economics and sociology, part cdlxvii: comments on a blog post by noah smith

with 15 comments

A few weeks ago, economics columnist Noah Smith wrote a blog post about how economics should raid sociology. This raises interesting questions about how academic disciplines influence each other. In this case, why has sociology not been a good a receptor for economics?

I start with an observation, which Smith also alludes to: Sociology has already been “raided” by economics with only moderate success. In contrast, economists have done very well raiding another discipline, political science. They have done fairly well in establishing pockets of influence in public policy programs and the law schools. By “success,” I do not mean publishing on sociological topics in economics journals. Rather, “success” means institutional success: economists should be routinely hired sociology programs, economic theory should become a major feature of research in graduate programs, and sociological journals should mimic economics journals. All of these have happened in political science but not sociology.

Here’s my explanation – Sociology does not conform to the stereotype that economists and other outsiders have of the field. According to the stereotype, sociology is a primarily qualitative field that has no sense of how causal inference works. In some accounts, sociologists are a bunch of drooling Foucault worshipers who babble endlessly in post-modern jargon. Therefore, a more mathematical and statistical discipline should easily establish its imprint, much as economics is now strongly imprinted on political science.

The truth is that sociology is a mixed quantitative/qualitative field that prefers verbal theory so that it can easily discuss an absurdly wide range of phenomena. Just open up a few issues of the American Sociological Review, the American Journal of Sociology or Social Forces. The modal article is an analysis of some big N data set. You also see historical case studies and ethnographic field work.

It is also a field that has import traditions of causal identification, but does not obsess over them. For example, in my department alone, there are three faculty who do experiments in their research and one who published a paper on propensity scores. Some departments specialize in social psychology which is heavily experimental, like Cornell. There are sociologists who work with data from natural experiments (like Oxford’s Dave Kirk), propensity scores (like IU’s Weihua an), and IV’s (I actually published one a while ago). The difference between economics and sociology is that we don’t reward people for clever identification strategies or dismiss observational data out of hand. If possible, we encourage identification if it makes sense. But if an argument can be made without it, that’s ok too.

So when economists think about sociology as a non-quantitative field, they simply haven’t taken the time to immerse themselves in the field and understand how it’s put together. Thus, a lot of the arguments for “economic imperialism” fall flat. You have regression analysis? So does sociology. You have big N surveys? We run the General Social Survey. You have identification? We’ve been running experiments for decades. One time an economist friend said that sociology does not have journals about statistical methods. And I said, have you heard of Sociological Methodology or Sociological Research and Methods? He’s making claims about a field that could easily be falsified with a brief Google search.

In my view, economics actually has one massive advantage over sociology but they have completely failed to sell it. Economists are very good at translating verbal models into mathematical models which then guide research. The reason they fail to sell it to sociology is for a few reasons.

First, economists seem to believe that the only model worth formalizing is the rational actor model. For better or worse, sociologists don’t like it. Many think “formal models = rational actor model.” They fail to understand that math can be used to formalize and study any model, not just rational choice models.

Second, rather than focus on basic insights derived from simple models, economists fetishize the most sophisticated models.* So economists love to get into some very hard stuff with limited applied value. That turns people off.

Third, a lot of sociologists have math anxiety because they aren’t good at math or had bad teachers. So when economists look down at them and dismiss sociology as whole or qualitative methods in particular, you loose a lot of people. Instead of dismissing people, economists should think more about how field work, interviews, and historical case studies can be integrated with economic methods.

I am a big believer in the idea that we are all searching for the truth. I am also a big believer in the idea that the social sciences should be a conversation not a contest of ego. That means that sociologists should take basic economic insights seriously, but that also means that economists should turn down the rhetoric and be willing to explore other fields with a charitable and open mind.

50+ chapters of grad skool advice goodness: Grad Skool Rulz ($2!!!!)/From Black Power/Party in the Street

** For example, I was once required to read papers about how to do equilibrium models in infinite dimensional Banach spaces. Cool math? Sure. Connection to reality? Not so sure.

Written by fabiorojas

May 18, 2016 at 12:15 am

three paths in sociology

leave a comment »

After reading a good deal in the sociology of race last month, I appreciated that progress in sociology happens in three ways:

  • Problem solving/topic concentration: Here, you choose a topic, say, urban poverty, and you attack it from multiple directions. In the course of solving the problem, you arrive at insights. It reminds me of what Richard Feynman said about becoming a genius. It’s not too hard. Just write down a few problems that would be amazing to solve. Then, every time you learn something new, apply it to that problem. Sooner or later, you’ll solve the problem and everyone will say you’re a genius.
  • Synthesis: There’s a reason we do research – we want publish our results so other people can build on it. So you make intellectual progress by reading tons and tons of empirical studies in your area and see what the major findings are.Add water and mix.
  • High theory: Ignore the tons of research that has come before and instead apply abstract and broad theories to new topics and see how far you get. The benefit is that the nitty gritty of normal science tends to do different things that high theory, so you are bound to say a lot of original things.

