Decisions based on verbal probabilities: Decision bias or decision by belief sampling?

We examined decisions based on verbal probability phrases, such as "small chance," "likely," or "doubtful" (we call these phrases verbal probabilities). Verbal probabilities have communicative functions called directionality and can be categorized into positive (e.g., "likely" or "probable") or negative (e.g., “unlikely,” “doubtful”) phrases in terms of their directionality. Previous studies have shown that the directionality of phrases affects decisions. Although such decisions seem biased, we argue that they are not. We hypothesize that since a speaker has the option to choose the directionality used during communication, the selected directionality becomes relevant information to a decision maker, and is taken into account in making decisions. We modeled these processes using the Decision by Belief Sampling (DbBS) model. We found that the observed data could be well explained by our hypothesis, and that the DbBS model could be one of the best potential models for decisions based on verbal probability information.