet, but I’ll just point out that this is the first time I’ve heard in the last 5 years that economics is essentially a class of solved problems and is therefore boring. (It’s a refreshing critique)
The link to an old post on suicide rates reminded me, my colleagues have recently released a great interactive tool to explore these relationships. I was unable to share it when I last saw that discussion.
Comments on the old post are closed, so I’ll share the link here: GBD-Compare Suicide Mortality Rates. The good news is that the rates are dropping. Unfortunately, 2010 Finland is just reaching the levels of 1990 Sweden.
When I talk about this to certain audiences, I have no reservations calling it “Computational Social Sciences”. I think “Big Data” is also accurate. But I agree, the distribution of illness is *not* something everybody knows, and it turns out to be reasonably challenging to estimate well.
Dear Andrew:
I admire and respect your work, and so I am surprised by the tone of some of your comments on your blog, above.
Have you had a chance to read any of our papers recently? We have many experiments and observational studies in major journals that have survived grueling peer review. We have also responded at length to our critics. Here is the recent response to our critics (with comments and rejoinders):
http://humannaturelab.net/wp-content/themes/human-nature-lab/media/pdf/publications/articles/SIM_special.pdf
And here are a few papers (from Nature and PNAS) that further explore social networks using diverse methods, both observational and experimental.
http://humannaturelab.net/wp-content/themes/human-nature-lab/media/pdf/publications/articles/127.pdf
http://humannaturelab.net/wp-content/themes/human-nature-lab/media/pdf/publications/articles/126.pdf
http://humannaturelab.net/wp-content/themes/human-nature-lab/media/pdf/publications/articles/106.pdf
http://fowler.ucsd.edu/massive_turnout.pdf
One cannot have it all, andrew, as you know very well. If we use observational data, nihilists (who you often take on!), criticize the difficulty (in the limit, the impossibility) of causal inference; if we use experiments, critics object to the lack of verisimilitude.
Finally, as someone who spends a lot of time trying to advance the public understanding of science (i mean you, not me!), i don’t think you really should be calling the kettle black.
As to your original complaint, let me say this: my main point (within a constrained 800-word format, for which the editors write the title, not the authors) is that we can learn something from the natural sciences about institutional forms and about ways of doing science (just as they surely can learn from us). This does not mean what we social scientists are doing is bad; but it would be arrogance to assume we have nothing to learn. And re-deploying talent to tackle new (and, yes, old!) topics is fine, but we should be judicious, and, yes, I think we should move to the scientific frontier. Is that a claim you don’t agree with?
Sincerely,
Nicholas
Thanks for joining the discussion. There are so many comments! I am glad that you linked to the response you wrote. I now see that Andrew had already linked to it but I would not have noticed this otherwise. I also appreciated some of your advice, particularly the idea that social scientists should incorporate undergraduates into their research. This seems particularly useful in my field of economics.
As for the tone of the comments on this blog: Comments are anonymous, and so people are less polite. I also suspect that commentators are angry because they feel that your op-ed is personally threatening to them (you called their disciplines “boring” and “counterproductive”, after all). The op-ed also seems to give justification to the recent NSF funding cuts to political science research. Their comments on this blog are ultimately harmless to you, but your op-ed in the NYT may cost them their job.
No doubt social scientists would be wise to adopt methods from the hard sciences. However, I suspect that the reverse is true to a far greater degree. Economics, sociology and psychology are all younger than biology, physics or medicine, so following your line of reasoning, should the older and more-established fields not adopt interdisciplinary methods from the social sciences? I find your research a lot more interesting than any of the articles in the most recent issue of Science! (“Pigeon hierarchies shift when the birds take flight”). Networks have been part of sociology and economics as long as those field shave existed. It seems to me that much of your work adopts methods that are rooted in the social sciences to answer questions from medicine. If anything, social scientists have had enormous success colonizing the hard sciences.
Moreover, several of the particular types of interdisciplinary research you would like to support have been unsuccessful. Networks, as I mentioned, have a long theoretical history in economics, but there have been powerful critiques of empirical work on networks, critiques which I think public health researchers would be wise to note. Also, last I checked (a while ago, admittedly), evolutionary psychology research has had enormous difficulty being more than just-so stories. When health economists want to publish something that is statistically dubious, they send it to medical journals (eg, J. Sachs’s retracted Lancet article on the Millenium Villages). There are notable cases of influence going in the other direction, like the success of Prospect Theory in economics – but note that even this was an adoption from psychology, another social science.
As someone with an economics background, I also have to ask: Where’s the market failure? The number of undergrads in health fields has increased, and the social sciences have maintained their popularity. It seems like students have done a good job deciding for themselves what is interesting.
Thanks for replying on this thread.
