1. Skip to navigation
  2. Skip to content
  3. Skip to sidebar
Source link: http://archive.mises.org/8173/empiricism-gone-mad-or-how-i-learned-to-love-regression-analysis/

Empiricism Gone Mad: or, how I learned to love regression analysis

June 5, 2008 by

Under the Weather: Health, Schooling, and Economic Consequences of Early-Life Rainfall is a recent offering from the National Bureau of Economics Research.

Who would have thought that the rainfall in your birth-year and birth-location determines (inter alia) your lung capacity, height, final education level, etc? Of course, there is the big caveat that this only holds for women born in rural Indonesia between 1953 and 1974. Regardless, it’s now empirical fact and ready to guide public policy — you understand, the rainfall gap and all.

The best I can assume is that authors Sharon L. Maccini and Dean Yang found the only remaining data sets that had not yet been correlated and had at it. Or, maybe they simply had government grant dollars to consume. Either way, their main contribution is additional proof that empiricism is nonsense.

{ 20 comments }

Fred Furash June 5, 2008 at 4:48 am

Isn’t this how the philips curve was “found”?

Koen Deconinck June 5, 2008 at 5:02 am

Actually, the original Phillips curve connected unemployment with changes in nominal (yes, nominal) money wages in the UK. That’s not completely useless, and you would indeed expect that there’s some connection between money wages and unemployment: if there’s a lot of unemployment, you’d expect wages to fall or at least not to rise as fast as they would when unemployment is lower. But, somewhere along the way, some economists replaced ‘changes in nominal money wages’ with ‘consumer price inflation’.

By the way, now that we’re talking about silly statistics:

http://www.nber.org/papers/w8158
KOPCZUK & SLEMROD, “Dying to Save Taxes: Evidence from Estate Tax Returns on the Death Elasticity”.

They were the proud receivers of the 2001 IgNobel Prize in Economics:
http://improbable.com/ig/winners/

Robert Wegner June 5, 2008 at 5:06 am

I’m currently reading “On the Accuracy of Economic Observations” by Oscar Morgenstern. I’m about half way through the book and amazed at how critical Morgenstern is of emperical methods from data collection, accuracy of data , “functionally false statistics and meaningless statistics” etc.

Inquisitor June 5, 2008 at 7:18 am

Morgenstern was an Austrian, no? He’s listed as an Austrian economist (not just in nationality.)

Garth Brazelton June 5, 2008 at 7:37 am

I think your attack on empiricim is misplaced. It’s not empiricism itself that is the fault, it is empiricism in the hands of researchers that jump to conclusions about causation without really having a formal theory to back up their work. That said, any blanket statement against empiricism seem out of hand to me.

Inquisitor June 5, 2008 at 7:53 am

Well empiricism is not synonymous with empirical research… it’s an entire epistemological position.

fundamentalist June 5, 2008 at 10:12 am

This is a perfect example of what statisticians teach you not to do—go fishing for correlations. That’s why real statisticians hate data mining and this is an example of the worst kind of data mining abuse. Real statisticians approach data with a hypothesis in mind and test that hypothesis with the data. This is very similar to Mises’s approach to data. Any results not predicted by the hypothesis should be ignored. You might cut some slack for an analyst who could invent a hypothesis to explain the cause/effect relations, but further testing would be required.

A similar problem occurs repeatedly with EMF radiation. Back in the 90′s some researchers found a statistically valid relationship between EMF radiation and cancer and caused a scare. But no scientist could explain how EMF radiation might cause cancer, so the statistical results were viewed as an anomaly, as they should. But like a boomerang, some journalist always uncovers the original statistical research and makes a run at it again.

The article requires coin, so I didn’t read it all. Do the researchers have a hypothesis as to how rainfall would affect lifestyles many years later? If not, the results are probably spurious and there are some tests that they could run to determine that.

Fephisto June 5, 2008 at 7:27 pm

This is not empiricism, this is what we call, “heuristics”.

KY Leong June 6, 2008 at 12:54 am

As Mr Jim Fedako learnt to love regression analysis, I have learnt to love this kind of “empiricism” for its rather strong appeal to one’s sense of humor.

Brainpolice June 8, 2008 at 6:06 am

I think that it’s “radical empiricism” (which places too much emphasis on and trust in the senses while completely dismissing the a priori out of hand) that is the problem, not necessarily empiricism in general.

Hornshiver June 14, 2008 at 8:36 am

Empiricism is a very useful technique to gather knowledge. But empiricism is seen negatively by some because irrational humans often abuse empirical techniques such as economics.

Without empiricism, newtonian physics, biology, and other useful knowledge would not be developed. Empiricism is useful if it is done correctly.

Jim Fedako seems that he wants to abolish all empiricism because he had inductively “proved” that all empiricical research is false just because he had seen some bad examples of empirical research. That is a kind of empiricism, even that he denies using it! Thus, Jim Fedako is contradictory.

Hornshiver June 14, 2008 at 8:45 am

Without empiricism — life-saving medicine, synthetic materials, technology, computer, would not be developed. Those who proved that all empirical science is incorrect just because there are some bad examples is a bigot.

Jim Fedako June 14, 2008 at 9:05 am

Hornshiver,

You are conflating empirical with empiricism. I suggest that you take a look at Hans-Hermann Hoppe’s Economic Science and the Austrian Method, Hayek’s The Counter-Revolution of Science, and David Gordon’s The Philosophical Origins of Austrian Economics. All available at the Mises Store.

Good luck on your intellectual journey!

Hornshiver June 14, 2008 at 1:02 pm

Jim Fedako, your suggestions are only criticisms on empirical methods on economics, not science in general.

Hornshiver June 14, 2008 at 1:02 pm

Jim Fedako, your suggestions are only criticisms on empirical methods on economics, not science in general.

Hornshiver June 14, 2008 at 1:02 pm

Jim Fedako, your suggestions are only criticisms on empirical methods on economics, not science in general.

Jim Fedako June 14, 2008 at 10:16 pm

Hornshiver,

You are conflating empirical research with empiricism. Until you accept the difference between the terms, there is nothing we can debate.

newson June 15, 2008 at 4:47 am

to hornshiver:

i’ve downloaded a free pdf copy of the counter-revolution in science from here:
http://www.archive.org/details/counterrevolutio030197mbp

jeffrey June 15, 2008 at 9:34 am

That’s very impressive. It was my understanding based on public records that this one had been (sadly) renewed.

Hornshiver June 15, 2008 at 11:36 am

I do not see any difference between the definitions of empiricism and empirical research, according to Wikipedia.

Comments on this entry are closed.

Previous post:

Next post: