We will exclude participants with incomplete data (due to technical difficulties etc.), participants who will answer that we should not use their answers (because they did not understand instructions or were not focused during the session), participants who answered that they were aware about the design of the study, and participants who will both notice the manipulation of the answers and know what choice blindness means. Our main variable of interest is the number of corrected manipulated trials, that is the number of scenarios where participant’s initial answer is reversed in the second presentation, but he or she chooses to change it. Data analysis of the primary hypothesis will be done using multilevel logistic regression (Gelman & Hill, 2006). The change of an answer by a participant on the second reading will serve as a dependent variable. We will treat situations and participants as random factors and include effects of group and manipulation of an answer as predictors. We will use both varying intercept and varying slope for the effect of the manipulation for situations and varying intercept only for participants. Depending on the overall number of changed answers, we may use the question asking whether they would change the original answer for exploratory analysis in which we would examine whether people are more likely to correct an answer they say was made intuitively or deliberatively. <br> References Gelman, A., & Hill, J. (2006). *Data analysis using regression and multilevel/hierarchical models.* Cambridge: Cambridge University Press. <br> The analysis script used for obtaining the reported results is included in Files (data_analysis.R). For the primary hypothesis, it used the function that was preregistered before data collection.