The Measurement of Subjective Probability

Beliefs come in degrees, and we often represent those degrees with numbers.We might say, for example, that we are 90% confident in the truth of some scientific hypothesis, or only 30% confident in the success of some risky endeavour.But what do these numbers mean? What, in other words, is the underlying psychological reality to which the numbers correspond? And what constitutes a meaningful difference between numerically distinct representations of belief?In this Element, we discuss the main approaches to the measurement of belief.These fall into two broad categories-epistemic and decision-theoretic-with divergent foundations in the theory of measurement.Epistemic approaches explain the measurement of belief by appeal to relations between belief states themselves, whereas decision-theoretic approaches appeal to relations between beliefs and desires in the production of choice and preferences.