France nationwide score-voting style poll!

21 views
Skip to first unread message

Warren D Smith

unread,
May 7, 2013, 7:31:48 PM5/7/13
to Claude Hillinger, electionscience, Jean-Francois Laslier, Antoinette Baujard, Frédéric Gavrel, Herrade Igersheim, Isabelle Lebon
WOW! See this!

http://www.nationspresse.info/wp-content/uploads/2011/04/Sondageopinionway.pdf

This is a professional polling agency doing a nationwide poll of
France using 7-point scale evaluative voting, including "majority
judgment" median-based method!
1000 respondents. The bad news is this poll was done in April 2011 --
1 year before of the real April 2012 presidential election. Therefore
the candidates were not the same set as those who actually ran...

The MJ winner (no tie-break needed) was (Socialist Party head) Martine Aubry.
In the actual election, Aubry did not run, but
Francois Hollande (also Socialist party) did run (beat Aubry 56-43 in a 2-person
runoff in the Socialist Party primary), and won the Presidency.
Hollande also was the approval voting and several kinds of
average-based score voting winner. The Condorcet winner either was
Hollande or Bayrou.
Asterisks (*) on candidates who did not actually run.

"Canddt".......Rated above MJ....MJ.... Rated below MJ

1 Aubry* 38,0% Assez Bien – 49,3%
2 Borloo* 46,0% Passable + 34,4%
3 Villepin* 40,1% Passable + 39,3%
4 Bayrou 37,9% Passable + 35,9%
5 Joly 29,9% Passable – 49,8%
6 Sarkozy 46,9% Insuffisant + 41,3%
7 Chevènement* 43,1% Insuffisant + 32,2%
8 Mélenchon 36,8% Insuffisant – 41,8%
9 Besancenot* 35,4% Insuffisant – 44,2%
10 Dupont-Aignan 25,5% Insuffisant – 46,7%
11 Arthaud 25,8% Insuffisant – 48,0%
12 Le Pen 44,4% A Rejeter -

Here are the score voting totals found in this poll for the candidates
who actually ran, plus Aubry (counts for scores start from max score
ordered toward min score on 7-point scale):
Melenchon: 13 27 50 112 165 214 418
Joly: 32 47 74 145 203 190 309
Aubry*: 82 129 170 126 196 114 184
Bayrou: 12 47 128 192 261 166 193
Sarkozy: 41 87 111 95 135 118 413
Dupont-Aignan: 5 14 27 70 139 277 467
LePen: 68 65 70 72 78 93 556
Arthaud: 1 9 34 77 137 261 480

The pollees were asked WHICH voting method they preferred:
1. Traditional (plurality plus 2nd round runoff): 63%
2. MJ (median-based 7-point scale): 36%
3. Other/don't know: 1%


--
Warren D. Smith
http://RangeVoting.org <-- add your endorsement (by clicking
"endorse" as 1st step)

Warren D Smith

unread,
May 7, 2013, 7:35:23 PM5/7/13
to Claude Hillinger, electionscience, Jean-Francois Laslier, Antoinette Baujard, Frédéric Gavrel, Herrade Igersheim, Isabelle Lebon
At the end of the poll, the pollees were asked WHICH voting method
they preferred:
1. Traditional (plurality plus 2nd round runoff): 63%
2. MJ (median-based with verbal 7-point scale): 36%

Dale Sheldon-Hess

unread,
May 7, 2013, 7:39:33 PM5/7/13
to electionscience
On Tue, May 7, 2013 at 3:31 PM, Warren D Smith <warre...@gmail.com> wrote:
> http://www.nationspresse.info/wp-content/uploads/2011/04/Sondageopinionway.pdf
>
> The Condorcet winner either was Hollande or Bayrou.

I knew Bayrou would do much better (as seen in previous evaluative
polls) but I'm (I guess pleasantly?) surprised that Hollande was a
likely evaluative and Condorcet winner, i.e., that ttr possibly got it
right in the end. (Less likely that it did so when Sarkozy won.)

--
Dale

Clay Shentrup

unread,
May 7, 2013, 7:39:48 PM5/7/13
to electio...@googlegroups.com
At the end of the poll, the pollees were asked WHICH voting method 
they preferred:
1. Traditional (plurality plus 2nd round runoff):  63%
2. MJ (median-based with verbal 7-point scale):  36%
3. Other/don't know:   1% 

How incredibly depressing. 

Warren D Smith

unread,
May 7, 2013, 9:13:38 PM5/7/13
to electio...@googlegroups.com
--I conjecture that the use of a verbal 7-point scale has a
puke-inducing effect.
Why in hell anybody would want that instead of (more concise and
meaningful) numbers, is beyond me -- but obviously this is a pure
guess. I also conjecture the use of a complex median+tiebreak scheme
revolted the voters as opposed to "highest average wins" which is
simple. This is also a pure guess.

