The probit technique for analysis of subjective assessments of thermal sensation is described. It enables transition temperatures from any selected thermal sensation to the adjacent thermal sensation (e.g. from “neutral” to “warm”) to be identified. A transition temperature is defined as that temperature at which the maximum number of people would change their assessment from one thermal sensation to the next. Thus if a seven-point scale of thermal sensation is used, six transition temperatures are possible. Increments between them will not necessarily be identical, as would be assumed in a linear regression analysis. The method has been applied in three studies: laboratory studies by the Kansas State University, field studies in Port Moresby, and field studies in Melbourne. In the first it is shown that men are more thermally tolerant than women, in the second it is shown that in Port Moresby the preferred temperatures of Melanesians are 2°C higher than those of Caucasians and in the third it is suggested that subjects in Melbourne have a slightly lower preferred temperature than predicted by Fanger (1972).