Why CAL-101 Will Have An Effect On Most Of Us

Матеріал з HistoryPedia
Перейти до: навігація, пошук

Analyzing the contrast between undergraduates and executives (junior and senior combined) separately, revealed a significant difference in the SKAP1 percentage condition (14.6 vs. 26.6%, B = 0.796, SE = 0.310, z = 2.57, p = 0.010), while the difference in the natural frequency condition was negligible and not significant (40.5 vs. 36.2%, B = 0.213, SE = 0.252, z = 1.12, p = 0.398). Finally, analyzing the contrast between the two representation formats revealed a significant difference between the percentage condition and the natural frequency condition within the undergraduate sample (14.6 vs. 40.5%, B = 1.48, SE = 0.28, z = 5.32, p displays, for each strategy separately, the coefficients (B) and the p-values of the five different logistic regressions, each with the main effect of representation format and participant sample, and the interaction between representation format and sample (after controlling for task and order, and with standard errors clustered for each participant). For each of the five cognitive strategies, except for providing the false-alarm rate as response, the number of participants who provided the corresponding numerical estimate significantly differed between the percentage condition and the natural frequency condition. In contrast, for none of the strategies, except for joint occurrence, we observed a significant effect of participant sample, and for none of the strategies, except for Bayesian, the interaction between representation format and participant sample reached significance. Did the Order Matter? No. Across all tasks, participants, and both representation formats, a response has been classified as Bayesian in 30.5% for tasks on the first page of the questionnaire and 28.0% for tasks on the second page (in a logistic regression, the difference was not significant; B = 0.131, SE = 0.115, z = 1.14, p = 0.254; after controlling for representation format, participant sample and task, and with standard errors clustered for each participant). Did Prior Knowledge of Bayes�� Rule Make a Difference? For none of our executive participants, but for 91 of our 259 undergraduate participants (62 bachelor and 29 master students) the booklet contained questions on demographics and on prior knowledge about Bayes�� rule.