Useful set of slides. I'd add that survey experiments are often also used for policy advice in ways that overlook the general political equilibrium. The two main issues are that (i) people self-select into the messages they receive (whereas in the experiment they are forced to engage with the
2 likes 1 replies
?
Replies
T
message) and (ii) there is counter-framing in the real world (which requires us to understand how people deal with conflicting cues/messages). It may well be that a message produces some desired effect in an experiment, but fails to do so in the real world because the GE effects trump the partial