Dienstag, November 10, 2009

Statistics, USA Today and ... Integrity

There are times I come close to despair (not that I succumb: despair is a mortal sin).

This is one of those times.

According to the headline, "Economists See Threat In Climate Change"

The article then quotes: 94% of economists surveyed think that the US should join climate agreements to limit global warming.


Nope: they surveyed 289 environmental economists who had published climate-related studies.

Talk about cherry-picking: this is the reason why surveys richly deserved the utterly abysmal reputation that they "enjoy".

This is like asking 289 murderers if they think the legal system is too harsh.

This is like asking 289 nurses if they think they are underpaid.

This is like asking 289 teenagers if they think their parents are being too restrictive.

This is like asking 289 terrorists if they think the War on Terror is overdone.

This is like asking 289 teachers if they think their course load is too heavy.

This is like asking 289 police officers if they think the courts are too lenient.

This is like asking 289 prostitutes if they think their pimps are abusive.

This is like asking 289 drug users if they think junk is too expensive.

This is like asking 289 students if they think their teachers shouldn't grade so hard.

See the point?

To underscore the complete and total lack of journalistic quality at USA Today, understand this quote:

"Many observers look at economists as skeptics of the need for (climate) mitigation," says economist Gary Yohe of Wesleyan University in Middletown, Conn. But "most accept the unquestionable consensus from the natural scientist that the planet is warming and humans are to blame."

The language is telling: it's an unquestionable consensus.

Things that are unquestionable are things of dogma, superstition and belief.

But not science.

A proper sampling would have used 289 names chosen at random from, say, the membership of the American Economic Association.

But that proper sampling wouldn't have given the folks who did the survey - The Institute For Policy Integrity - the results that they wanted.

But...wait a second...

The Institute For Policy Integrity????

This is what they say about themselves:

The Institute for Policy Integrity (IPI) is a non-partisan advocacy organization and think-tank dedicated to improving the quality of governmental decisionmaking.


Sponsored by the New York University School of Law, and housed in the Frank J. Guarini Center for Environmental and Land Use Law, IPI consists of a board of advisors of leaders in government, business, and policy; members of the Law School's world-class faculty; and a team of staff and fellows dedicated to the Institute's mission.

IPI advocates for reform before courts, legislatures, and executive agencies. By reaching out and building coalitions with traditional nongovernmental organizations, IPI facilitates their effective participation in the regulatory process. IPI is also dedicated to contributing original scholarly research in the areas of environmental, public health, and safety policy.

Given their extremely poor sampling practices, which they justify as representing the opinion of "experts", they are scarcely non-partisan: a non-partisan advocacy organization is perhaps the greatest oxymoron of them all.

Just what you get when you let lawyers wander around without supervision. Garbage in, garbage out.

Sorry: this is the work of hacks of the very worst kind. Not because I am a skeptic: it is an affront to statistical sampling and analysis. That is something that us empirical folks take very, very seriously.

At least 94% of us...

Keine Kommentare: