Choice blindness is a cognitive phenomenon describing that when people receive false feedback about a choice they just made, they often accept the outcome as their own. Little is known about what predisposes people to correct manipulations they are subjected to in choice blindness studies. In this study, 118 participants answered a political attitude survey and were then asked to explain some of their responses out of which three had been manipulated to indicate an opposite position. Just over half (58.4%) of the manipulations were corrected. We measured extremity, centrality and commitment for each attitude, and one week prior to the experiment we assessed participants’ preference for consistency, need for cognition and political awareness. Only extremity was able to predict correction. The results highlight the elusiveness of choice blindness and speak against dissonance and lack of motivation to engage in cognitively demanding tasks as explanations why the effect occurs.

False feedback on choices has been documented to inducelasting preference change. Here we extend such effects to thepolitical domainand investigatethetemporal persistence ofinduced preferences, as well as, the possible role the length ofconfabulatory justifications may play. We conducted a two-day choice blindness experiment using political statements,with sessions being roughly one week apart. Changes inpolitical preferences remained one week after initialresponses, and were most prominent in participants who wereallowed to confabulate freely. These findings, being the firstto demonstrate lasting preference change using choiceblindness, are discussed in light of constructivist approachesto attitude formation through a process of self-perception.