Leon Festinger and colleagues in their 1956 groundbreaking field study “When Prophecy Fails” followed the actions of a group committed to the belief that flying saucers would transport them to a special place. Those who committed to the idea gave away their possessions, separated from family, and otherwise gave themselves over. Festinger et al hypothesized five factors that would influence the outcome: (1) manifest conviction, (2) active commitment to the conviction, (3) the conviction must be amenable to unequivocal dis-confirmation, which (4) occurs, and (5) social support post-dis-confirmation. His researchers infiltrated the group and recorded their observations — btw, I’m quite certain this study would be considered unethical today because of its obvious potential to influence and embarrass participants (this was done in the Milgram era). After the dis-confirmation, most of the folks at the center of the group changed their behavior to become ever more flying saucer partisan, proselytizing as they had not before, “doubling down” on their commitment and seeing the like-actions of their comrades as confirmation of the correctness of their opinions. A factor key to the actions of partisans post-dis-confirmation was the strength of association to the social group. Those most committed to the message sustained others by amplifying the group message.
Festinger’s group is mostly interested in the behavior of the central group, and this research necessarily focused upon those in the center who were most easily tracked. But I was curious, and re-looked at the study to see what happened overall. Within those with less dramatic pre-event commitment, the report was best able to describe those who stayed. A few in that less-strongly committed group re-upped their adherence. Its a bit hard to track in their report, but to me it is notable that they were able to track down only a few who described their regret at joining. Some, when found, professed adherence to the central belief but made no further active commitments to it. The largest fraction of group disappeared. I’m particularly interested in this ascertainment problem. I suspect that it has importance to the present.
Trump is delighted by his polls, because he chooses to highlight the numbers on his side. Republican approval has “never been higher”. But I’m also aware of the discomfort that some of (my few) Trump-supporting friends now manifest at discussing politics. Those who experience disappointment in a professed belief aren’t inclined to proclaim a change of opinion as much as they are to be uncomfortable with the subject of their (previous) vigorous opinion. They won’t change their opinion so much as they will change the subject. I suspect they deny the present knowing that at a later time that original opinion will no longer be relevant to the questions of the day, and thus need neither be acknowledged nor denied.
Political polling requires voluntary cooperation. I’m guessing that a good number of the original marginal Trump supporters are now disinclined to answer political questions. It’s very difficult to track the rate of refusal to answer – ascertainment bias is the sneakiest confounder of data, as it undermines the denominator, not the numerator.
So in this backwards way, I have to say I’m very pleased to see that Trump’s favorable rating amongst Republicans is stable or increasing. It is most consistent with a dissolution of his previous voters. I suspect we aren’t seeing them in the polls of Republicans now, increasingly won’t see them reflected in the polls of the proximate future, and predict that we won’t see them at the polls in November.