The big problems with social science—The idea of a social science, part 1


Image from:
Image from:

With this article I’m launching an occasional series about the social sciences and their place in a democratic society. This is a huge topic, and I’m not going to try to cover it all at once. In this first post I’m only going to introduce the topic and mention some really big picture thoughts. I’ll have more specifics soon.

To be fair, I should acknowledge up front that I am skeptical/critical about a lot of social science, for reasons that will be discussed throughout this series. I am also not a trained social scientist, so I make no claims to expert knowledge. I do, however, have some experience that I think gives me a relatively informed perspective. I will do my best, but I am sure I’ll get things wrong. I enthusiastically welcome any corrections or alternative perspectives in the comments.

The title of this series, “The idea of a social science” is taken directly from Peter Winch’s 1958 book, The Idea of a Social Science and its Relation to Philosophy. I don’t agree with all of his arguments, but this book definitely started me thinking about these questions, and it’s a good read.


Why write about social science?

Social sciences are fundamental to our society; we make public policy decisions based on data about people that is gathered, analyzed and interpreted by social scientists and the work they produce is a major part of how we understand our society. Our biggest political debates, about issues like welfare, immigration, marriage equality and foreign policy, are all based on theories and assertions about how humans behave in different circumstances—i.e. “raising the minimum wage increases unemployment”, “student assessments improve education” or “gay marriage is good/bad for children”. These are all verifiable statements, and social scientists (sociologists, anthropologists, economists, psychologists, political scientists…) should be qualified to either confirm or refute them.

This is how things work in natural sciences anyway. For example, there was a time when biologists argued about what “genes” were made of. It was clear that some chemical in cells must be carrying heritable information that served as a blueprint for growth and development, but nobody knew what that chemical was. Most biologists assumed genes were made of protein because of the chemical complexity of protein molecules. We all know now, however, that the much simpler DNA molecule is what genes are made of. When the chemical nature of genes was resolved, it happened quickly and biologists who had been “protein supporters” accepted the DNA result and moved on to other things. This is how sciences are supposed to develop—over time more evidence is collected, bad ideas are rejected, more of the truth is discovered and big questions get answered.

Big questions about human behavior, however, don’t seem to resolve. For example, for many decades economists have been arguing about whether government benefit programs (welfare, food stamps…) affect employment rates, by de-incentivizing work. This debate shows no signs of ending and almost nobody ever admits they are wrong.

Every day on CNN and FOX, and even in testimony before Congress, you can see economists with PhD’s using objective, ‘sciency’ language to make contradictory predictions about the effects of public policies. Because of their training, these folks are considered experts and because of the ‘sciency’ language they use, their opinions are treated as facts. The statements these experts make, however, are not facts, they are opinions and theories that often have very little empirical support—because it’s simply not possible to use the scientific method to test most economic theories (this is the topic of the next post in this series). Consequently, I believe that the work of economists should not properly be called ‘science’ and that doing so has negative effects—it both muddies our public debates and damages the overall reputation of science.

This misuse of the language and reputation of science undermines public trust in science. When irresponsible scientists, or people calling themselves scientists, get public predictions wrong, people start to doubt the whole field. This problem isn’t unique to the social sciences, however—for example, while I personally accept the overwhelming consensus that anthropogenic climate change is a real phenomenon and ought to concern us, it’s also clear that specific predictions of disaster have been oversold. As a kid, I honestly thought that, by now, most animal species would be extinct and we would be living in a hot, flooded, toxic waste dump. Of course none of this happened—the most dire predictions have not occurred and nature appears to be somewhat more resilient than we thought.

The unfortunate result of these “cry wolf” predictions is that many, many people believe that climate scientists are just alarmists who are making stuff up. This assumption feeds directly into Limbaugh-esque narratives about how the whole field is just a cover for anti-capitalist, un-American, economic reforms. Thus, actual progress on climate change is now politically more difficult and the problem is worse.

Misusing the language of science also distorts debate by invalidating other good opinions. Obviously, some folks are experts about certain issues, and their ideas on these subjects should be considered to be more credible, or at least more informed. But, when opinions are wrapped in the language of science (charts, numbers, statistics…) they assume a veneer of objectivity they don’t deserve. This stifles public debate.

Overselling the objectivity and confidence of social science has major real-world effects; for example, nearly every economist with a national reputation failed to anticipate the housing bubble that wrecked our economy eight years ago. As Alan Greenspan, then chair of the Fed and the nation’s most prominent economist, said at the time:

“a ‘bubble’ in home prices for the nation as a whole does not appear likely….Although we certainly cannot rule out home price declines, especially in some local markets, these declines, were they to occur, likely would not have substantial macroeconomic implications.”

