Bryan Caplan  

A Hawk-Dove Ideological Turing Test

Does French 'exception culture... Efficiency, Equity, and Ideolo...
In Stereotype Accuracy, Clark McCauley describes a fascinating Ideological Turing Test from 1972:
Dawes, Singer, and Lemons (1972)... recruited students who were "hawks" and "doves" with regard to the Vietnam War and asked them to write opinion statements that the typical dove on campus as well as the typical hawk on campus would endorse.  Then they recruited a second group of hawks and doves, asked the hawks to agree or disagree with the hawk statements written by hawks and doves, and asked doves to agree or disagree with dove statements written by hawks and doves.  Hawks rejected more hawk statements written by doves than hawk statements written by hawks, and doves rejected more dove statements written by hawks than dove statements written by doves.  Both hawks and doves rejected statements mostly on the grounds that they were too extreme.
The degree of exaggeration was... neither large nor consistent.  Hawks rejected as too extreme 16 of 40 statements written by doves and 11 of 40 statements written by hawks; doves rejected as too extreme 9 of 40 statements written by hawks and 8 of 40 statements written by doves.  Thus, doves significantly exaggerated the extremity of the typical hawk, but hawks showed only a weak tendency to exaggerate the extremity of the typical dove.
Dawes et al. (1972) recognized two possible explanations of their results.  One is the information-processing version of exaggeration theory: We disregard information indicating moderation or neutrality because this information is more difficult to assimilate.  The other possible explanation... is motivational: Exaggerating the position of our opponents reduces the force of their arguments.
A well-designed Ideological Turing Test reverses the latter temptation - the more accurate your description of your opponents' position, the more credible your rejection of that position.

COMMENTS (10 to date)
John Thacker writes:

Stereotype Accuracy is a fascinating book. Glad that I got my copy before it went out of print.

Hazel Meade writes:

One game I've seen people play on message boards is to create a "spoof" character of their opponents. The ideal and best spoofs are those that are the most convincing. You want to make the spoof just accurate enough to convince people tyhat he's a real person, and just extreme enough to make him ridiculous. And thus hilarious to people who get it.

Hi, Hazel.

We've seen that game and variants of it here on EconLog since our inception in 2003. The game of arguing against oneself under multiple nicks/email-address/IP-address is one of several reasons that we require commenters to have valid email addresses--including responding to us when we send them email. Our policy requirement of an email response if we email a commenter doesn't by itself rule out someone's having multiple legitimate email addresses or multiple personas. If someone wants to change his nick/email-address/etc. in the long run, we generally don't want to stand in the way of that. But a cost of affording that flexibility is that if someone is committed to spending time gaming the system here, it surely can happen.

We think our email verification policies pare down down the gaming a bit, but we don't delude ourselves. So, for example: Hazel, are you one of those gamers? We don't know. :)

Games are a perpetual dance of mirrors in a hall of mirrors. Sometimes you just have to cut through with common sense.

RA writes:

In Jonathan Haidt's last book, he argues that conservatives are better at guessing how liberals would answer value questions than vice versa. The results from this experiment are consistent with that, as the hawks were less likely to exaggerate the opinions of their opponents.

Tom West writes:

I'll admit I was vastly entertained to find out that some years ago on a board I was reading that the two most insightful commentators on two *opposite* sides of the issue were the same person...

Unfortunately for the poster, I got the impression that the general community felt the sock puppet made a slightly better argument than the putative "real" person :-).

Hazel Meade writes:

Not under THIS persona! LOL.
Although actually what I mean is something different. "Spoof trolling" is the art of entertaining one's fellow commenters by doing a paraody of the opposition that only a few very smart people can detect. (You have to include a few "gives" to clue in the more perceptive people that it's a fake.) The comedy is watching other people get fooled by the spoof, or even people that the spoof is parodying agreeing with the spoof. There is really no arguing with oneself involved.

The point of this is that people can get much more sophisticated than the ideological Turing test, and are often very good at detecting imposters. It's a game about challanging other people to detect the "fake" hawks from the "real" hawks, as in the example above.

Daniel Kuehn writes:

Another option of course is that the set of people self-identifying as hawks or doves is different from the set of people identified by out-group members as hawks or doves. The out-group may be very accurately describing a somewhat different group.

A lot more people consider themselves "libertarian", for example, than regular readers of this blog would probably consider libertarian. As a non-libertarian I typically think of the sort of people who agree with what's on this blog as "libertarian" and would write that way on a Turing test. But if Greg Mankiw were reading it (he has referred to himself as a libertarian) it wouldn't describe him very well and he'd say it's extreme.

Ideology is not a characteristic like height or weight that can be precisely measured.

There may also be signaling with buzz words and the structuring of the language. It may be that hawks respond more to buzz words and the use of certain language than doves are so they are more clued in when someone tries to imitate that.

guthrie writes:


Your comments remind me of a radio personality who developed a whole show based upon this game.

Phil Hendrie used his gift of impressions and combined that with the standard radio delay (about 7 seconds) and developed outrageous characters with whom he would get into 'arguments'. He would then allow listeners to call in to join the 'debate'.

The fun comes when a listener realizes the 'house of mirrors' aspect to the show and joins in the joke. Or stops listening out of embarrassment.

It might be interesting to see how he, and perhaps his listeners who ‘got the joke’, would do on such an ideological Turing Test. It might also be interesting to see how adept he is/they are at picking out imposters.

MingoV writes:
Hawks rejected as too extreme 16 of 40 statements written by doves... doves rejected as too extreme 9 of 40 statements written by hawks... Thus, doves significantly exaggerated the extremity of the typical hawk...

The results are the opposite of what is stated. Hawks categorized dove statements as extreme almost twice as often (16 vs. 9) as doves categorized hawk statements as extreme. That's not surprising given some of the statements I've read and heard from pacifists who believe that chanting a mantra or wearing "Make love not war" T-shirts will stop invading armies. (Though I do agree with their opposition to the Vietnam war and other wars.)

Ken P writes:

Don't forget time period. Two years post Kent State shooting.

That may cause doves to have a pretty extreme assessment of how hawks look at things. Also, my guess is that more doves were activists than the hawks.

Comments for this entry have been closed
Return to top