[4D2-GS-12] Human interface, education aid: Fundamental human interface

[4D2-GS-12-03] Analysis of anchoring effects in rate estimation using a Bayesian update model

〇Tomoaki Hamada¹, Takashi Takekawa¹ (1. Kogakuin University of Technology & Engineering)

Keywords:anchoring effect, cognitive bias, Bayesian update, Beta distribution

There are phenomena in which later judgment / estimation is affected by the numbers presented in advance. This phenomenon is called the anchoring effect, and much research has been done. As an example, Ozawa and Takekawa quantitatively described the anchoring effect for subjects with a small amount of knowledge in 2018 using the log-t distribution. However, the log-t distribution derived by Ozawa et al. Cannot be used for ratio estimation where the estimated value is in the range of 0-1. The purpose of this study was to verify the anchoring effect in rate estimation and to construct a model. In general, a beta distribution is often used as a probability distribution for proportions. Therefore, fitting by beta distribution was tried and it was confirmed that fitting was possible. In addition, it was confirmed that the qualitative properties of the anchoring effect coincided with the ratio estimation.