The secret consensus among economists

[Aimless]

http://www.igmchicago.org/igm-economic-experts-panel

Whatcha reckon? Are these the wrong economists?

[wiggin]

Actually, there's a very interesting divide between opinions held by academic economists and economists who work in industry. This appears to be mostly if not entirely academics - so certainly worth looking at, and revealing, but not necessarily representative of a true consensus. Also US-centric, which I don't have an issue with, but at odds with some other prevailing economic models.

[RandBlade]

Any panel anywhere is likely to include a bias.

Another issue in that polling format is that there is (regardless of question) normally a systematic error towards agree.

EDIT:


An extreme sitcom example of why that can happen but a lot of such "agree/disagree" do introduce leading facts before asking the question. It's also psychologically believed that agreeing is preferred to disagreeing and demonstrated in studies that you can ask the same people to agree or disagree with polar opposite questions and a significant proportion will agree to both mutually incompatible statements.

[Loki]

The further you get from a social scientists immediate area of expertise, the more their opinion reflects their personal and ideological biases than it does their scientific training. I've seen similarly absurd results for international relations professors. What this poll does is the equivalent of asking all biologists whether they agree with evolution, instead of restricting the sample to evolutionary biologists.

I also wouldn't be surprised if there was a systematic response rate bias in favor of professors who think the government should have a more active role in the economy.

[Dreadnaught]

Just from skimming a few of the queries, their answers seem to be heavily weighted towards "agree".

Economists are yes men.

[RandBlade]

I told you it's a systematic error in almost all polls. Using the words agree and disagree leads to bias.

Thats why in the Scottish referendum the questions been changed from its initial "Do you agree that Scotland should be an independent country" to "Should Scotland be an independent country"

[Aimless]

The thought occurred to me, and I agree it may be a reasonable objection in various settings, which is why I have asked them to poll the panel with the following statement:

"Polling a panel of economic experts with statements about economic policy with which they can agree or disagree comes with a significant risk of biasing answers in favour of agreement."

We've got them now

[Aimless]

What I'm curious about is how the inclusion of confidence levels influences reliability and validity

[GGT]

Another reason to be skeptical of opinion surveys/polls.

[RandBlade]

Should always have a healthy skepticism of opinion polls but it also depends upon what is being measured. Agree/Disagree format questions are notoriously unreliable while other questions (especially Voting Intention) are typically far more accurate.

[Dreadnaught]

The thought occurred to me, and I agree it may be a reasonable objection in various settings, which is why I have asked them to poll the panel with the following statement:

"Polling a panel of economic experts with statements about economic policy with which they can agree or disagree comes with a significant risk of biasing answers in favour of agreement."

We've got them now

It would be neat if they actually took it. I wonder how many people actually submit to this thing.