Friday, May 02, 2008

Daily D: Alumni hate pro-parity polls, themselves

A hilariously biased front-page story from today's D:
Alumni have allegedly been subjected to “push polls” favoring the pro-lawsuit candidates in the Association of Alumni election over the last week, according to active alumni.

The article is especially hilarious in light of Daniel Belkin '08's opinion piece Wednesday that alleged pro-parity alumni have a "monopoly" on media attention.

44 comments:

Anonymous said...

How is it biased? Would you have preferred that it said "people posing as neutral researchers have been calling alumni to encourage them to vindicate their rights"?

lolocaust said...

Seriously. Bad press is not biased press. Those who financed the push-polling should have known that Dartmouth alumni are not stupid enough to put up with it.

Furthermore, the Executive Committee's complaints about the College denying them resources are preposterous. They're engaged in a friggin' legal battle, fer Chrissakes. It's in the College's best interests--and no rational agent would act otherwise--to aid its legal defense and deny its legal enemies.

Dreisbach's whining about "fairness" only serves to demonstrate the extent to which he has no real understanding of the principles and values that he so often bandies about. It's one thing to throw around terms like "fairness," "democracy," "parity," "rights," "free speech," etc.--all referring to things that few people would disagree with--and another thing to use them in a way that actually makes sense.

Anonymous said...

I thought this was a fair response re: the mailing lists:

“Why would the College hand their mailing list over to the people who are suing them?” -- Bill Hutchinson

Why, indeed?

Anonymous said...

The Alumni are all lifetime members of the Association. The Association is entitled to know the members and addresses of its own membership, even if it allows the money and database management to be handled by the College.

The College is interfering with an independent Association by preventing it from contacting ITS OWN MEMBERSHIP.

Truth Be Told said...

If the AoA's Secretary-Treasurer were not Dave "Conflict of Interest" Spalding, the Sec-Tres would automatically provide the duly elected leadership of the AoA with the mailing list of its members.

Would a Democrat Governor not provide a list of registered voters to the Repubican party on the grounds that the Repubicans were trying to unseat him in the next election.

Sometimes you wonder if people are living in the real world with these comments.

Anonymous said...

The problem is not the College's denial of email access privileges. Rather it is its discriminatory practice of providing this to some alumni leaders and not others. Even without a lawsuit, the College has used this resource control to influence alumni communications.

The result of such de facto censorship is to drive people into the hands of others who help with alternate communication sources. Provide access and the dependency on outsiders goes away.

karl said...

Democratic governor, not Democrat governor. Learn to speak English, turd blossom.

Anonymous said...

Blossom says:

Karl, a democratic governor is presumably one who espouses democratic principles, even if he or she is a Republican. If he or she is a Democrat, that would in fact be a Democrat governor.

Your followon "blossom" comment discredits all Dartmouth people. Pick up your English.

Anonymous said...

Frank Gado's letter to the editor of the D knocks down a straw man. No one is advancing "the contention that the current mode of nomination and election 'splits' the vote." He found that word "split" in a student-reported article that said "Following the governance change, the nominated candidates would no longer split the vote in favor of the petitioners." That's not good reporting and it does not accurately reflect the Board's report, and it certainly is not a "contention" advanced by the Board. By hanging his entire discussion on the word "split," Frank manages to avoid having to explain why he can't answer the real concern voiced by the Board, which is that approval voting is susceptible to bullet voting:

"Some alumni have noted that this permits a form of tactical voting referred to as “bullet voting” under which organized supporters of a particular candidate are encouraged to vote only for that candidate and no other." It is on page 9 of the report.

anony said...

@ anon. 3:59: why is the Association is entitled to know the members and addresses of its own membership when the College handles the money and database management? We know you believe this entitlement exists, but can you say why?

@ Truth Be Told: How could any secretary-treasurer obtain a mailing list from Dartmouth that Dartmouth did not want to turn over? You say the secretary "would automatically provide" this information – but how would he get it?

Sam 0 said...

anony is right. We're playing to win this time.

karl said...

"Karl, a democratic governor is presumably one who espouses democratic principles, even if he or she is a Republican. If he or she is a Democrat, that would in fact be a Democrat governor."

It's the Democratic Party, numbnuts. 'Democrat' is a noun, not an adjective. You're confused, because 'Republican' is both a noun and an adjective, whereas 'Democratic' is simply an adjective. From Wikipedia:

In the 20th and 21st centuries, "Democrat Party" is a political epithet that is sometimes used by opponents to refer to the party. The current official name of the party is the "Democratic Party."

