On Preferring A to B, while also preferring B to A

When people evaluate two or more goods separately versus jointly it’s common to see “preference reversals”. In a random survey, for example, people were asked to value the following dictionaries:

  • Dictionary A: 20,000 entries, torn cover but otherwise like new
  • Dictionary B: 10,000 entries, like new

When asked to value just one dictionary, either A or B, the average value was higher on Dictionary B. But when people were asked to evaluate both dictionaries together the average value was higher on Dictionary A.

What’s going on? Most people have no idea how many words a good dictionary has so telling them that a dictionary has 10K or 20K entries just fades into the background–it’s a dictionary of course it defines a lot of words. On the other hand, we all know that “like new” is better than “torn cover” so dictionary A gets the higher price. When confronted with the pair of dictionaries, however, we see that Dictionary A has twice as many entries as Dictionary B and it’s obvious that more entries makes for a better dictionary and in comparison to more entries, the sine qua non of a dictionary, the torn cover fades into importance.

Cass Sunstein collects a bunch of these examples (these two from List and Lowenstein respectively):

  • Baseball Card Package A: 10 valuable baseball cards, 3 not-so-valuable baseball cards
  • Baseball Card Package B: 10 valuable baseball cards
  • Congressional Candidate A: Would create 5000 jobs; has been convicted of a misdemeanor
  • Congressional Candidate B: Would create 1000 jobs; has no criminal convictions

In each case B tends to have a higher value when evaluated separately but A tends to evaluate higher with joint evaluation. When is separate evaluation better? When is joint evaluation better?

There is a tendency to think that joint evaluation is always better since it is the “full information” condition. Sunstein pushes against this interpretation because he argues that full information doesn’t mean full rationality. Even with full information we may still be biased. The factor that becomes salient when the goods are evaluated jointly, for example, need not be especially relevant. Is a dictionary with 20k entries actually better than one with 10k entries? Maybe 95% of the time it’s worse because it takes longer to find the word you need and the dictionary is less portable. We might let the seemingly irrefutable numerical betterness of A overwhelm what might actually be more relevant, the torn cover.

Sellers could take advantage of the bias of joint evaluation by emphasizing information that consumers might think is important but actually isn’t–our computer screen has 1.073 billion color combinations while our competitors only has 16.7 million–while making less salient 6 hours of battery life versus 8 which may in practice be more important.

Personally, I’d go for full information and trust myself to figure out what is truly important but maybe that is my bias. See the paper for more examples and thought-experiments.


Obviously what's best is full information with a high quality filter.

When I want to upgrade my phone I can look at raw specs, or at low quality information filters such as youtube reviewers. Fortunately there are high quality reviews at gsmarena and arstechnica that put the info into perspective that is useful for me.

There is unfortunately not a similar broad a reliable information filter for many other important areas .. such as evaluating behavior of politicians.

What is not clear to me is how you would judge them separately if you are shown both or how you would value one if there is no price context. Some people may value an item very high or very low. compared to others skewing the results.

I suspect there's a tendency to value the two options symmetrically even when they aren't. Thus, in our politics, political opponents, one a Democrat and one a Republican, are usually viewed symmetrically by those not committed to a particular party (i.e., the tribal voters), with only small policy differences; indeed, the usual complaint among many uncommitted voters is that all candidates are the same. Why? I suppose the explanation is that these people either can't or just don't pay close attention, the symmetric view offering an excuse to flip a coin: a dime no doubt, since there's not a dime's worth of difference. This is reinforced by the way political candidates are sold: opponents often claim to support the same things, national security, social security, children, etc. even when their specific policy preferences would provide radially different approaches. Products are sold in much the same way. Thus, competing cars in the same segment are sold not on specific differences but on such amorphous things as "quality" and "comfort" and "safety" and "image". Cars are sold this way for the same reason politicians are sold this way: consumers either can't or just don't pay close enough attention to evaluate real differences. Politicians, cars, dictionaries, there's not a dime's worth of difference. Ironically, isn't this the description of perfect competition: the product is essentially the same and buyers are well-informed, so sellers compete on price. What price does your favored political candidate charge? [Of course, the analogy breaks down on the "well-informed" buyers condition.]

This is one of the great conflicts of modern life. The liberal (of both left and right variety) who wants maximal information and liberty to choose, believing that he will succeed, and the great tradition of constraining people's options, both for the common good, for the benefit of some elites, and because in some meta-rational sense, the proletariat knows that they cannot be trusted to follow their own decisions and instincts. Scoff if you will, but this is exactly the problem that societies have evolved to try to deal with in the face of changing demographics, technology, and markets. Usually imperfectly, often terribly, but you can't wish it away saying that the Ubermensch works best unfettered by the chains of tradition.

No congressional candidate will ever create a job. They will destroy many.

One should hope so!

Economic growth only comes from moving resources from low valued uses to higher valued uses. Destroying low-value jobs is necessary for growth.