It is interesting to me that you are surprised that Professor Gelman has been sharply critical of your observational studies of social contagion, given that numerous statisticians and other methodologists who care about credible causal inference have found this work unconvincing. My sense is that many of these folks remain unconvinced by the SIM article.
Personally, I think positive peer effects (and thus social contagion) are ubiquitous, so I would expect that (at least for some populations) obesity is “contagious”. However, I have trouble thinking that the parameter estimates from many of these papers tell us much about the true size of these effects.* Other claims, such as about “three degrees of influence” are not easily made consistent with any reasonable data generating process (if three degrees, why not four? — this is just an artifact of the statistical significance filter and the periodicity of the data collection waves). I would guess that, even if not key to other results, these kinds of claims make statisticians (among others) averse to positively evaluating the rest of the work.
Dean
Dear Dean and Jack (not PQ):
Dean: These issues you note are all discussed in our Statistics in Medicine piece, and it’s not worth revisiting in this location; interested readers can see for themselves. If you have any recommendations for better methods to use, or any ideas for other experiments that James and I could run (beyond the many we have done, including with 61,000,000 people on facebook, and thousands of people in other online settings), we’d be happy to hear suggestions (you can email me or James at [email protected] or [email protected]. We are genuinely interested in advancing science in this area, as we note in our SIM piece..
Jack: yes, network methods were invented by sociologists! And that is a very good example of how sociologists have influenced physicists, economists, and others. Linton has written a fine history of this topic, and I discuss this in “Social Networks Are Like the Eye,” at http://www.edge.org. As I note above, I think the natural sciences have a lot to learn from the social sciences. Plus, I think even the distinction between the natural and social sciences is an artifact of Medieval religious thought — separating man from nature. We would do well to abandon that, I think. Just take a look at recent beautiful work by Frans de Waal on primate and elephant cooperation, for example. As for funding for the social sciences (e.g., the preposterous cuts to the NSF budget), I am, like you, appalled. But we have to get our house in order if we are to attract the funding we deserve. Or that is one man’s opinion.
Sincerely,
Nicholas
I agree that this might not be the place to rehash those points in detail. I was more making a “meta” comment that (to me, at least) it is unsurprising that many statisticians continue to balk at much of this analysis. It certainly seems that many of your interlocutors have not been convinced by the arguments in AOAS/SIM piece. I guess what I’m saying is: I was surprised at your surprise that many statisticians, even having read your SIM piece, remain unconvinced of the credibility of some key results.
I appreciate that you and Professor Fowler have done non-observational work in other domains.* But for those of those that expect peer effects in many places (I would hope most social scientists do), it is not clear that (even very solid) evidence of (e.g.) peer effects via Facebook on voting (my Facebook Data Science colleagues also worked on this experiment) is particularly informative about the size and importance of any peer effects in obesity-related behaviors. That is, I would expect that many folks don’t see this work as appreciably adding to evidence for many of the other public health claims.
* Not that there is anything categorically wrong with observational studies of social contagion.
From the first linked paper: “There are two broad classes of investigations of networks that we have undertaken: studies of network topology (and its determinants), and studies of the spread of phenomena across network ties. Although we have done work on the former [5, 7–13], here we will focus primarily on the latter, discussing analyses of the flow of behaviors, affective states, or germs.”
Nothing in your cited work justifies using nodes and flows between nodes to establish causal relationships, particularly on latent variables, such as behaviors and affective states. Practitioners of network analysis (in natural and social science) seem to think that everyone expects the underlying assumptions of network analysis , emergent properties arising from nodal connection. Most network analysis, regardless of disciple, as tautological. I see the value in employing system analysis to address complex problem, but network analysis is not a synonym for systems analysis.
My objection: There is a subtext to the article that I perceive to be defacto positivism. A desire to be scientific in the natural science sense that has led the social sciences down dead end paths before. I hear (rightly or wrongly) the scientism that Wittgenstein sought to resist. I have no doubt that science in all its many forms is the most important “tool” of progress, yet it is socially blind progress. It was Samuel Messick whose writings on test validity pointed out to me that measurement (and the resulting data) cannot find valid application without participating in a hermeneutic and value based process. Part of the social sciences must be devoted understanding and applying these processes that serve as a guide to progress.
please, there are enough real objections to the analysis here, no need to provide straw men.
I do believe there are valid measures of value that fall outside the realm of the scientific method, but it only muddies the water to bring them up here because they are not at all specific to this work. There are enough scientific issues with this work to focus on the critiques specific to this study, e.g. jsb’s comment above.
Nicholas:
I might be the dissenting voice on this thread, but I really liked your article.
In my view, there’s two reasons for the hostile tone of comments you rightly perceived on this thread:
(1) Andrew is (again, IMHO) chronically prone to excerpt the worst bits of articles. Taken out of context, these snippets might justifiably cause many readers to misinterpret your point.
