The "Unrepresentative Sample" Canard

Iain Murray is repeating the lame argument made by Darrell Issa at the Waxman hearing yesterday:

The UCS mailed out over 1600 survey forms to climate scientists and based their assertions of political interference on the 297 that got returned. That's a response rate of just 19 percent. OMB guidelines clearly state that a response rate of less than 80 percent requires an investigation of potential biases and an even closer investigation for a response rate lower than 70 percent. A response rate of lower than twenty percent is clearly vulnerable to the charge of a self-selecting sample, perhaps those with an axe to grind against their bosses, the politically motivated, and so on. In short, it proivides all sorts of legitimate reasons to dismiss the survey as utterly unrepresentative.

I will say it again: The UCS found over 100 government climate scientists claiming political interference with their work or the dissemination of that work. It is beyond perverse to attack this finding on the grounds that it isn't "representative"--that's not the point.

Suppose the UCS had never even called this a "survey" at all but just announced numbers of scientists reporting political interference. They would still have documented a very disturbing phenomenon. Or alternatively, imagine that the survey was representative after all, but only 10 percent of scientists reported political interference. That would still be a serious problem.

These considerations show that the "unrepresentative sample" argument, although perhaps the best the right can do, is utterly vacuous.

More like this

Well, I got off jury duty and managed to get over to the Hill after all. As I write this, the Waxman hearing is still ongoing, and you can watch it here. I have it on good authority from three people that in the opening statements, Rep. Jim Cooper, a Democrat, actually waved a copy of The…
Yesterday, the US House of Representatives Committee on Oversight and Government Reform held a hearing on âAllegations of Political interference with the Work of Government Climate Change Scientists.â As committee chair Henry Waxman noted in his opening statement, the committee had been…
When the Union of Concerned Scientists (UCS) last week released a report detailing widespread political interference in science at the Environmental Protection Agency (EPA), I almost didn't blog about it, since the fact that political interference runs rampant in the Bush Administration shouldn't…
The Union of Concerned Scientists has released another disturbing report about political interference with government science. For Interference at the EPA, they surveyed EPA scientists from all of the agencyâs scientific program offices and 10 regional offices, and from more than a dozen research…

See, Chris, it's not unrepresentative when you have an e-survey that finds things that conform to your worldview, but ANY survey IS unrepresentative when that survey is found to disagree with your chosen identity received worldview.

IOW, they are making sh*t up. Good to know they still have nothing and have to make stuff up.

Best,

D

"OMB guidelines clearly state that a response rate of less than 80 percent requires an investigation of potential biases and an even closer investigation for a response rate lower than 70 percent."

I could not agree more.

Bring on the investigation!

By Dark Tent (not verified) on 31 Jan 2007 #permalink

Its interesting that according to the UCS employees were instructed not to reply to the surveys. If you combine this with the OMB guidelines you can automatically get a survey you don't like branded as biased.

Wow thats impressive in a sleazy and underhanded kind of way. The Bush administration does have a machivellian way of getting what it wants.

By Joseph O'Sullivan (not verified) on 31 Jan 2007 #permalink

Roger Pielke, Jr., agrees that this sampling critique is weak and misses the point. However he points out that the UCS should not have included inadequate funding for climate research among its gripes -- which is totally spot on. I don't know why UCS does this, it totally obscures what the real issue is. Not giving enough research funding to scientists is not a misuse of science in any way, shape, or form.

"Not giving enough research funding to scientists is not a misuse of science in any way, shape, or form."

I would not go that far.

Withholding funds can be used to choke off research for political reasons. For example, adopting a policy that "we need to do more research on global warming to determine if humans are causing it" and then turning around and killing funding for a satellite program which might confirm that such a link exists certainly might be construed as a "misuse" -- if not of science itself, of the scientific funding process.

Such "misuse of funding" is admittedly difficult to prove since the whole subject of how much money should be spent on any given science is political in and of itself. But that does not mean that there are never cases where such misuse occurs.

By Dark Tent (not verified) on 31 Jan 2007 #permalink

"However [Pielke] points out that the UCS should not have included inadequate funding for climate research among its gripes -- which is totally spot on."

This point by Pielke and agreementy by Mooney reveals a bit of provincialism and lack of understanding of politics. There are numerous documented cases where officials have interferred with research by cutting off funding. It's a legitimate concern and the UCS was right to look at it, although it's difficult to find proof. And I'm not sure if you CAN find proof through something like a survey.

One case is the CDC's Injury and Prevention Center which became a target for budget cuts after research by Kellerman and others looking at the dangers of firearms. Look it up.

Federal researchers know that they are vulnerable for budget cuts and there are current congressional investigations into whether Bush is trying to cut off climate change research by cutting back on the funding for satellites.

It's not some great secret.