Merely creating jobs is not worthwhile if the task is of low value. A politician could create many jobs by passing legislation to build affordable housing while specifying that no mechanical equipment (e.g., backhoes to dig foundations or pneumatic hammers) are allowed. Shovels and old-fashioned hammers only. Yay! Jobs created. But what a waste.

If your local congressman isn't creating jobs, then maybe you shouldn't keep voting them in office.

How can one "prefer" a dictionary without a definition of purpose? If it's to look good on a shelf get the one that isn't torn. If it's to have more words defined, get the fatter. If it's to do the hard crosswords in the British papers get Chambers. If you want etymology get the Oxford.

Exactly. And the Congressman who magically "creates" jobs.... how? For who? Doing what? For how long? At what cost? To what end?

Economics is the discussion of poorly-designed thought experiments.

Came on to say the same - there is no parity in the choice of "has or hasn't a criminal record" and "will create x jobs (is or isn't a magician)". Indeed the lesser number of jobs might reasonably be preferred.

Such a question functions less well as a gotcha than a guide to evaluating the economists (three, so far) who found it meaningful.

Ergo, a useful link after all.

Correct. These choices cannot be divorced from their context, and no survey or experiment can create appropriate context.

In the meta experiment, Sunstein is just trying to convince us that Congressmen create jobs. Always nudging.

Sensible thoughts, so what value the Urban Dictionary?

And don't forget Ambrose Bierce's Devil's Dictionary

.... So "uninformed opinions" tend to be superficial & logically inconsistent -- this ain't no big newsflash.

"cognitive dissonance" also fits easily into this common human behavior.

(the referenced study is paywalled -- but the chances that it was based on a true "random survey" are zero)

If you use a dictionary frequently and it is important to you, you can rather quickly choose between options, especially if you include price information. If a dictionary is trivial to you the cost of acquiring the information to make an informed decision is a waste of time and you just pick based on some prejudice.

The definition of being middle-aged is that you often spend $200 worth of your time to make a decision about a $50 item. In a way that is irrational. But if you will be using that item on a frequent basis the proper $50 item may save you $1,000 over the lifetime use of that item. (My wife values her time highly and will buy two to four such items and then discards or gives away the ones that don't meet her needs. When I suggested such a practice might be wasteful she said that what was wasteful was obsessing over trivial purchases when an informed decision is best obtained through real-life evaluations of what works best for her. Within limits that makes sense. )

It is very hard for a third party to evaluate the rationality of the above actor unless they have a fuller understanding of what the person or persons are actually trading in.

Politicians often seek wedge issues because the voters may not have strong opinions on most issues. The cost of researching the pros and cons of a variety of issues is just too great, especially given that voting itself is mostly irrational. Absent a wedge issue you will tend to vote with the politician or political party that you generally agree with. Politics is about balancing competing issues and you trust that the group you identify with will negotiate in your general best interest most of the time. Or you just don't get that involved. Unless a party can find a wedge issue that will either force you to action or get you to change normal patterns.

I was at an event with the late Senator Fulbright. He was criticized for being progressive on most issues but weak on civil rights. He said that his voters had very strong opinions on civil rights that restricted him from doing what his heart would have preferred but that on other issues they were willing to allow him to lead. He would have had a hard time trying to lead on civil rights. Politics is often about such compromises.

Behavioral economist often seem like children debating whether Thor can defeat The Hulk. They may think there is a rational answer but most adults, if asked, will not spend a great deal of time formulating an answer. Then the behavioral economist will declare most adults irrational because they do not give consistent answers or they tend to disagree with what the behavioral economists have decided is the best answer.

If you want to understand preference reversals, you might want to read articles on salience, decoys, compromise effects, and shrouding. Business use these techniques to change preferences in choice sets or to elicit action.

A good book on this subject is The Handbook on Behavioral Industrial Organization edited by Trembly and chapter 4 on this subject which describes it applications in marketing and competitive strategy.

As for Tyler's wish for unlimited information, consumers often suffer from information overload; in fact, a strategy is to overload consumers with information so that they pick one salient fact, and ignore others. Sometimes a better choice is made with summary statistics, or models where consumers enter information about preferences and the program goes through and ranks, as happens in some software programs which assist consumer in picking health plan options.

Sorry, meant Alex.

As for politics preference reversals, there is probably less reversal due to cognitive dissonance. Ask yourself: How many Trump supporters would today vote for him, given what has transpired since, and had there been disclosure (Russia are you listening) whether the vote would have been the same if there had been disclosure of the various meetings and Jr.s comment.

I should add: cognitive dissonance AND identity.

Well, in line with this particular link you'd have to evaluate Trump and Hillary separately as well as Trump vs. Hillary either-or

The media tried to clinically diagnose Donald J. Trump. Let us use the inverse property: 35 million Americans suffer from obesity, 20 million Americans suffer from drug addiction, 20 million Americans suffer from harsh poverty, 17 million American women have been raped, 20,000 people a year get murdered in America, over 2 million Americans are in Jail. Meanwhile the Hollywood culture that normalizes this society is just as sick. Those that profit are often sicker.

This is to say, the majority of Americans are psychologically ill.

"Sellers could take advantage of the bias of joint evaluation by emphasizing information that consumers might think is important but actually isn’t–our computer screen has 1.073 billion color combinations while our competitors only has 16.7 million–while making less salient 6 hours of battery life versus 8 which may in practice be more important."

For several years this was common with digital cameras, advertising the number of pixels on their sensors. It's not irrelevant to picture quality, but is only one of many factors.

Look at information gained from one example, two examples and three examples of a set. I like that test.

The information in one example is unreferenced and low. Most of the information is revealed with two examples of the set, adding a third example is often not worth much information gain.

Ah, but if the second is a decoy, you might pick item 1 if its difference with item three is salient, and you used item two as a reference to pick item 1 based on a relative price/quality difference. Sometimes decoys are added to a set. Look up "Salience and Consumer Choice" by Bordalo, Gennaili, and Shliefer in the Journal of Political Economy.

Stores will often give three choices in quality knowing that consumers will often pick the middle one.

However not every consumer needs to be informed to make an efficient market. The marginal consumer sets the price. Others can free ride. Think of stock markets. The marginal investor sets the price which in an open market is a fair price. Other consumers can free ride on the information embedded in the price. Do opportunities appear? Yes but the market is quick to correct.

The equilibrium is at a higher price when you have the naive and sophisticate consumer in the same pool. You might want to read: Bounded Rationality and Industrial Organization by Ran Spiegler or Handbook of Behavioral Industrial Organization ed. by Tremblay, or attend a graduate MBA program which has a program on pricing and pricing strategy.

Sure just assume that MC=MR is not true. That markets are not competitive.

The type of price discrimination you describe is difficult to do over the long run. And is increasingly more difficult.

Firms betting that they can exploit an ignorant part of the population have a hard time surviving. They can try to differentiate but that is not seeking naive customers, that is seeking different tastes in the customers

Sure, sure. Ask yourself what assumptions you are making in your theoretical, as opposed to fact supported, claim.

You assume a rational economic actor.

By the way, Dan, the cases and data in fact use MR=MC, but apparently you are not familiar with the materials. As for competitive market claims, you might be surprised to find that the addition of one more payday lender does not lower prices--they just squeeze into the market with a different scheme.

So payday lenders are all earning profits above the risk adjusted rate of return. Their customers have many options but they choose to do business with these businesses?

Or payday lenders serve a segment of the market that others view as too difficult to serve. Or they just don’t like that kind of business. And the profit margins and regulatory burdens keep many out.

So the market may be underserved because this industry is always adjusting to new regulations. The are always gaming the system. New entrants are replacing some who exit.

Yet at the end of the day they do not attract super profits. Their customers are not so much stupid as they are desperate. Traditional lenders do not want to serve this market.

Once upon a time it was common to see stores in poor neighborhoods offer discounts for cash buyers. Why? It was a way to get around limits on the interest rate they could charge. Where the retailers looking for stupid customers or where they gaming the system to serve an underserved community.

You see new entrants finding a new way to play an old game. But risk adjustment returns don’t much change. They don’t compete on price, in many markets because of regulations, so they compete on things other than price. Looking at prices alone does not give you a full picture of how they compete, especially in regulated markets

Dan, I think you are speaking without any knowledge of the subject matter given the answer which you seemed to ignore. So, let me ask you a question that will require some knowledge of the subject matter: let's stay with payday lending ,and your belief that consumer rational economic actors and that not every consumer needs to be rational to make the market work to get to a competitive equilibria maintained by sophisticated and rational actors who protect others by their decsionmaking: so, let me ask: if the market is imperfect (composed of sophisticates and naivettes) why haven't one smart payday lender sought to debias and provide lower rates to attract those rational sophisticates, rather than stick with the naiveties, since you believe so strongly that the sophisticates protect the naiveties by their purchasing decisions. Why do they resist disclosure regulations.

One, payday lenders do not seem to do any market segmentation. They just don’t do much screening to identify differences in customers. Profit margins are so low that rigorous screening is a waste of time. Adding expenses without improving bottom line.

Consumers who go to payday lenders do so because they don’t have access to other lenders.

The market decides the rates and given how low the profit margins are the rates reflect the cost of making these loans.

They don’t compete on the rate because they aren’t making above market returns that can be competed away.

That isn’t too hard to understand

Go do some reading about this subject. Start with a book on behavioral finance.

I would suggest you begin with an average HS Econ textbook. Somewhere you learned a few buzzwords used to explain market anomalies which are pretty useless when you lack a basic understanding of how markets work.

Good luck with those baby steps on your path to understanding

your attacker has a gun. would you prefer that:

(i) he holds it to your head and pulls the trigger, OR
(ii) he hits you with the butt end.

solve for case when gun is
a) loaded
b) not loaded

Congressmen don't create jobs.

Presidents don't create jobs either.

My go to criterion is whether it has an entry for the word "panglossian."
Pocket dictionaries don't, but with handbook sizes, we begin to see this word.

Comments for this post are closed