(2) The other reason is excellently summed up by parts of “Jack (not PQ)”‘s comment in this thread. To quote him:
“commentators are angry because they feel that your op-ed is personally threatening to them….. The op-ed also seems to give justification to the recent NSF funding cuts to political science research. Their comments on this blog are ultimately harmless to you, but your op-ed in the NYT may cost them their job.”
In short, there’s a very understandable element of self-preservation in action here. Try getting responses to this article from non-social-scientists (or at least non-career-academics) and you might get a very different, probably more encouraging opinion of your Op-Ed. In my personal interactions too, the recent cuts in funding have made a lot of my Social Science peers extremely touchy and defensive of the slightest suggestion of change or any insinuation that all may not be right in their systems and setups.
But yes, I did like your article.
I can assure you my job wasn’t at stake because of this op-ed. Not the least because I don’t have a job, but that aside I’m pretty sure AG’s research is on pretty solid ground as far as value is concerned.
It’s one thing to say that there’s a bit of turf defense among social scientists – which there certainly is. Completely different to accuse any reasonable dissent to fear of job-loss.
I’d bet a handy sum that no one commenting here is threatened by this op-ed. And I think you’re missing Christakis’ point if you think he wants to threaten these jobs to begin with.
Holy non-sequitur. That should read: Practitioners of network analysis (in natural and social science) seem to think that everyone accepts that the underlying assumptions of network analysis can support meaningfully analysis of emergent properties of systems, such as “contagions” arising from nodal connections.
I read the NYT piece more as a call to knock down barriers in social science. Kind of the CalTech model. I’m all for it.
I can’t be asked to speak to my discipline (ok I don’t really have one). I want to speak to a substantive problem across as many disciplines as it is necessary to resolve it.
In an interconnected world of ginormous data and computing power there is no need to be parochial
Actually, I fear being parochial is a reasonable response to the competitiveness of the academic world, as I argue below.
Let’s be honest: if universities force social scientists and natural-physical scientists to create new, joint departments, the natural scientists will “devour” the social scientists and the new department will be just a bigger, natural-physical science department. My social scientist colleagues and I have seen it time and again in the context of grants, research centers, and other instances where the two worlds collide.
Yes, it is a question of self-preservation. And no, we social scientists (okay, my acquaintances and I) don’t want anything to do with natural-physical scientists, who already collect ~90 percent of grant and subsidy money and still push to squeeze social sciences even further.
I disagree, and I’m coming from a natural science background. Natural scientists are absolutely _terrible_ at interpreting observational data. They just don’t have the firsthand experience of being burned by the statistical paradoxes in ways that epidemiologists, economists, and statisticians have. In fact, Christakis’s optimism about “contagion” as a causal mechanism relative to serious veterans like Larry Wasserman is probably a small example of this.
I will say this, in epidemiology and a lot of social science, the skill is very bimodal. You have people like Sander Greenland, Miguel Hernan, Judea Pearl, and Andrew, who take this stuff seriously and respect the complexity of these systems. And then you have a lot of people who are just good at gathering “large data sets”. The latter often like to bring up how many million people they have in their data as a justification for why their interpretation is right.
Not to mention that closing departments is basically the only way to fire large numbers of tenured faculty. I doubt that any of the commentors are especially concerned about that particular part of the basket of proposals in the NYT piece, but I’d be surprised if that didn’t factor into some of the support Christakis has gotten from “politicians and university donors” and (I’m assuming) university aministrators.
@Anonymous and @Gray: All good points, but they do not contradict my basic position, which is that when you put social scientists and natural-physical scientists together, the latter take over and dismiss the value and concerns of social scientists. Maybe my experience (and that of my acquaintances) is unusual, but I doubt it.
I agree that both camps can (and should) learn from one another, but I’m being realistic here.
If you will insist on taking this closed, protectionist stance of non-engagement with the rest for fear of losing turf, I think you might have worrisome years ahead. It’s not a constant, guaranteed sized funding-pie that you must fight off the natural-physical scientists to retain your share.
I’d rather think of these collaborations as a way to strengthen your proposals and add relevance / differentiation. Novelty for novelty’s sake sucks, but OTOH, so does repetitiveness, stagnation and questionable relevance.
My hunch is, in the years ahead, tougher questions will be asked, and a lot more convincing will have to be done to justify current levels of Social Sci. funding. I’d rather collaborate and have natural scientists take some share, rather than push them afar and be starved of funding.
Of course, if you are one of the stars of the profession, none of this will impact you. But for more mediocre investigators all these options will have to be on the table.
Pingback: A week of links - Evolving Economics
Turns out some ideas are a lot better when sussed out in an Edge conversation than a NYT op-ed.
http://www.edge.org/conversation/a-21st-century-change-to-social-science
Still have my differences, but this is a fresh and relevant argument.
Pingback: Telling a new story… | olivianuamah