Let me put it this way. If your algorithm is complicated, then voters
either will not understand it, or will at least be afraid of it. What
voters do not fully understand, they
(mostly) oppose. For example it is clearly true that if voters do not
know anything about some candidate, they tend to give him "zero" score
vote, often. If you propose a complex law and try to get voters to
enact it by referendum, then the more complex it is, the more of an
uphill battle it will face from voters -- who think you are trying to
pull the wool over their eyes.

Similarly, if they do not understand a voting system, they I
conjecture would oppose it.

Those who like complicated voting systems should keep that in mind.
You are in a deep
hole, probably "-25% of the populace worth," as soon as you do that.
And in my opinion Jameson Quinn's "GMJ" scheme is simpler and clearer
than the original MJ scheme, but still complex enough it would be in
such a deep hole.

But this is all conjecture.

Here is the (clearly incomplete) description of MJ that was provided
to the pollees:
==============
Un nouveau mode de scrutin : présentation du concept

Imaginons maintenant que l’on vote selon un nouveau mode de scrutin,
qui fonctionnerait selon le principe suivant :
Le vote a lieu en un seul tour. Il vous est demandé d’évaluer chacun
des candidats, en lui attribuant l’une des mentions suivantes :
Excellent / Très bien / Bien / Assez Bien / Passable / Insuffisant / à
Rejeter. N’accorder aucune mention à un candidat est comptabilisé
comme « à Rejeter ». Naturellement, vous pouvez accorder une même
mention à plusieurs candidats.
A l’issue du vote, chaque candidat obtient un certain pourcentage de
chaque mention. Sa mention finale, appelée la « mention-majoritaire »,
est celle approuvée par au moins 50% des votants. Par exemple, un
candidat ayant obtenu 5% d’ « Excellent », 10% de « Très bien », 30%
de « Bien » et 10% d’« Assez bien » (et donc 45% de mentions moindres)
reçoit la mention-majoritaire « Assez bien ».
Le vainqueur est le candidat qui obtient la meilleure
mention-majoritaire. Si plusieurs candidats obtiennent la même, le
vainqueur est celui qui obtient le plus de meilleurs mentions et/ou le
moins de pires mentions.
===========

Jameson Quinn

unread,
May 7, 2013, 10:13:37 PM5/7/13
to electio...@googlegroups.com
That French description is not badly written, but still I think harder to understand than the Bucklin-like justification for GMJ. 

But really, we can't know why French people prefer the status quo without further evidence. I don't even really trust my own intuition here, and much less Warren's (no offense, Warren; obviously I'd trust my own more.)

Jameson

2013/5/7 Warren D Smith <warre...@gmail.com>

--
You received this message because you are subscribed to the Google Groups "The Center for Election Science" group.
To unsubscribe from this group and stop receiving emails from it, send an email to electionscien...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.



Clay Shentrup

unread,
May 7, 2013, 11:11:29 PM5/7/13
to electio...@googlegroups.com
I completely agree with Warren.

Jomo France

unread,
May 30, 2013, 7:48:42 AM5/30/13
to electio...@googlegroups.com
We've used two rounds for legislative elections since the 19th century, except for 5 or 6 elections. This is also what we used for local elections, including canton councillor and municipal council (as a list). And what was .

So, despite the system has enormous flaws, notably because the 50pc majority is reached artifically, people are used to it and have a sense of it. . . somewhat. The mindset of people is not as fixed towards 2RM as Plurality in the U.S., because virtually all election methods have been reformed these past 30 years which make the system unclear. But that's how I explain it.

Besides, MJ, a brand name for a median-based choice, does not seem serious, and experts showed it could be arbitrary. The guys behind it don't know what a campaign is. Between both I'd choose the devil I know.

Jameson Quinn

unread,
May 30, 2013, 9:04:17 AM5/30/13
to electio...@googlegroups.com
Besides, MJ, a brand name for a median-based choice, does not seem serious, and experts showed it could be arbitrary.

Which experts, when? Are you referring to Zahid, Manzoor Ahmad, and Harrie de Swart. Majority Judgment Theory and Paradoxical Results. Working paper). Tilburg University, Netherlands, 2009. http://www2.eco.uva.es/presad/SSEAC/documents/ZahidPaperfinal.pdf? That paper rests solely on essentially the same example that Unger recently brought up here. As I argued, this example is implausible in several ways, and one can construct "paradoxical" results for any system.
 
The guys behind it don't know what a campaign is.

I don't know exactly what you mean. But I think my response would be: "let he who is without sin..."
 
Between both I'd choose the devil I know.

Both what?
Reply all
Reply to author
Forward
0 new messages