著者

若井 大成 岡田 謙介

出版者

日本基礎心理学会

雑誌

基礎心理学研究 (ISSN:02877651)

巻号頁・発行日

pp.40.19, (Released:2022-04-12)

参考文献数

41

---

Among the models of the metacomprehension judgment process, the anchoring and adjustment model by Zhao and Linderholm (2008) is superior considering its generality and explanatory capacity. However, this model is theoretical rather than empirical. Therefore, the following two major problems need to be solved: (1) quantifying the anchoring and adjustment effect and (2) quantifying the influence of the factors that affect the adjustment process. To tackle these problems, we proposed a novel mathematical formulation of the anchoring and adjustment model by considering metacomprehension as a probabilistic distribution. Moreover, using data collected from an online experiment, we estimated the parameters of the proposed model. As a result, it is indicated that (a) the effect of anchoring by self-perception of ability is larger than that of adjustment by experience in reading metacomprehension, and (b) five factors such as familiarity of the contents can explain the adjustment process. Finally, based on the results, we discussed the explanatory capacity of the model, as well as the quantitative effect and influence of the above two points.

著者

塚村 祐希 若井 大成 下條 朝也 植田 一博

出版者

日本認知科学会

雑誌

認知科学 (ISSN:13417924)

巻号頁・発行日

vol.29, no.3, pp.494-508, 2022-09-01 (Released:2022-09-15)

参考文献数

30

---

Latent scope bias is a bias that arises when humans estimate how probable a causal explanation is. This bias is a tendency to underestimate the probability of explanations with latent scope, the set of unobserved events that may or may not be occurring. Previous studies proposed the “inferred evidence” account, in which the bias occurs because we underestimate the probability that the unobserved event is occurring and reason based on this probability using the Bayesian rule. However, no studies have examined whether humans estimate the probability of explanations based on the Bayesian rule. Therefore, the present study examined how humans estimate the probability of explanations under uncertainty using Bayesian cognitive modeling. Specifically, participants read two explanations with different latent scopes and responded to one of them with a probability of 0% to 100%. The results obtained indicate the following two points: First, humans estimate the probability of explanations based on the Bayesian rule, which supports the inferred evidence account. Second, there are individual differences in the occurrence of latent scope bias.