Despite this enormous failure to understand and predict major events in their field*, economists like Greenspan and Ben Bernanke are still treated like authorities on the topic. They can both be found on TV regularly, engaging in policy debates, and making empirical claims about the economy that are really just educated guesses. They speak confidently and include specific numerical predictions, and their policy recommendations are treated as facts**. This distorts the debate because the opinions of others (either folks from less empirical disciplines who don’t claim specific numerical predictions, or economists who are more humble and unwilling to oversell their knowledge) are granted less authority and we end up with something like “tyranny of the hubristic”.


So, to sum up, I believe there are deep reasons, both practical and fundamental, why social sciences don’t settle questions and progressively move toward the truth, the way that natural sciences do. Consequently, I will argue they shouldn’t really be considered sciences. But my concerns about this subject aren’t just theoretical; I believe that properly understanding the nature and limitations of social science will lead to better political discussion and ultimately to better public policies.


What’s ahead in this series: I’m not exactly sure how long this series will last or how often I’ll post on this subject, but suffice it to say that I have a lot of ideas. I have already outlined several more posts on the following topics:

Measuring humans is hard and impossible: This post will discuss some of the issues involved in measuring and predicting human behavior. I’ll discuss some practical limits that make it very, very difficult collect good data about people (cost, ethics…) as well as fundamental, epistemological problems with measuring human behavior that don’t exist in other sciences. I’ll also discuss examples of where I believe social science can be done well and how meaningful ‘experiments’ can actually be conducted on humans.

Reproducibility and fraud: I’ll discuss some of the major academic fraud scandals that have recently engulfed some fields of social science, particularly social psychology. These scandals illuminate what I consider to be some of the fundamental flaws in social science methodology. I also am personally interested in scientific ethics and fraud in general, and I’ll probably broaden the discussion to include some thoughts about fraud in research about vaccines and autism. I’ll also discuss how bad research results can persist as ‘facts’ in society and how difficult it can be for good scientists to dislodge them from the public consciousness.

Can big data save social science? I’m personally very interested in the idea of ‘big data’ and how we can best learn from the enormous amounts of information that are suddenly available because of advances in information technology. I’ll discuss a bit of my own research here and talk about how ‘big data’ is currently causing a huge revolution in biology. I’ll consider how social sciences may be able to benefit from this data explosion. I’ll also discuss some of the limitations of ‘big data’ and how what we are measuring isn’t necessarily what we want to know.

Stay tuned.


* Side note: This event, the glaring failure of nearly every esteemed economist to predict the housing crash, is really what got me hot-and-bothered about economics and social sciences in the first place. It reminded me a lot of the claims that the Bush administration so famously got wrong about Iraq: “we know where the WMD’s are”, “we’ll be greeted as liberators”. Except, I expect politicians to be full of it. But I had thought that economists knew what they were talking about and had a pretty good predictive understanding on how our economy worked. After the housing crash played out, it was abundantly clear to me that economists didn’t really know what they were talking about and we shouldn’t have let Greenspan’s cockiness lull us into complacency.

** To be fair, both Bernanke and Greenspan are excellent politicians and thus have mastered the vague ‘fedspeak’ style that generally avoids specifics. They had to get something really, really wrong before it was obvious that they were full of crap. The worst examples of what I’m describing are common in the media, but usually come from lesser-known economists, trying to establish themselves as media personalities.


2 thoughts on “The big problems with social science—The idea of a social science, part 1

  1. An interesting post. I think your skepticism about the social sciences is understandable, but saying that they’re not science is a bit excessive. Social sciences will never have the certitude that is often possible in the natural sciences, mostly because tightly controlled experiments are often simply not possible. But then meteorology doesn’t have the certitude of physics. Reduced certitude doesn’t mean that information from the social sciences isn’t useful.

    I’d argue that Bernanke’s and other policy maker’s understanding of the history of the Great Depression and how monetary and fiscal policy can affect the economy made the Great Recession far less severe than it would have been. Of course, what “would have been” is a counterfactual, so it’s impossible to prove to a determined skeptic. But any open minded person familiar with the policy and economic history of 1928-1933 and 2008-2012 can see the differences without arcane effort.

    I think the biggest problem economics has isn’t that it can’t be scientific. The biggest problem is that too many people who are allowed to call themselves “economists” simply ignore the science in favor of their ideological predispositions, and it’s difficult to judge who is who without learning a lot about economics. That’s certainly a problem for the profession, but not necessarily for the science itself.


    1. Thanks for the thoughtful detailed comment. You make several good points, and I agree with a lot of what you say. I am painting with a broad brush in this article, but I plan to explore in more detail many of the issues you raise. Also, I agree with you that informed intervention can (and probably did in the example you cite) positively effect the economy. I’m not anti regulation or intervention.

      I think the biggest issue is the point you make in your first paragraph: “tightly controlled experiments are often simply not possible.” and your comparison to meteorology is apt. I’m planning to explore that issue in a lot more depth in the post in this series.

      One last point–please don’t think of me as “a determined skeptic”, I’m totally open to the fact that I might be wrong and I’m happy to learn more from informed folks. Thanks for taking the time to engage and I hope you come back and give feedback on the rest of the series.

      Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s