Removing the Blank=X option from range voting hurts third parties by biasing the results against them

The effect of removing the blanks (Xs) score-option in range voting (for voters who wish to express "no opinion" about that candidate) is harmful to third parties in a biased way, i.e. it hurts them but does not hurt major party candidates.

Why? Well, this is true both experimentally and theoretically.

Experimentally: Demanding blanks (Xs) be replaced by 0s would have had this effect on the 2004 US presidential range voting study (#82 here): It would have reduced third-party candidates' vote counts by a substantial factor.

WDS was using blanks in his poll but the other pollster Doug Greene was using 0s. Their results may not be fully comparable since in different locations, but ignoring that, Greene's Badnarik average range-score was 7.7 versus WDS's was 9.4 (versus with AV was 0.64 versus with Plurality was 0.32), a 20% decrease.

But Badnarik suffered the least from the blanks-->0 switch: Nader, Peroutka, and Cobb (and probably Calero too) all suffered about a factor=2 drop under this policy switch.

Even with this drop, RV still was giving them all (except possibly for Nader) more votes than AV, i.e. the Nursery effect is true for range voting either with or without Xs, but is more true with Xs.

Theoretically: It makes sense that only the third party (i.e. less well known) candidates are going to be much affected by the X-->0 policy switch, because nobody is going to leave a well known candidate blank!

It also makes sense that all the third party candidates are going to be hurt by the blanks-->zero policy switch, because 0 is the lowest possible score and all the voters who wanted blanks to be blanks, not 0s, are now going to be countermanded. (Meanwhile, all the voters who felt blanks should be 0s are not going to change the picture, since they would have put 0s anyhow if told that blanks are Xs with no effect, not 0s.)

Tactically: What does this tell us? Well, if we want unified third party support behind range voting, it tells me we better go with the blanks=Xs policy. (Without third-party troops we probably cannot hope to get, say, ballot initiatives for RV onto state ballots.)

As far as top Democrat and Republican party leaders are concerned re Iowa 08 – I suspect blanks=blanks leads to higher-quality winners and the possibility of a less-known but better-quality candidate winning is fine from their point of view (he'd immediately become better known!) so I think it is more in the top D's and R's interests to want blanks=blanks in Iowa 08 because its better quality output will maximize their later presidential chances. This is not completely clear, but it is my best guess.


Return to main page