Anonymous said...

Anony:

Dartmouth College Alumni Relations dept is the agent for the Alumni Association.

The Association used to run several things for itself but slowly these associations have delegated tasks to the colleges: magazines, fundraising for the college, mailing lists, voting...

As an agent, the College raises money for the alumni fund and the alumni endowment in the name of alumni but then withholds the alumni leadership the lists of alumni.

Sounds like Alumni Relations Office is breaching its duty to the Alumni association...

If they wanted to purchase or take over the association, then they are free to try!

Oops, I guess that's what the merger and board-stacking plan was all about.

Since the association is still legally and tactically free - if only for a little while - then the College is on thin ice, as agent, if it wants to withhold the independent association's own membership list.


.

Anonymous said...

Gado Defender (amazing) says:

Anon 5:13 criticizes Gado for raising vote-splitting as a red herring. Yet Gado's entire letter was in response to a D news (not opinion; just the facts ma'am) article that reported exactly that. While perhaps not in the Governance Report, the charge of vote splitting continues to be heard.

Then again, perhaps it is just the D and its reporting. The same article stated "When speaking before the Alumni Council on May 19, Neukom said that the Board’s newly created governance committee would examine questions of the “size and composition of the Board.”" Neukom actually stated that the governance committee had been meeting over the past year on the matter, which would be considered by the full Board in June. A decent guess is that the subsequent letter they received on May 30 from the AoA put them on notice that perhaps, just maybe, they should consult, or appear to consult, alumni before finalizing their recommendations. Thus the deferal to the September meeting.

Anonymous said...

@ anonymous 7:11:

Agency is not a matter of feeling, it involves specific legal requirements. If you could prove that Alumni Relations was an agent of the Association, then you could make the Association liable for acts of the Alumni Relations Department after the fact. What good would that do? An agency relationship does not create some obligation of the agent to follow orders of the principal going forward.

The DAM was never run by the association, it was run by the Class Secretaries Association.

The College never raises money "in the name of alumni." It raises money in the name of the College.

"If they wanted to purchase or take over the association, then they are free to try!" No, it's already been done by others.

Anonymous said...

@ Gado Defender [sic]:

The charge of vote-splitting does not continue to be heard. It came up only once recently, in the poor reporting and poor use of a single word in that one article.

Gado did not respond to the news story as a story, point out the poor reporting, or try to correct the record. He opportunistically latched onto the error and attributed it to his opponents as a way of setting up a straw man.

We can assume that Gado has no response to the real complaint of alumni regarding approval voting, which is that it is susceptible to bullet voting. Wikipedia sayeth --

Bullet voting is a tactic where the voter only selects one candidate, despite having the option to indicate some preference for other candidates. They might do this either because it is easier than evaluating all the candidates, or (depending on the voting system in use) as a form of tactical voting.

If enough voters bullet vote then almost any voting system functions like plurality voting system. This is generally considered to be a poor result, as many voting systems are intended as reforms or improvements which avoid or minimize some of the alleged disadvantages of plurality (aka First Past the Post). However, if voters bullet vote as a conscious strategy to express their meta-preference that candidates with strong support are preferable to "compromise" candidates with broad support, then it may not be a problem.


Note that nobody has said that bullet voting necessarily allowed Smith to win, only that the system is subject to bullet voting.

Truth Be Told said...

Let's recall first that SMITH RECEIVED VOTES FROM 55% OF THE VOTERS IN THE LAST ELECTION. So if there was bullet voting, that's a lot of bullets!

That said, the College refuses to release the results of the voting for the losing candidates. Why don't they do that? If they did, we could settle once and for all whether the losers split the votes or not.

Recall also that Smith was the only candidate who did not support the failed constitution, and he was the only one who did not parrot ad nauseam Jim Wright's "the College is in great shape" mantra. So it is natural that people critical of the administration and its management of the College concentrated their votes on Smith.

Once again, it is worth making the point that the fault lies with the AC Nominating Committee, which refuses to nominate anyone other than down-the-line supporters of the Administration.

If it did so, it could choose responsible critics of the College and provide the alumni with a real choice, rather than obliging the petition movement to put someone forward.

Anonymous said...

