1、NA-04-2-4 Do People Like to Feel “Neutral”? Response to the ASHRAE Scale of Subjective Warmth in Relation to Thermal Preference, Indoor and Outdoor Temperature Michael A. Humphreys J. Fergus Nicol ABSTRACT This analysis explores a discrepancy between “neutral on the ASHRAE scale and refer no change”
2、 on a thermal preference scale. The data employed are from numerous field studies drawn from two large databases. Multivariate analysis showed thai the ASHRAE scale vote depended not only on the recognized thermal variables but also on the outdoor temper- ature, while the thermal preference vote was
3、 relatively little influenced by outdoor temperature. Of respondents who desired “no change” in their thermal state, many had a ther- malsensation other than neutral. Ifthe indoor temperature was warm, their preferred sensation tended to be slightly warmer than “neutral, while ifthe outdoor temperat
4、ure was high, the preferredsensation tended to be slightly cooler than “neutral. Each eflect was about one-third of a unit of the ASHRAE scale. Some implications for research and practice are discussed. INTRODUCTION Scales of Subjective Warmth Early researchers in thermal comfort used a variety of s
5、cales for recording warmth, but, by the 1970s, two scales had become dominant, and both of them had seven categories. The ASHRAE scale was developed during research for ASHVE (American Society of Heating and Ventilating Engi- neers; later became ASHRAE). The central category was orig- inally labeled
6、 “comfortable,” but this was later changed to “neutral,” and so, in its current form, the scale has no overt reference to comfort. Nevertheless, it is usually assumed that the three central categories indicate the zone of thermal comfort, an assumption incorporated in the construction of the PMV and
7、 PPD indices (Fanger 1970). The Bedford scale was normally used in British and Commonwealth countries. It arose from Bedfords magisterial field study of the winter warmth of workers in light industry (Bedford 1936). The scale (Table 2) combines warmth and comfort. Each ofthe three central categories
8、 carries a label that indicates comfort, while the others indicate discomfort. McIntyre (1 978) compared the properties of various seven-point scales of warmth, using data from numerous experiments conducted in a climate-controlled room. He Table 1. The Categories of the ASHRAE Scale 0 Hot O Warm 0
9、Slightly warm 0 Neutral 0 Slightly cool o Cool 0 Cold Table 2. The Categories of the Bedford Scale 0 Much too warm 0 Too warm 0 Comfortably warm 0 Comfortable 0 Comfortably cool 0 Too cool 0 Much too cool Michael A. Humphreys is a research professor at the Oxford Centre for Sustainable Development,
10、Oxford Brookes University, U.K., and a research fellow at the Centre for Christianity and Culture, Regents Park College, University of Oxford. J. Fergus Nicol is a research professor both at the Oxford Centre for Sustainable Development, Oxford Brookes University, U.K., and at London Metropolitan Un
11、iversity. 02004 ASHRAE. 569 concluded that the properties of the ASHRAE scale and the Bedford scale were similar. De Dear (e.g., de Dear 1985) compared the two scales during fieldwork in Australia and found that in practice the scales were interchangeable. These results strengthened the presumption
12、that the three centrai categories of the ASHRAE scale could be taken to indicate a state of thermal comfort, for they corresponded to the three “comfortable” categories of the Bedford scale. Scales of Thermal Preference A scale of thermal preference indicates whether people would prefer a warmer or
13、a cooler condition, and so it may be used to locate more precisely where within in the comfort zone the optimum point might be. Fox used the Bedford scale in a study of the elderly in the U.K. winter and supplemented it with a question to probe their thermal preferences (Fox et al. 1973). The scale
14、had five categories. Fox noticed that some people who said they were “comfortable” nevertheless would have preferred a warmer condition; the central category of the Bedford scale was not necessarily optimal. Humphreys (1975) compared the findings of numerous field studies of thermal comfort from var
15、ious climates. He suggested that discrepancies between the temperature the vari- ous populations found comfortable and the expectations from thermal physiology could be reconciled to some extent if people in hot climates preferred to feel slightly cooler than neutral, while people in cold climates p
16、referred to feel slightly warmer. McIntyre (1980) tested this suggestion and found that a sample of British office workers who voted “neutral” in winter would, on the whole, have preferred a condition slightly warmer than neutral, while a sample of Americans in summer would have preferred a conditio
17、n slightly cooler. McIntyre used a three-point scale of thermal preference (Table 3). Scales of thermal preference have not settled down to an agreed number of categories. Most researchers use the three- point McIntyre scale, but Nicol, following Fox, prefers a five- point preference scale (e.g., Ni
18、col et al. 1994), while Oseland (1 997) used a seven-point scale. Recent research has confirmed that a preferred tempera- ture obtained from the preference scale sometimes differs from a corresponding neutral temperature obtained from a seven-point scale of warmth. Nicol et al. (1994) reported diffe
19、rences of a degree centigrade or so for summer data from Pakistan. He had used a seven-point semantic differential scale of subjective warmth, with the central category labeled “comfort.” Feriadi and Wong (2002), using the ASHRAE scale, reported a larger difference, some 3”C, for people in their own
