At 11:15 PM 5/7/2013, Clay Shentrup wrote:
>We need your votes! You can sign in with Facebook or Github.
>
>The current top 10 are:
>Solutions To Strengthening America's Democracy / 158
>Civic Hacking in Pursuit of Democracy / 119
>Dear Democracy… / 97
>Bring It to The Table / 92
>Spin / 73
>DIVE IN: This is What Democracy Looks Like / 64
>Exciting / 61
>Carrots and Sticks for Democracy / 60
>Agenda Setters / 55
>Democracy is US! / 50
So, how are voter finding what to vote for out of
375 videos? I think it's probably networks of
friends. In other words, most of the voters have
an a priori bias. Most voters have not even seen
our video. This is a structural defect in the process.
Another problem is that we don't know the total
number of voters. That's important! If nearly
everyone is bullet voting, we can readily guess
that they are partisans. I'm pretty sure there is
more than one very good video in there! Ours is one, of course.
But is there a systemic bias other than this?
There are three sorts from the main page: Name, Recent, Comments.
Name is alphabetical. One of the top ten is in
that first screen of 30/375 videos. Agenda Setters.
Recent is probably a reverse date sort, so that
someone who wants to at least see all the videos
and who has been following the site can see
what's been added. And one video from Recent is in the top 10, DIVE IN.
Comments is obviously based on comments, but
exactly how isn't clear. Most web sites would
give a priority to the most recent comments. In
any case, we are at the top at this moment on
Comments. If this is quickly see as us talking
about our own video, this could backfire, but it
*is* getting our image up there, or it would have
been completely buried in the pile of 375.
In the first screen for Comments, I find
Our video first.
Carrots and Sticks for Democracy.
Agenda Setters
Spin
Dear Democracy
Civic Hacking
DIVE IN
(I also notice that the MN FairVote video is now
visible in the Comments first screen. *We created
that* by commenting there. Remember, bad
publicity is better than no publicity, and what
doesn't kill you makes you stronger....)
So, what do I get from this? Well, that 30/375
shows very less than 1/10 videos. We might expect
to see roughly one from chance. However, I'd also
expect the default screen, especially, for
readers, to attract more actual viewers and
therefore more approvals. All this can lead to
strategies to give one's video a leg up.
Possibly, upload it just before the deadline,
just enough time to allow for something to go
wrong and fix it. Slightly risky, then. Our video
could have been called A is for Approval. Would
have been top. They *could* decide to, next time,
reverse the order.... probably not.
(The best thing they could do is to accumulate
range ratings for each page, from upload. If they
did that, optimal strategy would be to get one's
video up sooner, you'd actually get rated by more
people because you would be seen by more people.
Or they could simply start voting as they did,
and could have a display that presents videos in
sequence of average rating. And a button that
jumps to random video as well, or "trending,"
etc. I would think that *their goals* would be
best served by encouraging every video to be seen
by as many people as possible....)
Are Comments (simply the number of comments,
without voters even looking at the actual
comments) driving votes, or are motivated voters
driving comments? 7/10 of the top ten are also on
the Comments first screen. That's a strong correlation. Cause?
Carrots and Sticks for Democracy is also awful as
to content. It's about treating politicians as
objects to be manipulated by "carrots and
sticks," and it's similar, then, to the proposal
prominent in #1 to fine politicians for "not
keeping campaign promises," never mind that this
would then set up a conflict between an oath of
office and personal interest, wherever it became
clear to the politician, *in context*, that his
duty required something other than what he
promised. I.e., it would remove the deliberation
from our deliberative democracy, and push it toward mob rule.
One of the *best* features of our democracy is
that demagogues can break their promises!
The solutions proposed are actually worse than
the problems. That's common when people think and
speak from *outside a system.* They have no
*personal relationship* with politicians, nor do
the politicians have a personal relationship with
them. The scale is too large for such
relationships to be anything other than very
unusual. And that's why Asset, blah, blah, blah.