The concern about "bullet voting" giving one candidate an advantage is a red herring. It can easily be offset by "shotgun" voting wherein those opposed to one candidate can encourage the electorate to vote for everyone except their opponent. This form of "non-approval" voting simply says vote for all those you find acceptable and reject those who are not. With this approach, if a majority of voters had been truly opposed to a petition candidate, he would not have won.

As noted above, Smith received approval by 55% of the voters, not just the votes cast. A clear win. The system is not broken; people are just unhappy with the outcomes.

anony said...

Note that nobody has said that bullet voting necessarily allowed Smith to win, only that the system is subject to bullet voting.

That part was missed by the last two writers.

The issue is not "splitting" votes, Truth Be Told, unless you're Frank Gado looking for a red herring.

"Shotgun voting" is possible, but many alumni would consider it distasteful or unethical, something only the Hanover Institute would propose. It is totally inappropriate for elections involving a trustee instead of a political campaigner like Smith.

The short-lived experiment in approval voting has allowed an apparently unexpected growth in campaigning. It makes sense to try a different system.

Robert Mugabe '46 said...

If the establishment wants to win an election, it should nominate someone who can win and stop making excuses about the voting method.

Anonymous said...

If the board is dissatisfied with the process alumni use to select their nominees, it should instruct alumni to improve the process.

Anonymous said...

Is the board allowed to "instruct" alumni?

Truth Be Told said...

Has everyone forgotten the referendum on the constitution in the fall of 2006, which proposed to revise the system of voting in trustee elections?

The Board endorsed the constitution put forward by the AGTF - and the alumni, in their infinite wisdom (and with a 38% turnout!) decided that they liked the present system of approval voting just fine, thank you.

I guess that the Trustees only trust the alums when the alums do what the trustees want them to do.

Morgan Tsvangirai '74 said...

Nominate a winner.

Anonymous said...

The board is allowed to instruct alumni. Alumni who don't wish to adopt the proposal, won't.

Anonymous said...

Except that the Board has stated that if the alumni do not agree to the changes the Board wants in the alumni's election rules, it will administer the alumni elections themselves.

Anonymous said...

That's what they said, but realistically they are probably reconciled to giving eight seats to petitioners if they get their way with the lawsuit.

Anonymous said...

Actually the board said if the changes are not made before the next election, the board will have the College run the election.

BROKEN LADDER said...

"Bullet voting" is not a problem with Approval Voting, since the ideal strategy is to vote for the person you'd vote for with plurality ("pick just one") voting, plus everyone else you like better.

So if there are two candidates you like pretty well, and one that you strongly dislike, it can be in your best interest to vote for your second favorite, since the fear of helping your second favorite beat your favorite is smaller than your fear of getting your least favorite.

Just imagine you're one of those Nader voters back in 2000 who felt that showing support for Nader was more important than helping Gore beat Bush. With Approval Voting you could have done both, and that could easily have made all the difference.

The great thing about Range Voting (Approval is the simplest form of Range) is that it is so resistant to strategic voting.

Robert Z. Norman, a Dartmouth professor emeritus of mathematics and an Approval Voting advocate, does not believe this had anything to do with problems with Approval Voting. He says:

"the claims about bullet voting in the Dartmouth Alumni
election remind me that with a per voter average of voting for 1.8 candidates,
the proportion of bullet votes has to be fairly small. The alternative, as I
mentioned in a communcation a day or two ago, is that nearly everyone voted for
one or three candidates but not two. Unlikely as that might be, it would suggest
that most of those who voted followed a strategy of either voting for the
petition candidate or voting for all nominated candidates, in which case
Richie's claim that the opposition was disorganized falls apart, as does the
claim by some of the Alumni Council people that in a 1 on 1 situation the
petition candidate would been defeated."

Anonymous said...

Approval voting indeed can result in vote-splitting because it presents an inescapable conflict when you as a voter prefer one person on the board-nominated slate, but also really want the petition candidate to lose. If you bullet vote for your preferred candidate, you won't help another board-nominated candidate win -- but you might help the petition candidate win. If you vote for all the board-nominated candidates, you will help defeat the petition candidate -- but you could cause your preferred choice to lose.

55% for the petition candidate Stephen Smith sounds like a lot of support, but in fact any one of the board-nominated candidates might well have defeated the petition candidate in a one-one race or with the preferential voting system used to elect Dartmouth's student body president.

Anonymous said...

