To economists someone’s answer to a question about value shouldn’t change based on how they were questioned. But, subjective value depends on the question asked.

Casino experiments

If you visit a Las Vegas Casino today you’ll see computer video screens everywhere. Every casino floor is covered with computers, video poker machines, video slot machines, even video simulations of table games like blackjack. Each one of those games seeks to capture your attention with brilliant colors engaging animations and a host of cues that signal money and prizes.

Their sounds merge into a cacophonous symphony of computerized noise, but if you would walk through Las Vegas casino floors in March 1969 you would’ve found only one computer off by itself on a balcony in the 4 Queens Casino. There stood a large cabinet containing a PDP-7 Minicomputer and attached a much smaller video screen and a gaming table.

Like other tables in this casino this one displayed a roulette wheel, but sitting on the table was a computer keyboard that allowed casino patrons to gamble electronically as part of a remarkable experiment. This experiment was run by the psychologist Sarah Lichtenstein and Paul Slovic who wanted to test a particular phenomenon identified in their laboratory.

Subjective value variability

They had found that the subjective value per dispense assigned some economic gamble but the span of a five dollar roulette wheel seem to depend on how that gamble was presented. Such variability didn’t make sense to economists who thought that value was something more fundamental and invariant.

To economists someone’s answer to a question about value shouldn’t change based on how they were questioned any more than the price of a pain in old Sears Roebuck catalog to change each time someone flipped to its page. Lichtenstein and Slovic knew that economists would be more accepting of their results if real money was involved. So they brought a mini laboratory to the casino and recruited real casino patrons to play gambling games for real money.

At the beginning of each game the player walked up to the computer and chose their stakes, from $0.05-$0.25 per chip. Remember this was 1969, then the players played two games. In the first game which are called the choosing game the players were shown pairs of beds were match for expected value. In each pair one of the beds had a higher probability of winning while the other was associated with larger potential rewards.

The player chose one of the two bets in a casino employee was serving as a dealer spun the roulette wheel to resolve the vets. In the second game which I’ll call the pricing game. The players were just shown one bed at a time say a 75% chance of winning 10 chips in a 25% chance of losing 2 chips. They’ve then indicated that what price they be willing to sell the bet back to the dealer and give up the chance to play.

So they might think a 75% chance of winning $10.25 percent chance of losing to that’s pretty good I’ll give it back to you for five chips. These games went on for as much as several hours and some people won or lost as much as $80 again remember that this was 1969 so $80 was quite a lot of money. Now I want to establish some intuitions about how people should play these two different games.

Reversing preferences

In the choosing game the players just indicate whether they choose gamble A, gamble B, gamble C or gamble D the and so on. If they choose A over B and that implies that they must be the more valuable gamble and it should command a higher price in the pricing game. In short each players choices should line up with their prices, but they don’t.  The players often shows a high probability gamble over a high reward gamble, but then later in the experiment they sold the high reward gamble for the greater price.

Their preferences often reversed, depending on how the question was being asked. Over the next several decades psychologists and economists conducted study after study of what we now call preference reversals the psychologist showed repeatedly that people reverse their preferences depending on whether they were choosing or pricing.

In general, people are more likely to choose gambles with the highest probability of winning, but people also reported that gambles with the highest magnitude of winning should command higher prices. Economists tried really couldn’t make these preference reversals go away with relatively little success. Subjective value depends on the way we ask  the question.