Constructing an alert system of violations of the normative axiom of probability using a text-corpus

本研究旨在通过语料库数据分析，从确认度（Crupi, Tentori, & Gonzalez, 2007）的角度解释连接谬误（Tversky & Kahneman, 1983）。为此，我们计算了来自英国国家语料库的确认度指数，并将其与先前研究的数据进行拟合（Shafir et al., 1990）。结果表明，确认度这一主要指标（Crupi et al., 2007）可以显著预测连接谬误，表明连接谬误与确认度之间存在关系，以及语料库数据对解释判断偏差的重要性。

This study aims to explain the conjunction fallacy (Tversky & Kahneman, 1983) in terms of degree of confirmation (Crupi, Tentori, & Gonzalez, 2007) by employing corpus data analysis. To accomplish this, we calculated indexes of the degrees of confirmation from the British National Corpus and fitted them to data of the previous study (Shafir et al., 1990). The results show that a major index of the degree of confirmation (Crupi et al., 2007) can significantly predict the conjunction fallacy, indicating a relationship between the conjunction fallacy and degree of confirmation as well as the importance of corpus data to explain biases in judgment.