"The importance of irrelevant alternatives
- Posted by:
- Economist.com | WASHINGTON
HOW educational is our humble newspaper? So educational, it appears, that even our subscription pitches embed important psychological truths. In a lively and illuminating TED talk, behavioural economist Dan Ariely explains that he was intrigued by a puzzling set of options our website once offered prospective subscribers: You could get access to all our web content for $59, a subscription to the print edition for $125, or a combined print and web subscription, also for $125. Intuitively, the offer of the print-alone option seemed absurd. Nobody, even a staunch luddite, would rationally choose to forego web access when it costs nothing extra, so why even list it? Why not just say that the print subscription also includes access to the web archives?
I'm not sure whether this is what the paper's marketing people had in mind—though subscription offers are no longer structured this way—but Mr Ariely discovered something intriguing when he surveyed students about which option they preferred. Predictably, nobody chose print subscription alone; 84% opted for the combination deal, and 16% for the web subscription. But then he repeated the poll without offering the unpopular print-only alternative—after all, nobody was choosing it, so what difference could it make to leave it out? The second time around 32% wanted the print subscription, while 68% preferred to go web-only. It appears that the presence of the clearly-inferior option altered the decision process by making the combined web and print subscription seem like a better deal. And Mr Ariely has discovered the same effect in a range of situations, covering travel decisions and even our ratings of physical attractiveness. (Apparently, you'll seem more attractive if you are accompanied by a sibling who looks similar to you, but a bit uglier.)
This is intriguing from a political perspective because it seems to cut in the opposite direction from a well-known result in political science, the third-party spoiler effect. This is most familiar from the 2000 presidential election, when it's generally believed that Ralph Nader cost Al Gore a victory in Florida by drawing off support. While this is intuitive enough, the formal account that explains this effect invokes rational voters who are selecting candidates closest to them in ideological space. Of course, reflective voters are also aware of this effect, which is one reason third parties have so much trouble in voting systems like the one employed in the United States. American voters are wary of throwing a race to the candidate furthest from them ideologically by splitting support among several more closely situated candidates. This model surely captures a big part of the truth, but Mr Ariely's comparison effect suggests that the presence of these dark horse candidates may also work in the other direction to some extent. That is, centrist voters for whom Mr Nader is too far left might nevertheless be induced by his presence on the ballot to look more favourably upon Mr Gore.
This is, in many ways, an old insight. An old saw around Washington has it that the function of Greenpeace is to make the Sierra Club look mainstream, while Earth First! exists to do the same for Greenpeace. I once put it to an acquaintance who works with People for the Ethical Treatment of Animals that their outlandish and extreme ad campaigns needlessly alienated many potential sympathisers. He replied that the goal of PETA was not to make people love PETA, but to set down a marker defining the far pole of discourse. More moderate reformers might cringe at the thought of being associated with the purist animal-rights position, but whatever they propose is likely to sound reasonable by comparison. Such is the uneasy symbiosis of every political movement with its own fringe.
Still, if this comparison effect has anything like the magnitude in other contexts that Mr Ariely found with his students, this is obviously a potentially fruitful empirical programme. It suggests, for instance, that a general election candidate might sometimes benefit more from winning a primary campaign against a less appealing member of his own party than from sailing to an unchallenged nomination, even if primary fights inevitably leave contenders a bit bruised and sapped of resources. It also serves as one more reminder that opinion polls—and the democratic process in general—can't necessarily be thought of as measuring some pre-existing popular preference.
In a way, perhaps, that is Mr Ariely's most important point. We think of these forms of evolved irrationality as cognitive "biases"—but "bias" implies an objective reaction away from which one is biased, the "true preference" that a perfectly designed poll or voting system would uncover. The work of political scientist Kenneth Arrow cast serious doubt on that enterprise at the level of collective choice, but behavioural economics ought to leave us with questions about it even at the individual level. Is the choice people make between two subscription options more genuine than their choice from three? Or is it sometimes a mistake to speak of a "true" preference that exists within the individual, divorced from the contingencies of a particular choice scenario? If so, this seems to have some interesting implications for democratic theory. The extraordinary malleability of our expressed preferences in the face of trivial alterations in the choice context sits uneasily with the idea that there is such a thing as "what the American people want", the job of our political system being (ideally) to discern what this is and bring it about."Me:
Don the libertarian Democrat wrote: