Monday, April 02, 2007

The Decoy Effect

The Washington Post has a useful column today attempting to introduce an important marketing principle in the context of the Presidential candidates, now that the field in both parties has clearly separated into two major contenders and a host of minor players.

The Post columnist makes the case that the also-rans in both parties have utility to both main contenders because the also-rans can be painted in such a way as to create dominance for one of the main contenders, crystallising support from those voters who are concerned about the features emphasised in the comparison. To an extent this is true, in that this creates a comparison which the relatively superiour candidate will dominate; but the real power of the decoy effect is that it can resolve the apples and oranges analysis paralysis between two strong contenders when one of the two clearly dominates a third choice at least weakly--that is, is better in at least one aspect and at least equal on all other aspects. To the extent that the major players both weakly dominate another candidate, there is theoretical room for both to use that third candidate as a decoy, but in practice the advantage will go to the major candidate who dominates the third candidate more strongly.

Unfortunately, the column doesn't really clarify the most significant component of the decoy effect--relative inferiority on one variable as the driver in resolving the multivariable conflict. The point of the decoy is that it crystalises attraction to the candidate who dominates the decoy at least weakly on all aspects and the other major contender on the issue seen as the decoy's strength. It isn't enough to dominate a decoy on some qualification which isn't seen as the decoy's main attraction; one must be perceived as dominating the decoy on some relative point of strength.

For example, drawing attention to Kucinich's military inexperience in contrast to Clinton's relatively superiour competence (marginal as that may be) only resolves the conflict for those people torn between Kucinich and Clinton; it doesn't help resolve the apples-to-oranges comparison between those undecided between Obama and Clinton on other issues unless Kucinich dominates Obama on the military issue as well. The point of the decoy effect is that the comparison which shows a clear advantage to one major candidate over a third candidate without similar relative advantage to the other major contender draws support for the relatively superiour candidate out of proportion to the importance of that particular issue to the consumers' overall menu of concerns.

Vedantam's column also confuses the issue by bringing another marketing phenomenon into the analysis without clarifying it: the compromise effect, in which consumers will gravitate to the intermediate option of any three offered. Talking about Nader vs. Gore vs. Bush possibly raising support for Gore over Bush vs. Gore alone is an example of the compromise effect, not the decoy effect, since Gore did not dominate Nader more clearly than Bush dominated Nader on most issues; instead Nader's more leftist positions generally made Gore seem a more moderate choice and increased his overall appeal.

No comments: