Political Scientist Analyzes Why Chris Daggett Received Fewer Votes than Polls Had Predicted

David P. Redlawsk, a Political Science professor at Rutgers, has this op-ed in the Newark Star-Ledger, analyzing why New Jersey independent gubernatorial candidate Chris Daggett only got 5.7% of the vote when polls close to the election had him at 20%. Redlawsk is also Director of the Rutgers-Eagleton Poll. He finds, in post-election polling, that 45% of the people who had thought they would vote for Daggett did not actually vote for him because they thought he couldn’t win.


Comments

Political Scientist Analyzes Why Chris Daggett Received Fewer Votes than Polls Had Predicted — 4 Comments

  1. Approval voting NOW for single offices – vote for 1 or more, highest win = more likely to be a *moderate* in a time of super-extremist party hack Donkeys and Elephants.

  2. While approval voting allows for a genuine first choice, it encourages an approval for the favorite of the top two polling, regardless of actual satisfaction. Of course, however, ANYTHING is better than plurality when more than two candidates run. So, Approval >> Plurality.

    This plague of disingenuous voting and fear of voting for the long shot pulls at my heartstrings. Though, my favorite single member election system has to be Condorcet. I think the name is unfortunate, however. Should have named it instant round robin, so at least then laypersons have a clue what’s going on. Since the notorious paradox is rare and unpredictable people have no incentive to vote insincerely since the paradox conditions are the ones that allow insincere voting to help.

    Just refer to Nicolaus Tideman in Collective Decisions and Voting(pg 154). “If voters have such expectations [paradox is rare/unpredictable] and … Condorcet [is used], then they have no incentive to vote strategically.”

    The Nanson method can be handcounted (combo of Borda and IRV) and is a Condorcet system. It’s not as compatible with some of the fair voting criteria, but remember that this mainly applies to when there are cycles, which is rare.

    Condorcet, like Approval, also picks the candidate that lingers along the median voter. However, this is with genuine voting in the modeling. With insincere voting, Approval may not be able to accomplish this.

  3. I think it warrants saying again that anything is better than plurality when we’re forced with a single office election with more than two candidates. We should all support IRV over plurality.

    There are several things I like about IRV. It’s more resistant to vote splitting than plurality, it’s more expressive, and the winner has greater than 50% of the vote (at least among the remaining active ballots).

    I’m a bit careful on discussing failed/passed voting criteria, because I think that can be misleading. The idea that two methods fail a specific criteria does not mean that they fail it with equal prevalence. IRV is frequently slammed for violating monotonicity, but I think it’s more important to look at the frequency of this event. I think this idea falls over into tactical voting as well, though clearly IRV is more susceptible than any Condorcet method to tactical voting. The question though is how often.

    I also recognize that IRV has a following in the US in terms of alternative system momentum, but I think it’s important not to marry ourselves to a particular system without evaluating it critically and looking at alternatives.

    Some criticisms of IRV:
    (1) While an expressive system since it allows the voter preference, voters may not always get to have the expression shown if transfers go to candidates that have already been eliminated. With Condorcet-Nanson all preferences are instantly active.
    (2)IRV does not always pick the Condorcet winner. (While likely not to be an issue in most races where the winner is dominant, this will become important in races that are closer with multiple candidates)
    (3) IRV can behave erratically, especially with >3 candidates. See voting simulation (assumes genuine voting): http://zesty.ca/voting/sim/

    On a final reference to Approval voting, the method is extremely easy to calculate, and is a definite perk. One downside is that it is less expressive. The benefit of the ease of calculation may be mitigated however by the fact that insincere votes are being tallied. What’s the point of the result if it is not a genuine one?

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.