Bryan Caplan  

Statistics is Not Right-Wing, But It is Elitist

PRINT
Medical Innumeracy in Super... Two New Books...

Here's a thought-provoking passage from Ian Ayres' Super Crunchers:

Like me, Ben Polak is passionate about the need to inculcate a basic understanding of statistics in the general public. "We have to get students to learn this stuff," he says. "We have to get over this view that somehow statistics is illiberal. There is this crazy view out there that statistics are right-wing." The stories in this book refute the idea that Super Crunching is part of some flattening right-wing conspiracy (or any other ideological hegemony).
Not so fast, Ian. There is a common theme that runs through the stories in Super Crunchers: anti-populism - or to be more positive - elitism. Yes, it is possible for statistics to support a populist view. But if you rely heavily on statistics, you will be forever double-checking whether the popular view is correct, and periodically saying it isn't. Statistics is anti-populist in the same sense that science is anti-religious: It questions dogmas instead of accepting them on faith. And like Darth Vader, most people find this lack of faith disturbing.

Furthermore, most people cannot and will never be able to use statistics to check their views. If you rely heavily on statistics, you effectively deny the competence of the man in the street to form his own opinions about the major questions of the day.

So what Polak and Ayres call the "crazy view" that statistics undermines popular ideologies is actually true. If we took the message of Super Crunchers to heart, we would lose our respect for the common man and the politicians who pander to him. Of course, if you are as elitist as I am, this is not a bug of Super Crunching, but a feature: As I told my lazy 12th-grade English teacher, "I'll give you all the respect you deserve."


Comments and Sharing





COMMENTS (8 to date)

Personally, I prefer the term "meritocratic". Although it doesn't have the same virtue of brevity as "elitist", it seems to get the same idea across without the baggage associated with elitist.

"Elite" can refer to people who are elite by virtue of their competence, but it also often refers to people who are elite by virtue of their membership in certain social cliques, which many people justifiably resent.

Steve Sailer writes:

Historically, the development of statistics is closely linked with the study of heredity. At least three crucial figures in the development of the tools Ian Ayres uses -- Francis Galton, Karl Pearson, and Ronald A. Fisher -- were leading advocates of eugenics.

Buzzcut writes:

Here's the problem in a nutshell:

I'm an engineer. I've been exposed to statistics in college, and I work with them every day in my job.

Yet, when you get into the hard statistics of, say, Freakonomics, you've gone somewhere that even I don't fully understand.

I've downloaded Levitt's papers, and I can't make heads or tails of what he's doing. I think that, as a result, his conclusions are suspect.

Once you get beyond simple linear regression, which can be done in Excel, you are into something that the average person simply can't fathom. No amount of education is going to change that.

Floccina writes:

Sorry for being off topic but...
I think that more libertarianism will only come if and when it becomes impossible for government to enforce the laws.

mk writes:

Interesting point but a few objections/responses:

1) It's important to have a definition of populism. Is populism "whatever the people want and emotionally respond to?" Is it "a type of politics railing against the upper classes?"

2) On second, thought, maybe it's not important because you could replace "populism" in your post with any other ideology (conservatism, liberalism, libertarianism, etc.). Empiricism (or "statistics" as you prefer to call it) will sometimes be at odds with any ideology, because none of them have a monopoly on truth.

3) However, statistics/empiricism can't settle all moral debates. Or at least, it is not trivial that it can. Thus, a large part of populism (or any other ideology) remains intact because ideology is typically defined in part by moral concerns.

4) Elitists, broadly construed, hold sometimes-incorrect views just as the "common man" does. So the empiricism you are advocating is an example of a particular strand of elitism -- "empirical elitism," you could call it -- which is at odds with every other formulation of elitism.

5) Having not read Super Crunchers I can't respond to the argument that the common man could never understand the use of statistics involved. Regardless however, arguments of the kind "most of the time, X results in Y" are readily understandably by anyone. Many if not most statistical arguments can be stated in this or similar form.

Buzzcut writes:

arguments of the kind "most of the time, X results in Y" are readily understandably by anyone. Many if not most statistical arguments can be stated in this or similar form.

I don't disagree. I mean, if you actually read Freakonomics, not Levitt's papers, he has a wonderful narative. He can put words to the numbers and make a story that is quite compelling.

But at some point someone needs to check that narative against the numbers. I am a Levitt skeptic. I could not make heads or tails of the numbers. That is problematic.

What Levitt is really saying is, "trust me". Sorry, that isn't good enough.

Steve Sailer writes:

Right. Levitt's most famous theory -- that legalizing abortion cut crime -- was exposed Foote and Goetz a half year after Freakonomics came out as being based on two technical errors Levitt had made in his programming.

Steve Sailer writes:

You could argue the exact opposite -- that well-done sophisticated statistical analyses (i.e., not Levitt's Abortion-Crime study) typical confirm popular prejudices, which the elite abhor.

Comments for this entry have been closed
Return to top