At this point, most of the stuff in a corrected edition of _Influence_ would have a bunch of asterisks in it or have to be deleted for failing to replicate or being fraudulent. I wonder how well this old social psych priming/influence experiment has held up...
This kind of human interaction is way too quick to have any deeper thinking involved. Because brains are prediction machines, if something unexpected happens that requires a quick response, it will be fully handled via autopilot.
If it was done via text messages I would expect a different outcome.
It's more effort to figure out whether a requested small favour is really needed than it is to just go along with it. Also the extra effort spent to give a reason, regardless of the quality of the reason given, is itself a signal indicating the importance of the request to the requester. You're more likely to help someone, all else being equal, if they really need the help. This seems pretty straight forward to me.
Their model of scripts and mindlessness is only one of many possible explanations of the behavior they're seeing. And I don't even know where to start with the sending letters to random doctors thing.
Is this actually a thing that was/is taken seriously?