Is Facebook “Naturally Occurring”?

December 22, 2011 at 11:07 am 16 comments

| Gabriel |

Lewis, Gonzalez, and Kaufman have a forthcoming paper in PNAS on “Social selection and peer influence in an online social network.” The project uses Facebook data from the entire college experience of a single cohort of undergrads at one school in order to pick at the perennial homophily/influence question. (Also see earlier papers from this project).

Overall it’s an excellent study. The data collection and modeling efforts are extremely impressive. Moreover I’m very sympathetic to (and plan to regularly cite) the conclusion that contagious diffusion is over-rated and we need to consider the micro-motives and mechanisms underlying contagion. I especially liked how they synthesize the Bourdieu tradition with diffusion to argue that diffusion is most likely for taste markers that are distinctive in both sense of the term. As is often the case with PNAS or Science, the really good stuff is in the appendix and in this case it gets downright comical as they apply some very heavy analytical firepower to trying to understand why hipsters are such pretentious assholes before giving up and delegating the issue to ethnography.

The thing that really got me thinking though was a claim they make in the methods section:

Because data on Facebook are naturally occurring, we avoided interviewer effects, recall limitations, and other sources of measurement error endemic to survey-based network research

That is, the authors are reifying Facebook as “natural.” If all they mean is that they’re taking a fly on the wall observational approach, without even the intervention of survey interviews, then yes, this is naturally occurring data. However I don’t think that observational necessarily means natural. If researchers themselves imposed reciprocity, used a triadic closure algorithm to prime recall, and discouraged the deletion of old ties; we’d recognize this as a measurement issue. It’s debatable whether it’s any more natural if Mark Zuckerberg is the one making these operational measurement decisions instead of Kevin Lewis.

Another way to put this is to ask where does social reality end and observation of it begin? In asking the question I’m not saying that there’s a clean answer. On one end of the spectrum we might have your basic random-digit dialing opinion survey that asks people to answer ambiguously-worded Likert-scale questions about issues they don’t otherwise think about. On the other end of the spectrum we might have well-executed ethnography. Sure, scraping Facebook isn’t as unnatural as the survey but neither is it as natural as the ethnography. Of course, as the information regimes literature suggests to us, you can’t really say that polls aren’t natural either insofar as their unnatural results leak out of the ivory tower and become a part of society themselves. (This is most obviously true for things like the unemployment rate and presidential approval ratings).

At a certain point something goes from figure to ground and it becomes practical, and perhaps even ontologically valid, to treat it as natural. You can make a very good argument that market exchange is a social construction that was either entirely unknown or only marginally important for most of human history. However at the present the market so thoroughly structures and saturates our lives that it’s practical to more or less take it for granted when understanding modern societies and only invoke the market’s contingent nature as a scope condition to avoid excessive generalization of economics beyond modern life and into the past, across cultures, and the deep grammar of human nature.

We are, God help us, rapidly approaching a situation where online social networks structure and constitute interaction. Once we do, the biases built into these systems are no longer measurement issues but will be constitutive of social structure. During the transitional period we find ourselves in though, let’s recognize that these networks are human artifices that are in the process of being incorporated into social life. We need a middle ground between “worthless” and “natural” for understanding social media data.

Entry filed under: Uncategorized. Tags: , , , , .

Scraping Using twitteR (updated) How the poor debtors still sell their daughters, How in the drought men still grow fat

16 Comments

  • 1. Jesse  |  December 22, 2011 at 3:01 pm

    Facebook data collection is ‘unobtrusive’. I would just call it poor word choice.

    It’s like where a museum is trying to figure out the most popular exhibit. So instead of surveying people, they just keep track of how frequently they have to replace floor tiles in each of the exhibit galleries due to wear. That’s ‘naturally occurring’ right?

    • 2. gabrielrossman  |  December 22, 2011 at 3:12 pm

      What you’re calling “unobtrusive” is the same as what I’m calling “observational,” though I actually think “unobtrusive” is the better term.

      The deeper issue is whether we want the term “naturally occurring” to mean any and all data not resulting from the researchers’ actions or if we also think that the proprietors of spaces can introduce sufficiently strange constraints that we wouldn’t think of them as natural.

      For instance, suppose that the museum had a system of maps with suggested routes, audio guides, cordons, guards who chastise you for standing in the wrong place, etc. If we examine the floor tile wear (or security system motion detector data, or whatever) this will tell us what people do but only under the constraints the museum itself has imposed on their motion. It is necessarily true that this informs us how people move about an environment with those constraints, but may or may not be informative about how people move about in a structured environment. I don’t know about you, but in my everyday life it is pretty rare for me to make a circuit around the interior of a room where I always face the wall (which is always one meter in front of me) and where I pause for one to ten seconds every few meters.