A tag on that last post. The reason 55% with approval voting doesn't necessarily mean that this candidate would win in a one-person, one-vote election is that several candidates could gain the acceptance of more than 50% of voters. Imagine the Republican Party presidential field in January, say. You might have had McCain with 75% acceptance, Romney with 70% acceptance, Giuliani with 67% acceptance, Huckabee with 55% acceptance and Paul with 37% acceptance. But if you had a dynamic where half of Huckabee supporters cast bullet votes for their guy rather than vote sincerely, suddenly that 55% could become a victory.

Anonymous said...

Broken Ladder, Professor Emeritus Norman was talking about voting "splits," not bullet voting when he said that, wasn't he? Your bullet voting excuse depends on having "one that you strongly dislike" and having people vote against that candidate. As a matter of fact, many alumni are not aware that they must do this for their voting to accurately reflect their preference (the instructions say to vote for as many as you like but do not say to vote for people you do not care about if there is one you dislike), and of those who are aware, many consider it improper to vote against candidates they dislike, especially in a race for "trustee," i.e. an honorary and non-representative, non-constituent position outside of government.

Greg said...

Regardless of whether Approval Voting elected the "right" winner in this election, it seems to have left the impression that some voters are gaming the system at the expense of others. Whether or not this is true, the system appears to be very susceptible to this impression. If the impression generates so much distrust and dissatisfaction with the process, that might be grounds enough to try out another system. I personally recommend giving IRV a shot to see how it compares in practice.

BROKEN LADDER said...

Approval voting indeed can result in vote-splitting because it presents an inescapable conflict when you as a voter prefer one person on the board-nominated slate, but also really want the petition candidate to lose.

Approval Voting is free from the vote-splitting problem in that if a zillion clones of your approved candidates join the race, that does not cause them to lose. The conflict you speak of is a strength of Approval Voting, since it causes "revealed preference" about real strength of preference. That is part of why Approval Voting is so resistant to the negative effects of strategic voting.

55% for the petition candidate Stephen Smith sounds like a lot of support, but in fact any one of the board-nominated candidates might well have defeated the petition candidate in a one-one race or with the preferential voting system used to elect Dartmouth's student body president.

The results of a one-on-one race are not conclusive about which candidate is socially better. In fact it is even possible to have a cycle of preference where X is preferred by a majority to Y, and Y to Z, and Z to X - so the argument that X must be better than Y if preferred by a majority to Y is wrong. Approval Voting is usually better than "Condorcet" methods, where we pick the candidate who would have beaten all others by a majority - especially if there are many strategic voters.
http://rangevoting.org/BayRegDum.html
http://rangevoting.org/StratHonMix.html

I suspect that he system used to elect Dartmouth's student body president is Instant Runoff, which is far worse, and much more susceptible to strategic voting.

Broken Ladder, Professor Emeritus Norman was talking about voting "splits," not bullet voting when he said that, wasn't he?

I asked him about the prevalence of bullet voting, and that was part of his response, so...

Imagine the Republican Party presidential field in January, say. You might have had McCain with 75% acceptance, Romney with 70% acceptance, Giuliani with 67% acceptance, Huckabee with 55% acceptance and Paul with 37% acceptance. But if you had a dynamic where half of Huckabee supporters cast bullet votes for their guy rather than vote sincerely, suddenly that 55% could become a victory.

Or, imagine a bunch of Nader fans who "approved" of Nader and Gore, but decided to vote only for Nader, and then got Bush, whereas they could have had Gore by also voting for him. Bullet voting hurts them in that case. It's only really safe to do it if you're bullet voting for your favorite of the apparent front-runner candidates. If Huckabee was a front-runner under sincere Approval Voting, then he'd already be an excellent candidate as far as the electorate was concerned.

Plus this scenario is totally unrealistic, since in real life you'd have some Huckabee supporters bullet voting - but certainly not all - and you'd have some supporters of other candidates bullet voting. That would mitigate the effect.

Regardless of whether Approval Voting elected the "right" winner in this election, it seems to have left the impression that some voters are gaming the system at the expense of others. Whether or not this is true, the system appears to be very susceptible to this impression. If the impression generates so much distrust and dissatisfaction with the process, that might be grounds enough to try out another system. I personally recommend giving IRV a shot to see how it compares in practice.

Greg is a rather unscrupulous IRV supporter with little regard for facts or science. IRV is a rather horrible voting method that just barely improves over ordinary plurality voting, but comes in far beyond Approval Voting.