Personally, I normally operate in the problem solving mode of research, which is why I am viewed as an atheoretical researcher by some. While this holds a kernel truth, it’s not quite right as I am doing “ground up” theory. I really want to learn how things work before moving onto broad theory. I am also not a “grounded theory”/induction type of person. I don’t believe theory just speaks for itself. What I do believe is that theory needs constraint from data and you just can’t cheat your way out of that requirement. You have to really, really learn how the world works before you can do that.

50+ chapters of grad skool advice goodness: Grad Skool Rulz ($2!!!!)/From Black Power/Party in the Street

Written by fabiorojas

May 17, 2016 at 12:01 am

new york, new york

leave a comment »

Hey, orgheads! I will be in New York this coming Saturday and Sunday. If you want to hang out, email me. I have to give a talk on Saturday afternoon, but other times might be open.

50+ chapters of grad skool advice goodness: Grad Skool Rulz ($2!!!!)/From Black Power/Party in the Street

Written by fabiorojas

May 16, 2016 at 3:28 am

Posted in uncategorized

sykes-picot is 100 (with a nod at fanon at the end)

with 3 comments

As we approach the 100th anniversary of Sykes-Picot, some interesting analysis (and defenses) from around the web:

From the Economist (they have a larger section but you have to be a subscriber):

A second wrong-headed notion is that redrawing the borders of Arab countries will create more stable states that match the ethnic and religious contours of the population. Not so: there are no neat lines in a region where ethnic groups and sects can change from one village or one street to the next. A new Sykes-Picot risks creating as many injustices as it resolves, and may provoke more bloodshed as all try to grab land and expel rivals. Perhaps the Kurds in Iraq and Syria will go their own way: denied statehood by the colonisers and oppressed by later regimes, they have proved doughty fighters against IS. For the most part, though, decentralisation and federalism offer better answers, and might convince the Kurds to remain within the Arab system. Reducing the powers of the central government should not be seen as further dividing a land that has been unjustly divided. It should instead be seen as the means to reunite states that have already been splintered; the alternative to a looser structure is permanent break-up.

From The New Yorker:

For a century, the bitter reaction to the Sykes-Picot process has been reflected in the most politically powerful ideologies to emerge—Nasserism, in Egypt, and Baathism, in Iraq and Syria—based on a single nationalism covering the entire Arab world. For three years, Egypt and Syria, despite being on different continents, actually tried it, by merging into the United Arab Republic; the experiment disintegrated after a 1961 coup in Damascus.

Even the Islamic State seeks to undo the old borders. After sweeping across Syria and Iraq in 2014, Caliph Abu Bakr al-Baghdadi announced, “This blessed advance will not stop until we hit the last nail in the coffin of the Sykes-Picot conspiracy.”

From The New York Times:

That Western imperialism had a malignant influence on the course of Middle Eastern history is without a doubt. But is Sykes-Picot the right target for this ire?

The borders that exist today — the ones the Islamic State claims to be erasing — actually emerged in 1920 and were modified over the following decades. They reflect not any one plan but a series of opportunistic proposals by competing strategists in Paris and London as well as local leaders in the Middle East. For whatever problems those schemes have caused, the alternative ideas for dividing up the region probably weren’t much better. Creating countries out of diverse territories is a violent, imperfect process.

From Foreign Policy:

The “end of Sykes-Picot” argument is almost always followed with an exposition of the artificial nature of the countries in the region. Their borders do not make sense, according to this argument, because there are people of different religions, sects, and ethnicities within them. The current fragmentation of the Middle East is thus the result of hatreds and conflicts — struggles that “date back millennia,” as U.S. President Barack Obama said— that Sykes and Picot unwittingly released by creating these unnatural states. The answer is new borders, which will resolve all the unnecessary damage the two diplomats wrought over the previous century.

Yet this focus on Sykes-Picot is a combination of bad history and shoddy social science. And it is setting up the United States, once again, for failure in the Middle East.

There’s a lot more.  This is an important anniversary.

The Foreign Policy article above is, at least for me, the more interesting one, especially as it ties into important sociological conversations about the invention of tradition and imagined communities.

For what it’s worth, it’s of course true that Sykes-Picot can sometimes get too much blame (or at least be given too much causal power) for the entirety of the problems in the Arab world (and the broader Middle East). For that matter, I’m quite sure there are many other borders that would have been just as bad.  Yet it’s sometimes easy to forget that even more than the lines themselves, it was the imperialist capacity to render those lines that has caused so much anger.  Someone like Fanon (and postcolonial theory more broadly) helps show there’s something important about the power dynamics in which you are named and recognized, and sometimes discussions of Sykes-Picot (and, for that matter, talk of the United States drawing up the map again) utterly ignore this distinction, framing a problem of recognition as simply a problem of categorization. Also for what it’s worth, I’ve written a bit about how Edward Said helps us think about this stuff here.