  • 3. Exiled Antipodean  |  December 22, 2011 at 3:06 pm

    The difference is the floor tiles probably don’t structure what exhibits people choose to see.

    Facebook actively tries to shape who you interact with, as we’ve seen from the complaints when they change the interface. When Facebook promotes posts from people you interact with more to the top of your feed, and tries to predict who you want to see it doesn’t become a “natural” observation of interactions.

    • 4. Graham Peterson  |  December 4, 2012 at 10:40 pm

      “When Facebook promotes posts from people you interact with more to the top of your feed, and tries to predict who you want to see it doesn’t become a “natural” observation of interactions.”

      The problem here reminds me of Shrodinger’s Cat. And I do not think there’s any resolution other than to accept that there is no such thing as “data” ultimately, though in one context or another one variable will approach a “given [or constant]” end of a spectrum and another the “variable.” That will always be a relative distinction subject to human judgement in research, though.

      As I understand it, which is not well, the sorts of conditional probability trees Google, Netflix, and Facebook use to predict user preferences do a remarkably good job of – predicting user preferences. As evidenced by hoards of people volunteering to use them over other interfaces designed to predict their preferences. It’s not clear at all Facebook central or Google central can systematically manipulate that network — there’s quite a bit of human ooph built into all the voluntary action taking place within it.

      These systems are designed to predict what people want. Organizational hierarchies emerging on the internet are ostensibly as close to “organic natural experiment in emergence” as you’re going to get — Google after all won the day by mimicking the structure of academic citation trees, which I suppose one could argue merely represent another level of structure, in which there was never an original “what scholars wanted,” but merely a response to the paths carved for them by the academy.

      We should I think expect interwebs as they evolve to take the same shape a vast majority of human institutions so far have, if there is any sort of equilibrium or logic built into socialization (God I hope so – uncoordinated behavior sounds like a mess).

  • 5. Graham Peterson  |  December 4, 2012 at 3:49 pm

    “If researchers themselves imposed reciprocity, used a triadic closure algorithm to prime recall, and discouraged the deletion of old ties; we’d recognize this as a measurement issue. It’s debatable whether it’s any more natural if Mark Zuckerberg is the one making these operational measurement decisions instead of Kevin Lewis.”

    Crushing point.

    “You can make a very good argument that market exchange is a social construction that was either entirely unknown or only marginally important for most of human history . . . . [therefore we should] only invoke the market’s contingent nature as a scope condition to avoid excessive generalization of economics beyond modern life and into the past, across cultures.”

    Terrible point.

    Economic historians have shown in voluminous detail that markets pervaded societies globally for thousands of years, in long-distance trade (archeological evidence nearly as old as language), in wage work (routinized production centers or “factories” dating to ancient civilizations), in monetary transactions (clearly as old as the bible, and looking at the archeological evidence, even pre-settled-agriculture), in banking (ibid.), and in just about any other metric one can dig for.

    And all of that evidence gathered and presented for the mere existence of markets, saying nothing about the methodological individualism one may or may not claim motivates markets. Though it is a wonder why wages trebled during the Black Death, if supply and demand did not exist before 1850.

    • 6. gabrielrossman  |  December 4, 2012 at 4:02 pm

      On the history of markets, everything you said is consistent with my claim that markets are historically marginal. I agree that there has been some trade, especially over long distances, throughout history. This is not the same thing as having a market society. Some of the things you’re describing were organized as oikos, not agora. There’s no contradiction between recognizing this and still saying that the fundamentals of supply and demand can have effects that roughly approximate what we’d expect from econ theory, even if exchange is organized as patronage or primitive communism. That is, showing that wages trebled with the changing ratio of land/capital to labor doesn’t mean that in the 14th century Joe Sixpack got paid by ye olde paychecke which he spente at WalMarte.

      Read Graeber’s “Debt” to see what I mean.

      • 7. Graham Peterson  |  December 4, 2012 at 4:31 pm

        Oikos and agora are not disjoint sets. See Zelizer, and for instance the reciprocity and ethical, social-role-defining relationship you have with the guy at the sub shop you go to regularly. Nor were they disjoint in history. Mesopotamian clay tablets with trading orders written on them, roughly 4,000 years ago-ish, express a deference to ritual and sympathy.

        See Malinowski’s ethnography of Trobriand Islanders: he confronts directly the supposition that they and other indigenous cultures were communistic, and knocks that idea out of the park. Count on Mauss to read the reports of actual primary data in a way that satisfies his noble savage priors, though.

        To the extent you’ll allow me to limit the definition of Oikos to command organizations, you might see that in fact the overwhelming majority of behavior in the modern market is guided by Oikos, within the firm. See Kenneth Arrow, Ronald Coase, Herbert Simon, on the prevalence of institutions in the economy and the actually-paltry amount of market exchange that goes on. Pareto-improving trades do not a market make — technology does — and technology is a commons, developed inside the firm and on the bowling league over Marlboros.

