In preparation for job interviews, I've been brushing up a bit on communication theory. The classic Influence: Science and Practice by Robert Cialdini is my current leisure reading.
Cialdini reports on a study done in the '70s which I think is easily worth blogging about.
"Consider what happened when researchers staged thefts on a New York City beach to
see if onlookers would risk personal harm to halt the crime. In the study, an accomplice
of the researchers would put a beach blanket down five feet from the blanket of a
randomly chosen individual—the experimental subject. After several minutes of
relaxing on the blanket and listening to music from a portable radio, the accomplice
would stand up and leave the blanket to stroll down the beach. Soon thereafter, a
researcher, pretending to be a thief, would approach, grab the radio, and try to hurry
away with it. As you might guess, under normal conditions, subjects were very reluctant
to put themselves in harm's way by challenging the thief—only four people did so in the
20 times that the theft was staged. But when the same procedure was tried another 20
times with a slight twist, the results were drastically different. In these incidents, before
leaving the blanket, the accomplice would simply ask the subject to please "watch my
things," something everyone agreed to do. Now, propelled by the rule for consistency,
19 of the 20 subjects became virtual vigilantes, running after and stopping the thief,
demanding an explanation, often restraining the thief physically or snatching the radio
away (Moriarty, 1975)."
I have to wonder how this study would fare with the modern Institutional Review Board, but that aside, it's a pretty strong result, which also happens to coincide with personal experience -- I have both asked and been asked by others to watch their things, and always felt pretty safe in doing so.
The psychologists theorize this is a result of desire for consistency -- if I said I'd watch, my behavior becomes consistent with that of a guardian.
I'd suggest a rational choice explanation that is just as, or perhaps more persuasive.
If I ask someone to watch my things and then they get stolen, the person assigned to watch is either a witness to the theft or is themselves the thief/accomplice. Knowing that I know this, the watcher has an incentive to be vigilant to avoid being blamed if the item goes missing.
So which is correct, psychology or game theory? Probably a little of both. I'd like to see a follow-up on this study performed in different countries with varying levels of trust for strangers. Presumably in a low trust institutional setting, the "watch my stuff" request would be less likely to work -- unless the incentives provided by game theory are sufficiently strong.
In any case, it's nice to read a study that basically affirms the trustworthiness of perfect strangers!