The lab, the truth, and the real world
Does the lab tell the truth about behavior in the real world? That is a question that concerns economists ever since the lab is used as tool to understand human behavior. The first properly designed experiments were conducted by Vernon Smith, who later was rewarded the Nobel Prize for it. These experiments, using students as subjects, tested behavior in an artificially constructed market game. It turned out that laws of supply and demand held pretty well in the lab, and so the lab proved to be a fruitful area to test economic theories. Critics argued that it is hard to transfer quantitative outcomes of the lab to the real world. After all, the lab is different from the real world in so many dimensions. Why would general laws of behavior differ in the lab and field? Why would the laboratory not tell the truth?
The Dictator Game
Well, one area in which the laboratory may be off in its predictions is when it comes to measuring pro-social behavior. Consider the Dictator Game. In this game, a subject receives 10 euros and can choose to share anything with an anonymous receiver. Classical economic theory, assuming everyone cares only about him or herself, predicts that no money is given by the dictator. It turns out that about sixty percent of the dictators give something more than zero, and these gifts average roughly 2,50 euros. So, despite all the incentives not to give, subjects (students and non-students alike) do sacrifice some of their earnings for someone else. At first sight, this seems like evidence showing the existence of altruism. However, the “general laws of behavior” may not hold in this case. A fundamental problem in the lab is that the subjects know that their behavior is being watched and recorded by a scientist. Could this explain why gifts in the Dictator Game are so high?
The researcher designs a set of very similar experiments that start in the lab with student subjects, changing one aspect a time, to end in the field with “real people”, unaware of the experimenter. This method is new, so the literature offers only a few of these studies. One study has found that students donate as much money to charity in the lab as they do in a real life setup of a charity event. In another study, subjects in the lab could send a thank-you-letter with cash to a volunteer of a university. They returned these as often as subjects who “accidentally” got a similar letter in their mailbox. On the other hand, baseball card dealers who are invited in a lab show to be trustworthy in a game that measures trust. These same dealers, however, rip off customers who rely on their trustworthiness when buying baseball cards.
'As more field experiments are conducted, other challenges are exposed'
The debate continues
It is too early to tell if the lab systematically overestimates pro-social behavior, because different settings need to be tested. In the meantime, however, it seems as if the experimenter demand effect is not always of big concern. But, as more field experiments are conducted, other challenges are exposed. For example, a popular topic in economics and psychology is to compare altruism between the rich and poor. Lab findings indicate that the poor are more altruistic. But, a recent field experiment that “accidentally” sent letters with and without cash to members of both social classes, showed the rich returned more. This is not to say that the rich are thus more altruistic. It turned out that the poor suffered from stress that hindered them to return envelopes. Such effects can only play a role when subjects have to plan a real activity, and this is typically absent in the lab. To conclude, the debate on whether the lab tells the truth about the real world continue