        And I’m not trying to pull economic-theory trump cards, because we both probably agree what a stupid game that is, but “That is, showing that wages trebled with the changing ratio of land/capital to labor,” is incorrect. It would be the exogenous drop in labor supply relative to labor demand, determining wages. We would expect in fact the shortage of labor relative to land to have increased the marginal productivity of the individual worker (more of the best land available to him) — but aggregate output to have declined.

        Finally, I fail to see what is in any way desirable about societies organizes by patronage, as they presented dastardly measures of both relative (of concern to sociologists) and absolute (economists) poverty compared to now. See Fogel Escape From Premature Hunger and Death for exactly how much. I the secular trend since Early Modern Britain has been a reduction in relative inequality by a factor of about 3. Add to that the absolute rise of average national income by about a factor of 10 at the lower and 30 at the upper estimated bound, and patronage starts to look like a pretty crappy system from any angle.

      • 8. gabrielrossman  |  December 4, 2012 at 4:43 pm

        The difference between oikos and firm is that in your classic oikos consumption is mostly in-kind whereas with firms in a modern market you get paid in cash and then buy what you want on the market. Another way to put it is that both firms and oikos are islands of command in a sea of commerce, but the inhabitant of an oikos visits the shore a lot less than the worker in a firm. Max Weber for instance was all about bureaucracy but he still saw it as distinctive that now we get a paycheck whereas formerly we would “eat from the master’s table.”

        I never said there was anything normatively attractive about patronage (or primitive communism). For what it’s worth, I also like market societies, I just think it’s a stretch to characterize most pre-modern societies as being oriented around markets.

        For the most part I think we’re arguing semantics. We agree that pre-modern societies had things that were meaningfully understood as some kind of exchange, it’s just debatable how meaningful it was that a relatively low proportion of this exchange took the form of cash transactions.

      • 9. Graham Peterson  |  December 4, 2012 at 4:59 pm

        Considering semantic definition is the crux of trouble between economics and economic sociology, I don’t see that such a debate is trivial.

        I have seen no evidence that money corrupts social exchange, in fact much the opposite, as offers in Ultimatum and Dictator games converge to 50/50 in direct proportion to the extent of market exchange of the societies the experiments are run in. And therefore I do not see that intricately specialized wage labor reduces the degree of Oikos in the firm. Nor that monetary transactions anywhere else reduce it.

        “Both firms and oikos are islands of command in a sea of commerce.” Not so. Not at all. Take for instance the amount of time the modal person in a modern society actually spends conducting monetized exchange. Some errands after work? Even among middle and upper management — the proportion of the number of deals-made to other Busyness is very low.

        Herbert Simon’s famous little thought experiment was that if aliens arrived at Earth with a machine that mapped market transactions as green lines and non-market transactions as red blobs — the earth would look like a mass of red with itsy bits of green hashes connecting them.

        This, at least for me, is a central issue in my work. It needs to be shown empirically that a majority of behavior *in* the market is properly defined as “non-market,” that is not submitting the the theory of games or constrained optimization.

        Graeber is a hack. <– Sorry, I just couldn't resist it. I know it's not scholarly.

      • 10. gabrielrossman  |  December 4, 2012 at 5:12 pm

        You seem to be under the misapprehension that you’re talking to Michael Sandel or something. I don’t think markets “corrupt” (with some minor exceptions, like prostitution) although I do think they change things in nontrivial ways, often for the better. For instance, I would say that there are both differences and similarities between a farm worked by wage labor and one worked by slaves, serfs, or peons. Similar in the sense that even a slave has to be compensated at subsistence and may have an opportunity cost to his owner, different in the sense that wage labors are treated with greater dignity, can more easily demand producer surplus, can more easily reallocate themselves to a more productive role, etc.

        You yourself were talking about productivity gains and inequality reductions with growth in markets. That’s a difference. Likewise, if you go with Fiske’s relational models theory you see differences in how people understand exchange depending on how its framed.

        Also, FWIW, I studied w Zelizer in grad school and recently have been getting more interested in these issues.

        Finally, I have some reservations about “Debt” (specifically, I think chapter 12 is paranoid) but I still think it’s a valuable book with a lot to offer.

  • 11. Graham Peterson  |  December 4, 2012 at 3:53 pm

    Not to argue that supply and demand negate the validity of a volume of other models of behavior hinging on interdependency, and doesn’t itself need amending with such models. But ‘chu know.

  • 12. Graham Peterson  |  December 4, 2012 at 5:29 pm

    Similar in the sense that even a slave has to be compensated at subsistence and may have an opportunity cost to his owner

    The pecuniary welfare of slaves in fact went down after abolition, though existential utility arguments (I really ought to let go of arguing that way) most likely went up. Racism seems in fact to have gotten much worse after abolition, and excluding people from voluntary trades is as easy as, well, not volunteering to trade with them.