Addition: My colleague Kevan Harris pointed me to the really impressive work of Sara Pursley.  Here’s an interview with her about Sykes-Picot, among other things:

But I’m not just contesting the Sykes-Picot narrative. I’m contesting all the narratives that say Iraq’s borders were “drawn” by Europeans in the years around World  War I, whether they locate that moment in Sykes Picot, or the Paris Peace Conference, or San Remo, or the Cairo Conference. These last three tend to be more popular with scholars and Iraq experts, who often know that Sykes-Picot doesn’t really work. But actually none of them work. The supposed map the Europeans drew of the Middle East—it doesn’t exist. Iraq’s borders were created like most nation-state borders have been created, through a drawn-out of process of resolving competing claims to territory through war, diplomacy, and other uses of power. It took many years and involved many actors. To begin with, a border requires mutual recognition of the authorities on both sides—that’s what a border is. You can’t just create one by yourself.

Written by jeffguhin

May 15, 2016 at 12:22 am

orianthi/beat it

leave a comment »

50+ chapters of grad skool advice goodness: Grad Skool Rulz ($2!!!!)/From Black Power/Party in the Street

Written by fabiorojas

May 15, 2016 at 12:01 am

sociology in the new york times

with 2 comments

More sociologists in prominent publications!

This morning I posted about two recent articles from sociologists in The Atlantic, and then Beth Berman reminded me about Corey Fields‘s recent piece in The New York Times on black Republicans:

A common thread that links high-profile black Republicans like Mr. Carson is their commitment to “colorblind” politics. You can also see this with recently elected politicians like Representative Mia Love of Utah and Senator Tim Scott of South Carolina. They acknowledge that being black is part of their life experience but reject the idea that racial identity should orient their political decision making. They often decry efforts, like affirmative action, to address racial inequality explicitly, claiming that such policies undermine black success.

Though colorblind black Republicans get a lot of attention, they fail to represent an important, if overlooked, type of black Republican, which I call “race conscious.”

And then, in the Times today, this piece from Neil Gross on why the highly educated are so liberal:

What explains the consolidation of the highly educated into a liberal bloc? The growing number of women with advanced degrees is part of it, as well-educated women tend to be especially left-leaning. Equally important is the Republican Party’s move to the right since the 1980s — at odds with the social liberalism that has long characterized the well educated — alongside the perception that conservatives are anti-intellectual, hostile to science and at war with the university.

I do hope this trend continues.  If you know of something or someone else I should mention here, please e-mail me at or contact me on twitter.

Written by jeffguhin

May 13, 2016 at 11:38 pm

Posted in uncategorized

sociology in the atlantic

with 3 comments

Two articles about sociology in four days, written by the sociologists themselves! These both look like great books.  It’d be wonderful to see more projects like these excerpted/previewed in The Atlantic or similar magazines. Obviously Matt Desmond was working with a publisher that made those kinds of connections a bit easier, but I don’t see why this can’t be a more standard practice. It’s great for the authors, the discipline, and the reading public.

To the work! First, a few days ago Steve Viscelli had a piece related to his research on truckers, and the way that companies convince drivers to become contractors usually against their best interests:

So, truck driving has its appeal—for a few months anyway. The money is better than many workers can earn in service jobs, but in most cases that’s only because they are working the equivalent of two full-time jobs. There are laws that are supposed to limit how much drivers work, and they record their hours on paper or in an electronic log. The rules are complicated, but limit drivers’ working hours to roughly 60 hours per week. But those laws aren’t effective, partly because 90 percent of the typical driver’s compensation is based on the number of miles they drive, so they only count the hours they absolutely have to. While the number of hours they spend driving on public roads is counted relatively accurately, many, many of the hours drivers work are spent waiting while their trucks are loaded and unloaded, and doing all kinds of unpaid, non-driving work that ranges from filling out paperwork to fueling their truck. Considering all the unpaid work drivers are putting in, they tend to earn little more than minimum wage. Over the four months I spent on the road, I averaged less than $10 per hour worked.

Then, today, Laura Hamilton has a piece related to her research on how higher ed uses helicopter parents:

Universities now rely, in part, on parents, particularly those with money, time, and connections, to meet their basic needs. Solvency is the most pressing one—net tuition now accounts for 47 percent of all public higher-education revenue, so schools necessarily prefer applicants who don’t require financial aid. Most public institutions, like the one I studied, are not need-blind, and take student funding into account. They particularly value out-of-state and international families who pay top dollar.

However, paying parents typically bring more than funds alone. They often help promote the university; conduct admissions interviews; interface with donating alumni; assist with their own students’ emotional, cognitive, and physical needs; and help place graduates (both related and not) in valuable internships and jobs. Competition to attract these parents is stiff—and administrators’ complaints about parental “meddling” are now tempered with interest in a “partnership relationship” with parents. As such, four-year schools structure their classes, activities, and living options around traditional students and expect parents to do the work of maintaining them, even as the financial, physical, and emotional costs of doing so continue to escalate.


Written by jeffguhin

May 13, 2016 at 5:01 pm

Posted in uncategorized


Get every new post delivered to your Inbox.

Join 3,648 other followers