I don't know as much about how sociologists get around "cheap talk" issues, but it is interesting how psychologists and economists have adopted radically different methodological approaches to it.

Psychology experiments notoriously depend on deception. They get the subjects doing one activity and convince them that it's the experiment, but in reality something else is going on that the researchers are interested in observing. The idea is that if the subjects are unaware of the true experiment they won't be dishonest in their responses.

Econ experiments explicitly eschew deception and instead rely on designing incentive compatible experiments. Where true responses are elicited through a payment structure that rewards them.