[EM] lying to pollsters (was comparative effectiveness ...)

[Bart Ingles]

Russ Paielli wrote:

> Forest Simmons simmonfo-at-up.edu |EMlist| wrote:
> 
>>> From: Russ Paielli <6049awj02 at sneakemail.com>
> 
> 
>>> So what is the optimal strategy in responding to an Approval poll? Do
>>> any or all voters have an incentive to lie about their cutoff point --
>>> or perhaps to even rearrange their preference order before drawing the
>>> line? And how would such strategy affect convergence if everyone adopted
>>> it? Will honest respondents be at a disadvantage?
>>>
>>> Obviously that question is a lot easier to ask than it is to answer, but
>>> I think some sort of answer in necessary before Approval can be fully
>>> evaluated.
>>>
>>
>> I think that the average person would be apt to play up support for 
>> his favorite and down play support for compromise, and that most folks 
>> would take this into account when interpreting the polls.
> 
> 
> I'd really like to know what the optimal strategy is for replying to a 
> pre-election Approval poll. If anyone is capable of determining that, 
> you are probably the one, Mr. Simmons. I suggest you put your capable 
> mind to it.
> 
> Maybe that strategy is just too darn complicated to formulate. Then 
> again, maybe the optimal strategy in replying to a poll is not much 
> different from the optimal strategy in the actual election. If that is 
> the case, then the "disinformation" problem may not be too bad.

The thing about polls is that they're nonbinding, in which case there 
will be incentive to lie under just about any method.

If both the poll and the election are approval-style, I think the 
incentive would be to bullet vote in the poll and then use whatever 
Approval strategy is applicable in the real election.  In particular, 
candidates in a close race for 2nd place would each want to appear to be 
best positioned to challenge the front-runner, hopefully to induce other 
voters to approve two candidates.


If the final election is Condorcet, then the incentive during the poll 
might be (given the following "typical" competitive 3-way race):

30 ABC
15 BAC
10 BCA
45 CBA

(1) for CW (B) supporters to try to make an extreme candidate 
(preferably C) look stronger, thereby incenting other voters (or at 
least the A voters) to rank fully and sincerely.  If they don't know 
which of A and C is stronger, they should collectively pick one at 
random, and then all rank this candidate 1st and their own 2nd.  In the 
real election, they should either rank sincerely or bullet vote for B.

(2) for the C supporters to make their candidate look artificially weak 
in the poll, to try to make the A voters think they have a shot at 
winning a 2-way race between A and C, and either truncate or bury B in 
the election (at which time some of the C voters might then bury B in 
order to create a cycle).

Bart