Abstract: Numerous studies on how people reason with statistical data suggest that human judgment often fails to approximate rational probabilistic (Bayesian) norms. Tversky and Kahneman (1983) studied people’s probabilistic inference under uncertainty using the Linda problem and several particular scenarios. According to the empirical results, in some situations when subjects are asked to assign the likelihood of several alternatives, including single and joint events, they tend to rate a probability to a conjunction of two events larger than that they assign to one of the constituent events, which anomalous phenomenon is called “conjunction fallacy”, and such fallacious behavior on conjunctive probability judgment was explained in terms of the “representativeness heuristic”. However, the heuristic has been criticized heavily as being too vague to account for explanations. In this paper, the phenomenon and several explanations on the conjunction fallacy were briefly reviewed, and a new exposition that people apply equate-to-differentiate decision rule (Li, 1994, 2004) to judge in the Linda problem was proposed to interpret this anomalous phenomenon

Key words: subjective probability, reasoning, conjunction fallacy, equate-to-differentiate approach