IRV is also subject to strategic exaggeration, and is much more harmed by it. But because it is a ranked method, it takes a much deeper analysis to build a case that there was strategic voting with IRV, whereas with Approval Voting, people will tend to look at bullet voting as a clear case of strategy (even if it is not, or if it represents strategy that wouldn't have made a difference anyway).

So Greg is basically arguing that ignorance is bliss. We should get a less representative election outcome so long as it obscures strategic voting and makes us feel warm and fuzzy about our election. That is the level that IRV supporters will stoop to.

Anonymous said...

Is that what "IRV" stands for, Instant Runoff Voting?

Is that what the Board said the Association must use if it is to produce acceptable nominees? If not, the debate is pointless.

Anonymous said...

Om page 35 the report, the board recommends elimination of approval voting, given that so many people have found it confusing and divisive (page 8) and prone to bullet voting (page 11). It suggests moving to a plurality vote, with one vote per person, with one board-nominated candidate.

Of course what could get tricky is if there are two petition candidates, allowing split votes and non-majority outcomes. Imitating the student government with its "instant runoff" preferential voitng system that has proven itself well there and many other institutions would make a lot of sense.

BROKEN LADDER said...

Om page 35 the report, the board recommends elimination of approval voting, given that so many people have found it confusing

Approval Voting is very simple. It's just like how we vote ordinarily, except we change the rule from "vote for one" to "vote for one, or more". This is empirically supported by the fact that Approval Voting experimentally leads to fewer spoiled ballots, whereas IRV increases them by about a factor of seven.
http://rangevoting.org/SPRates.html

..and divisive (page 8)

What exactly does that vague criticism even mean? Voting is about picking the most representative options, given that people are naturally already divided on their preferences.

If you want to talk about divisiveness, consider that IRV tends to elect extremists.
http://rangevoting.org/IrvExtreme.html

and prone to bullet voting (page 11). It suggests moving to a plurality vote, with one vote per person, with one board-nominated candidate.

Oh the painful irony. Even though the average was 1.8 votes per voter, they falsely claim that bullet voting was a problem - so the solution is to force bullet voting.

Imitating the student government with its "instant runoff" preferential voitng system that has proven itself well there and many other institutions would make a lot of sense.

How has IRV "proven" itself? It is more complicated, more susceptible to strategic voting, and far less representative than Approval Voting. Ranked ballots help to conceal strategic voting, but that's just sweeping the dirt under the rug.

Anonymous said...

With all due respect, Broken Ladder, approval voting is not working well for these elections.

BROKEN LADDER said...

With all due respect, Broken Ladder, approval voting is not working well for these elections.

What evidence do you have that approval voting wasn't highly successful at Dartmouth? All the criticisms I've seen are based on misinformation and misunderstanding, such as the totally false claim about bullet voting.

Anonymous said...

Reason for Misunderstanding:

Those who believe approval voting is not working well are those whose candidates are not winning. In order to justify changing the system, they spew misunderstanding.

Last time around they tried to empower a select group with managing who ran against whom, claiming this would result in head-to-head elections when it would not. (The failed AGTF proposal).

This time the governance committee tried to split outsider votes by making it easier to have multiple petitioners against a single nominated candidate, with a traditional "one vote" system. Thankfully the trustees did not approve this recommendation in its entirety. What was adopted is still sausage-meat that asks for clear majority winners when this cannot always be accomplished in multi-candidate elections.

BROKEN LADDER said...

Hats off to you for that dose of reality. It basically sounds like approval voting was done away with because it does work. Those in power don't like that.

democratic not "democrat" said...

The system really did fail in the last two elections, since it selected individuals who are in actual fact unqualified for the position. Those two managed to hide or disguise their disqualifications, however, and there is little reason to believe any other voting system would have produced a different result.

BROKEN LADDER said...
This comment has been removed by the author.
BROKEN LADDER said...

democratic:

Obviously if voters are ignorant about the candidates, they will make poor decisions. In some cases, an ignorant electorate could pick a candidate with Approval Voting who was unqualified, but who would have lost under plurality voting, because of vote splitting between himself and a similar candidate. In other cases, the situation could be the opposite. An unqualified candidate could run against two qualified and fairly similar candidates and win with plurality (again because of vote splitting) but lose with Approval Voting. So this problem is pretty independent of the voting method used.

However, Warren Smith's Bayesian regret calculations did employ "ignorance factors", and found that Range and Approval voting did well even with large amounts of ignorance.
http://rangevoting.org/BayRegDum.html