[EM] Range Voting strategy

[David Cary]

There are several elements of the Beaumont poll that bias it away
from real-world behavior.

1.  The poll was a poll, not an election.  There were no consequences
to the answers, hence no incentive to answer tactically.
2.  The poll asks for an indication of preferences, not a vote.
3.  The poll incompletely, arguably to the point of being inaccurate,
describes how to maximize the effect of a ballot.  
4.  The poll presented people with an unfamiliar voting method and
required them to respond in an ad hoc manner.  Respondents had little
time or incentive to consider what the possible tactics might be,
what the risks and benefits of various tactics might be, or which
tactics might be useful. 
5.  There was no opportunity for seeking or receiving outside advice
about how to use tactics.
6.  There was no information available to respondents about what the
likely range-voting specific behavior of other voters in that
election might be.  For example, there was nothing equivalent to
polling results about who the range-voting front-runners were and by
what margins.

In general, these deviations from real-world election behavior all
tend to bias the results towards non-tactical responses.  The 2004
Presidential poll suffers from similar deviations and bias.  (It gave
the following scoring instructions: 100=“great,” 0=“terrible,”
50=“middling.”)

Also, I'll note that neither poll was an exit poll in the usual
sense.  It was not asking people how they had just voted.  As such,
it would be unfair to attribute to this poll the accuracy normally
associated with exit polls.  

I'd suggest that there are a number of factors that would influence
the extent and severity of tactical voting in the "real world". 
Those factors vary depending on the circumstances of the election. 
That is true of any type of voting, not just range voting.  Claims
that there would always be some single extent and severity of
tactical voting are overly simplistic.

-- David Cary

--- Warren Smith <wds at math.temple.edu> wrote:

> >Laatu:
> You used the same word "poll" that I used. People obviously 1)
> didn't  
> have any major reason to try to force the results in any direction 
> 
> and 2) probably were not told and did not understand the strategic 
> 
> possibilities of Range Voting. People may also typically want to  
> answer sincerely in opinion polls.
> 
> --actually, anyone examining    
> http://rangevoting.org/Beaumont.html
> would see that right up front voters are instructed
> "To maximize the effect of your ballot, start by giving your
> favorite candidate a 10, and your least
> favorite a 0, and scoring the rest relative to that."
> 
> Also, my old 2004 exit poll said right ontop of its ballot also
> "Please vote the way you actually would if the real election were
> being held using range voting" 
> but this was evidently not felt to be explicitly meeting Laatu's
> criticism enough, which is
> exactly why Clay Shentrup's ballots in 2006 went further as I just
> said.
> 
> Also, exit polls are normally regarded as the most accurate kind of
> poll
> since they are of real voters who just voted.
> 
> So in conclusion:
> (a) cease these guesses and aspersions and actually read the
> ballots
> (b) Yes, I am willing to believe real elections would induce more
> approval-style range voting
> than polls, but a poll result this enormous and massive (80%, 97%)
> is NOT, I put
> it to you, likely to swing to over 90% the other way like Laatu
> seems to think it
> will.  I challenge Laatu to cite any poll on any occasion in the
> entire history
> of the world, where such a massive inversion happened.
> 
> Warren D. Smith
> http://rangevoting.org
> ----
> election-methods mailing list - see http://electorama.com/em for
> list info
> 



 
____________________________________________________________________________________
Do you Yahoo!?
Everyone is raving about the all-new Yahoo! Mail beta.
http://new.mail.yahoo.com