Today I heard about the following experiment.  Subjects were given a number to memorize.  Half of the subjects were given 7 digit numbers and half were given 2 digit numbers.  The subjects were asked to walk across a hallway to another room and report the number to the person waiting there.  If they reported the correct number they were going to earn some money.  On the way, there was a cart with coupons available that could be redeemed for a snack.  There were coupons for chocolate cake and coupons for fruit salad.  Subjects could take only one or the other before proceeding to the end of the corridor and completing their participation in the experiment.

63% of the subjects memorizing 7 digit numbers picked the chocolate cake.

Only 49% of the subjects memorizing 2 digit numbers picked the chocolate cake.

I can see two possible explanations of this.  One is very interesting one is more prosaic.  What’s your explanation?  I will post mine, and more information tomorrow.

Update: The experiment is in the paper “Heart and Mind in Conflict:  The Interplay of Affect and Cognition in Consumer Decision Making” by Shiv and Fedorikhin.  Unfortunately I cannot find an ungated version.  It is published in the Journal of Consumer Research 1999.  I heard about the experiment from a seminar given by David LevineHere is the paper he presented which is partially motivated by this and other experiments.

Our interpretations are similar.  The interesting interpretation is that we have an impulse to pick the chocolate cake and we moderate that impulse with a part of the brain which is also typically engaged in conscious high-level thinking.  When it is occuppied by memorizing 7 digit numbers the impulse runs wild.

The less interesting interpretation is that when we dont have the capacity to think about what to choose we just choose whatever catches our attention first or most prominently, independent of how “tempting” it is.  One aspect of the study which raises suspicion is the following.  In the main treatment, the coupons were on a table where threre was displayed an actual piece of chocolate cake and and a bowl of fruit salad.  This treatment gave the results I quoted above.  In a separate treatment, there was just a photograph of the two.  In that treatment the number of digits being memorized made no difference in the coupon taken.

The authors explain this by saying that the actual cake is more tempting than a picture.  That’s plausible, but it would be nice to have something more convincing.  Would we get the same result as in the main treatment if instead of chocolate cake and fruit salad we had yogurt and fruit salad?