    To destroy any fiber of ethos I had here: I have not yet read Graeber’s book. I understand the thesis to be that States created money as an efficient device to tax and oppress people. Actually I take that back — I’ve read citations and quotations of it elsewhere. Those indicate that he attempts to take-down introductory economics textbooks. That is the sort of behavior that gets economic anthropology, and economic sociology laughed at. Indeed I clicked the link on Rojas’ blog to some economics sociology blog that was attacking the assumption of perfect information in economics, proffering that it needed to become more complete. The author must have missed that Joseph Stiglitz, George Stigler, and George Akerlof all received nobel prizes for dealing with information (poorly, in my opinion, but still).

    Money predates states. I mean, it’s just incontrovertible fact. I suppose I’ll have to go in and do the ancient economic history properly because the work economic history grads are doing these days almost exclusively turns on margins where data are available for multivariate analysis. But the idea that money is an efficient solution to coincidence of wants and a store of value is not a just-so story.

    All of this trade-in-kind and reciprocity nonsense has got to stop. It’s a social advocacy program for the hyper left academy. As I said — reciprocity everywhere and already pervades modern markets in the systematic dynamics of organizational evolution. High-school treatments of economics which attempt to show that trade is zero-sum, and therefore always exploitative, that the State created markets as an extension of its hegemony, get us nowhere. I understand I’m not arguing against Sandel. I’m venting some, maybe. And you cited Graeber, who if only for his Occupying infuriates me (not for having a problem with plutocracy, but selling such a ridiculous theory of it to impressionable pot smokers).

    Polanyi and Graeber just do not have their facts straight. Unfortunately secular mutations of Garden of Eden creation stories will never cease to emerge from Judeo-Christian descended cultures, nor the secular conclusion of such an Abrahamic vision — that one day we will all die (revolt) and go to heaven (socialism).

  • 13. gabrielrossman  |  December 4, 2012 at 5:44 pm

    As to Graeber, all I can say is read the book without worrying too much about his connection to OWS. He’s not at all naive about what primitive societies are like and I didn’t see him as engaging in Rousseauian nonsense. One of the points of the book is a critique of the “double coincidence of wants” problem but this isn’t a straw man, I checked and it does actually appear in econ textbooks.

    I understand why you think that talking about gift exchange, etc, is a polemic against the market since a lot of people treat it that way but that’s not my view. I see it just as a way to appreciate the rich variety of human life.

    >Indeed I clicked the link on Rojas’ blog to some
    >economics sociology blog that was attacking the
    >assumption of perfect information in economics

    If you’re talking about Soc2Econ that’s a satire that Shamus Kahn, Fabio Rojas, and myself wrote to make fun of sociologists who make silly straw man caricatures of economics.

    • 14. Graham Peterson  |  December 4, 2012 at 5:57 pm

      I don’t have time to reply in detail about dual coincidence of wants, but I didn’t say it was a straw man he set up — I said it’s a correct description of the emergence of money. And yes, it is included in every introductory and intermediate macroeconomics textbook written.

      “If you’re talking about Soc2Econ that’s a satire that Shamus Kahn, Fabio Rojas, and myself wrote to make fun of sociologists who make silly straw man caricatures of economics.”

      Thank God. But you should understand that Graeber’s reading of mainstream economic history is one such straw man. Check out the first three chapters of McCloskey’s forthcoming Bourgeois book on her website – the draft is available. She deals with Polanyi handily.

      Note also that the Medici and Hanseatic Merchants were thoroughly market-ized societies well before the Industrial Revolution. As was, I forget the name, the Icelandic Free-State the anarcho-capitalists are always talking about (private court system). It’s difficult to appreciate just how much trade took place in Rome, among the Greek city states, in China, and everywhere else.

  • 15. Graham Peterson  |  December 4, 2012 at 6:14 pm

    Nice chat.

  • 16. Graham Peterson  |  December 4, 2012 at 11:09 pm

    Reflecting:

    I see how the above looks like I was arguing (with a large fire in my pants) that markets have occupied a historical constant of social space since The Beginning. And I admit that that would be a stupid thing to argue. I guess I was saying that markets were in fact considerable, and not marginal in the past, and that the material amount of trade in history has been systematically played down by Marxian historians (also of note is that just about every philosopher in history despised markets – so reading the history of *thought* would have given early modern and even later scholars the impression that markets were a small part of the actual *material* history).

    An interesting question is the degree to which social space is finite or not (certainly constrained by time to interact — but can immaterial meaning accumulate without bound?). Also: to what degree do markets “encroach” on non-market social interactions in the Sandel and Commodification sense? To what degree do they potentially expand social space or scope by enhancing existential opportunity? And to what degree even given a fixed social space do they intermingle with “non market social interactions?”



%d bloggers like this: