Wednesday, May 30, 2012

Comparison and choice: When something is better than nothing


Several years ago, I was buying a car, and the dealer had a promotion where the interest rate for financing the car was 0.99%.  It seems a little strange to charge such a low interest rate.  The finance company does not make a lot of money on that rate.  You might think they would attract even more buyers if they interest rate was 0%.  That increase in sales ought to make up for any difference in interest payments on the cars that are financed.

There is some reason to think, though, that a 0.99% finance rate would actually attract more buyers than the 0% rate.  This strange idea comes out of theories of the way people make comparisons when they make choices. 

In a 1993 paper in the Journal of Memory and Language, Dedre Gentner and I found that when people compare two things, they tend to focus on properties that the objects share over those that one object has that the other is missing.  For example, when comparing two cars from different dealers, you might focus on the size of the engines in each car or the trunk space.  If one car has extra rollover protection and the other doesn’t, you might focus less attention on that detail than on the properties that both cars share.  Some studies I did with Doug Medin published in a 1995 issue of Organizational Behavior and Human Decision Processes confirmed that people often ignore these unique properties when making choices.

A paper by Mauricio Palmeira scheduled to appear in the July, 2011 issue of the Journal of Consumer Research carries this work a step further.  He finds that making a unique property of an option comparable can increase the attention given to that property, even if the change makes the option objectively worse.

Let me unpack that.

Let’s go back to the interest rate example.  Suppose you are comparing two cars with a similar price.  They might be very similar along most dimensions.  Brand A might have a slightly larger engine than Brand B.  Based on the engine size alone, then, you might prefer Brand A.  The dealer for Brand A charges a 6% interest rate to finance the car.  If the dealer for Brand B charges 0% interest, then you would think that would be very appealing.  But 0% interest also makes the comparison of interest rates feel like a comparison of something against nothing.  Following all the work I just described, people might actually ignore this dimension and choose just based on the size of the engine. 

But what if you made the finance charge 0.99%?  This interest rate is clearly worse than 0%.  It is better to pay nothing than to pay something.  But, it does make the interest rate a point of comparison for the brands.  That means that offering a slightly higher interest rate might actually entice more people to buy the car.

Does this really happen?

In a number of studies involving choices between products like CD players, credit cards, and yogurts, he found that people tended to ignore a dimension when it had a value of 0.  Making the value slightly higher got people to pay attention to the value and increased choice of the option.

For example, in one study people chose between two credit cards.  Brand A had a higher annual fee than Brand B.  One group was told that Brand A offered a 0% annual interest rate and Brand B offered a 20% interest rate.  About half the people chose Brand A in this group.  A second group was told that Brand A offered a 1% annual interest rate and Brand B offered a 20% rate.  In this group about 3/4 of the people selected Brand A.  So a slightly higher interest rate was actually more effective at attracting people toward Brand A.

What does this mean for you? 

When you are making important choices, it is useful to make a table of the key properties of each option.  The table will have zero values in some of the entries.  You should be aware that you may be prone to ignore those zero values, so a table gives you a way to focus on the values that are missing.

This table might also make you do a bit more research about the choice you are making.  Research by Chris Hsee and his colleagues has found that one reason why people have trouble with missing values is that they find those dimensions hard to evaluate.  Suppose you find out that a particular car has rollover protection and the other doesn’t.  Is rollover protection an important feature to have in that kind of car?  Answering that question requires doing some more work to find out what experts think. 

Ultimately, you want to make sure that you use as much of the information about the options as you can in this decision situations.