The journalist's fallacy

The journalist's fallacy is the name I gave to the habit of drawing strong inferences from only one or two data-points. We've seen two nice examples of it this week, from left and right.

From the left, we have Seamus Milne. He points to Manchester University students who are campaigning for a more pluralistic economics curriculum as a sign that "orthodox economist have failed their own market test. Granted, he could have widened the evidence base; there's a similar movement at Cambridge, at least. But he omits two facts:

- Demand to study ("orthodox") economics at university has risen 33% in the last six years, whilst applications to university generally are up by only 19%.

 - At least two universities  - Oxford and Manchester; I know coz I took 'em - used to offer courses in the history of economic thought, but dropped them in part for lack of demand.  

These two facts suggest that demand for "orthodox" economics is strong. It's too damned strong in my book, but things are as they are, not as we'd like them to be. (And anyway, the truth or not of "orthodox" economics is not to be decided by what undergraduates think.)

From the right, we have Melanie Phillips' attempt to tell us that the Paul Flowers affair tells us something about the broader liberal-left. But she doesn't point that many Labour and Coop members are not financially illiterate drug-takers, nor that - perhaps! - some on the right might be. She doesn't even try to fill in the Bayesian boxes. Granted, she might have cited Hayek's "Why the worst get on top", but that could apply - as he says - to "any society."  

There are thousands of coops in the UK, and only one (or a few) Paul Flowers. The question - which is of fundamental importance - is: are coops more efficient than capitalistic firms? This cannot be settled merely by pointing to any individual personal failure, but by wider statistical research.

Seamus and Melanie are committing a similar error. Both are inferring too much from individual examples - an error exacerbated by wishful thinking. They are neglecting the very useful advice offered in this great piece in nature - that "data can be dredged or cherry picked."

Now, you might object here that pointing out that Seamus and Melanie have limited powers of ratiocination is the epitome of lazy blogging. It is, but there's a more general point here. It's that Seamus and Melanie have successful journalistic careers whereas I don't. One reason (of several!) for this is that the lively human interest anecdote is more vivid and interesting than dull statistics; admit it, you'd rather read about Flowers' porn habit than about the relative productivity of worker- versus capitalist-owned firms. And journalism is about giving readers what they want, not about high-minded logic and evidence.

All of which makes me suspect that Noah Smith might have added something here.

Comments

[Martin S]

You're a Marxist and therefore a materialist.

A materialist reading is this: demand for orthodox economics at university is driven by what 18-year-olds think will result in gainful employment (history of economic thought courses don't impress Goldman Sachs).

Seumas and Melanie do this stuff because it's their job. They're paid to generate controversial opinions. They're not paid to be right or to get things in proportion. They're paid to pander to their readers' prejudices, not to respect logic and evidence. They're both good at what they do.

[pablopatito]

"admit it, you'd rather read about Flowers' porn habit than about the relative productivity of worker- versus capitalist-owned firms."

As a stumblingandmumbling fan, I can say no, we wouldn't. But we're obviously in a minority.

[Metatone]

IAWPP

(I Agree With pablopatito)

The problem is that micro payments either don't work yet, or can't work (depending who you listen to) and I'm unlikely to pay enough per blog to make credit card charges make sense...

[Paul Walker]

But as Henry Hansmann points out capitalist firms are just a form of cooperative.

[Andrew]

If they are giving readers what they want, they certainly don't belong on a list of non-value-adding rent-seekers!

[Tom Slee]

Orthodox economics also tells us that the demand for economics degrees is likely to be based on the value of the degree as a costly signal to future employers and has little to say about the intrinsic value of the degree (that is, its relation to reality, or its effectiveness as a tool in driving good policies). It is consistent with economic rationality to sign up for the degree, knowing it's a useful signal, and simultaneously to campaign for a change in the content of the courses to better fit politico-economic reality.

[Z]

Tom: That's a good explanation, but doesn't explain the fact that the kinds of courses these students are protesting for end up getting canceled from lack of demand.

Chris, glad to see I'm not the only one tired of the Guardian's relentless, poorly written anti-economics writing of late. Worst recently is the "Academics back students in protests against economics teaching" one, in which somehow a group of 7 unknown faculty is equivalent to "a prominent group of academic economists."

The variant on your journalist's fallacy, of course, is the other journalist's fallacy, where any source is believable as long as it sufficiently agrees with the journalist's priors.

[Ken Clark]

A small point of clarification: Manchester still offers a course module on the history of economic thought and student demand for this is strong (I am a former head of department at Manchester). It may not run this year in its usual form: that is nothing to do with demand but rather an internal staffing matter.

A slightly larger point, perhaps, which I don't actually agree with but in the interests of putting the "rebellious" students' case...

The growing demand for economics degrees could be interpreted as a demand for greater understanding of the post-crisis world in which we find ourselves. It's only once these students get to our (top) universities that they are disappointed to find that orthodox approaches fail to deliver. Hence the demand for curriculum change.

Wren-Lewis, while rightly critical of the semi-illiterate Guardian campaign against the discipline, notes some aspects of economics teaching which could benefit from change in his recent blog: http://mainlymacro.blogspot.co.uk/

[JSeydl]

I totally agree that journalists are too quick to make sensationalist claims based off limited data. But your Seamus example in a non-sequitar.

Just because more students are taking economics courses, it doesn't mean they are endorsing orthodox views. A first year student doesn't know about the divisions that exist in the discipline. It's not as though these students thought about the decision, and then decided to endorse the orthodox view. Indeed, I see the recent spike in enrollment as evidence that students want to know more about the economy, probably because its collapse has affected them greatly in recent years. Still, I think it's too early to tell whether orthodox teaching fulfills their thirst for knowledge.

Oh, and two more points. It's not just Manchester and Cambridge where student movements are forming. There was a huge student conference at LSE called Rethinking Economics, which was all about questioning orthodox teaching. This movement is also now growing in NYC quickly.

And second, it does absolutely matter what undergraduates think. Pressure from them will influence the curriculum. INET notices this truth, which is why it is largely molding its online curriculum based on what students want.

So I agree with you in principle, but you picked a bad example, which you don't seem to know much about.

[Z]

JSeydl: Not to put words in Chris's mouth, but the point of the Milne article is that somehow orthodox economcis is losing out by its own measurements because a group of students are demanding something different. Chris is saying, well, if you go down that rabbit hole, then there's far more students who seem perfectly content with what they're getting. Not a slam dunk, but the original piece is ridiculous to begin with, anyway.