I'll know that we are really cooking with gas
when our video is "A is for Asset Voting."
Until then, we have a no-brainer improvement that
can, in principle, improve any other voting
system *including IRV.* (i.e, simply allow
approval voting within each rank, and a *ballot*
is not exhausted at a rank until or unless *all
approvals at that rank have been eliminated. It's
still IRV and most people might vote it as IRV,
but it is *also* Approval and could be voted as
pure approval, or pure approval with fallback
votes. Better method than raw IRV. Flawed, *but
the flaws can be ameliorated by how the voters vote.*)
Of the top eight as to comments, four are in the
top ten as to voting. The exceptions are
Our video
http://lookingatdemocracy.org/submissions/14935-the-united-states-needs-approval-voting
Mostly our comments, some others.
http://lookingatdemocracy.org/submissions/15135-myvotenow-org
Look like independent comments, short comments.
http://lookingatdemocracy.org/submissions/15178-the-american-walrus-party
Appear to be the comments of college friends. Satire.
http://lookingatdemocracy.org/submissions/14527-governmental
Comments very short, possibly social circle,
video is platitudes, mostly. Very young, high school or younger.
So, from the top 10:
http://lookingatdemocracy.org/submissions/15084-carrots-and-sticks-for-democracy
Comments appear to be from the general
public. Comments brief, but longer than the
"social" comments, and show real interest.
http://lookingatdemocracy.org/submissions/14839-agenda-setters
Comments are lengthier, appear independent,
mostly. Reading them, I'm interested in seeing
the video. It seems to be about a Votorola-type
proposal, but where they go with it will be important.
http://lookingatdemocracy.org/submissions/14184-spin
Short compliment comments. Spin is again a Votorola type proposal.
http://lookingatdemocracy.org/submissions/15175-dive-in-this-is-what-democracy-looks-like
Substantive comments, longer, that don't
appear to have been made by the makers of the
video. Again, seeing the comments, I want to see the video.
Something we can learn from this? I do have some ideas....
A little more:
http://lookingatdemocracy.org/submissions/14932-solutions-to-strengthening-america-s-democracy
is the number one in votes, 158 reported. It
has *one* comment that *resulted* from it being
number one. So it got those votes without
*anything* on the site driving viewers to it.
This video is from a person who made three submissions.
http://challengepost.com/users/trjr/submissions
Frankly, I don't see any reason why this video
would be so well-known, just from the site..
I.e., suppose it is *really excellent* (remember,
I haven't seen it yet.) That still would not
explain the rapid jump into the lead. Two
possibilities: external mention in some place
seen by a lot of people, or socking. The former
is legitimate, but it may also burn out. (I.e., not continue to increase.)
Being in the top ten will almost certainly lead
to more views and thus some increased level of
voting. To avoid misleading cascades, they should
probably list the top ten, if they are going to
announce it at all, in reverse order with links.
I'd suggest something quite different.
Hah! They need systems that are probably covered
by submissions, to some degree. That is, if they
really want a true public choice, rather than the
public following accidental cues and
manipulation, they need a more efficient and
thorough evaluation as part of a deliberative process.
Really! Voting systems people, what do you think
of an election that presents, with no
prefiltering, 375 candidates? 3 minute videos, 19
hours to look at all of them. Ain't gonna happen,
my guess is *nobody* will do this. Approval is
much better than Plurality, but .... a
single-stage process is very likely to produce suboptimal results.
https://www.facebook.com/LookingAtDemocracy?group_id=0
Looking at Democracy has a Facebook page and
some videos have been linked from there by
commentors. Carrots and Sticks is listed there.
Bring it to the Table is in the top ten and there
is a solicitation for votes there. (The 92 votes
it had at last canvass -- this system isn't
real-time -- wasn't caused by the mention on that
Facebook page, because it mentions that standing.)
Chatting up the video and the contest in various
fora is apparently legitimate, and is probably what they want to happen.