Analyzing the Impact of Bayesian Inference Analysis of Illusion Types on Using Learning Strategies

The purpose of this study is to use Bayesian inference(BI) to update prior probabilities(PP) of the use of four study strategies in light of the illusion score(IS) and to suggest pedagogical implications of the results. 231 university students were surveyed. After establishing four quartiles based on the IS(expected - actual score), the actual score(AS) of each quartile was calculated to see if there was a difference between the ASs, and the ASs were lower as the higher the IS, but the ASs of the two groups with the lowest ISs were not significantly different. The posterior probabilities(PoP) of the use of four learning strategies were updated after applying the ISs from the survey to the PP of their use, and the results showed that, for close illusion, the use of the ‘high’ useful strategy decreased and the use of the ‘low’ strategy(LS) increased, i.e., students who most accurately judge their competence are using less of the ‘high’ strategy(HS) and more of the LS. The underestimators are made up of relatively high achieving students, who are found to use fewer HS and fewer LS. The overestimators(OS) were relatively poorer students, who used more of the HS and more of the LS, including ‘rereading(R)’, but less of ‘highlighting/underlining(H/U)’ and ‘summarising(S)’. In other words, the high-achievers were using less of the LS, and the near-achievers, who were in the middle of the academic spectrum, were using more of the LS. OS are more likely to use ‘R’ but less likely to use ‘H/U’ and ‘S’. In conclusion, the PoP of learning strategy use, updated by BI after accounting for the degree of illusion (over, near, or under), suggest that students tend to use learning strategies inconsistently with their illusion type.