Laboratory experiments fail to capture the long time horizons of real cooperation.

Evidence

The first reflects the case where the studies in question overcome:by their very design a particular criticism of laboratory work. All laboratory studies suffer from the fact that they are very short, and the subjects are not tied in any way whatsoever to the enduring consequences of their actions as they would be in everyday life. Rarely are experiments scheduled to last more than a couple of hours, and at most they last for no more than a couple of weeks as in, for example, the Haney, Banks, and Zimbardo (1973) study of a simulated prison. This point would obviously be very serious for any generalizations from laboratory studies where an analogue of long-term political or economic activity was being sought. Apart from brief encounters with itinerant merchants, individuals are inevitably concerned with a time scale greater than two hours. — Diego Gambetta, Trust - Making and Breaking Cooperative…, 550

Categorization

• XREF: This time-horizon point parallels Milgrom’s repeated-game argument that infrequent interaction makes cooperation infeasible — the same structural condition in both lab and long-distance trade. — connects to: Laboratory emergency experiments lack real threats and so offer low ecological validity. — John Drury, The role of social identity processes…, 58