Abstract
The experimental basis of the adaptive approach to thermal comfort is the field study. Since the publication of Bedford's study in 1936, many researchers have collected data on people's thermal comfort in everyday conditions. The principal assemblages of data are those of Humphreys (1975; 1978; 1981), de Dear et al (1997), de Dear (1998), de Dear and Brager (1998) and McCartney and Nicol (2001). In addition there exist numerous studies not included in any assemblage. The data come from a variety of climates and countries, from people in buildings that are heated or cooled mechanically, and from buildings operating without either heating or cooling. This paper compares the more recent data with the early patterns revealed by Humphreys' 1975–1981 analyses of the worldwide data. The comparison is made for indices of thermal comfort, mean warmth response, temperatures found to be comfortable, sensitivity to temperature change, interpersonal variation, distribution of comfort with temperature, and the influence of outdoor temperature in the presence or absence of heating and cooling. The empirical findings are compared with predictions from the predicted mean vote (PMV) model of comfort. The role of the adaptive approach in the formation of standards and guidelines is discussed.