Robin writes (in reference to my model explaining extremeness and

"A problem with this model is that if all these people are getting
about the same thing, their signals should be correlated.  The people
become more confident and move from the initial consensus position
also tend to move in the same direction from that consensus."

        Robin then concludes "But we observe similar levels of confidence
moving both directions away from the consensus."

     The latter observation, however, is not what correlation predicts. 
Correlation predicts that more people should be on one extreme than the
other - it says nothing about the firmness of those on either extreme. 
I would argue that on factual matters this is often what we see, i.e. an
ignorant consensus which one might read about in say Time magazine or
ABC news or hear about in the street and a quite different view among
those who have seriously studied the issue.  [Bryan's research perhaps
provides some evidence - Bryan?]

        Moreover, on a number of issues I don't think correlation is that high
- the world is not so kind as to produce all true signals or to make
every signal equally available to every person.  If that were true, then
obviously everyone would agree with me.  A better world no doubt, but
not the one we have.    

        Robin also writes "Another problem is that people have to be unaware of
other people's positions, otherwise they could update on that and then
should all end up at the same place."  

      Not a bug, a feature!  People *are* unaware of other people's
positions!  Moreover, even if I knew everyone's position it would still
be difficult to judge the relative credibility/expertise of those


Dr. Alexander Tabarrok
Vice President and Director of Research
The Independent Institute
100 Swan Way
Oakland, CA, 94621-1428
Tel. 510-632-1366, FAX: 510-568-6040

Reply via email to