20、 homes in Indonesia. De Dear (de Dear et al. 1997), Table 3. Mclntyres Scale of Thermal Preference I would like to be: OWarmer UNO change OCooier comparing numerous field studies of thermal comfort from various countries, found that, although the differences were usually small, there could be a diff
21、erence of up to 3C between the neutral and the preferred temperatures in a particular set of data. The systematic dependence ofthese discrepancies on the climate was slight. For people in air-conditioned buildings, the effect of climate was statistically significant but very small, while for people
22、in naturally ventilated buildings, it was perhaps larger, but more scattered and not statistically signif- icant. The difference did, however, vary substantially with the mean indoor temperature for people in air-conditioned build- ings, but again it was not significant for those in naturally ventil
23、ated buildings. De Dears analysis pooled studies that used the ASHRAE scale, the Bedford scale, and the seven-point semantic differ- ential scale. Although their broad similarity has indeed been established, these scales differ in the wording of their central category (“neutral,” “comfortable,” and
24、“comfort,” respec- tively), and each may therefore bear a slightly different rela- tion to the preference scale. It could be that a clearer pattern would emerge if the scales were examined separately. This paper aims to quanti the relation between climate, indoor temperature, and the preferred point
25、 on the ASHRAE scale. There are insufficient data for a parallel treatment of the Bedford scale or of the semantic differential scale. Why the Preferred Point on the ASHRAE Scale Might Vary It seems natural to expect that people would prefer to feel warm in a cold climate and cool in a hot climate,
26、and there is anecdotal support for this expectation. Considering possible reasons for this leads to a speculative model for the relation between the preferred sensation of warmth, the indoor temper- ature, and the climate. In an equatorial climate, people outdoors often feel uncomfortably warm and s
27、eek cooler places for comfort. It could be, then, that they develop an association between words for comfort and for coolness. Conversely, in a cold climate, people outdoors often feel uncomfortably cold and seek warmer places for comfort. They would be likely to develop an association between words
28、 for comfort and for warmth. Thus, in the U.K. in winter, the phrase “comfortably warm” seems natural, and it is usual to ask a guest if he or she is warm enough. It would be eccentric to inquire if they were cool enough. Anecdotally, the reverse is true in hot climates. If this understanding is cor
29、rect, it implies that, while the physiolog- ical state associated with comfort would not necessarily depend on the climate, the description applied to it might well do so. In regions having cold winters and hot summers, the preferred point on the scale (if McIntyres result is capable of generalizati
30、on) undergoes a seasonal shift. This argues a seasonally sliding association of comfort with either warmth or coolness, rather than a fixed effect of a localitys climate on the use of language. 570 ASHRAE Transactions: Symposia If this supposition is correct, a similar linguistic response would be e
31、xpected to occur if the indoor conditions made the occupants feel persistently too cold or too hot. So, for exam- ple, ifthe ofice were unexpectedly cold on a Monday moming and the occupants became progressively chilled, they would perhaps associate comfort with warmth, while the opposite could be t
32、rue if the building were unusually hot on a sunny day in spring. This could lead to a negative association between room temperature and the desired point on the ASHRAE scale. There is yet another consideration, which would tend in the opposite direction. It is probable that, for reasons other than c
33、limate, people sometimes prefer to feel warm or cool rather than “neutral.” So, if the occupants of a building could control their thermal environment, it would be probable that if the room was warm, it was because people were preferring it warm at that time. So within the zones of comfort, apositiv
34、e association between room temperature and the desired point on the ASHRAE scale might be expected. We might thus expect a rather complicated yet explicable variation of the preferred point on the ASHRAE scale in rela- tion to both the indoor and the outdoor temperatures. THE DATA Two large database
35、s of field studies of thermal comfort were available to us and enabled a test to be made of the suggested model. It is not necessary here to give comprehen- sive descriptions of these extensive databases; readers needing fuller information will find it in the cited references. We confine ourselves t
36、o a brief description of the principal features. The SCATs Data These data come from a Europe-wide survey of the office environment. During the SCATs project (smart controls and thermal comfort), measurements of the indoor climate were made at selected work areas in 26 ofice buildings in five Euro-
37、pean countries (France, Greece, Portugal, Sweden, and the United Kingdom). The participating people were visited monthly, and more than 4500 “desk visits” were made between June 1998 and October 1999. A mobile automated data-acquisition system recorded air temperature, globe temperature, relative hu
38、midity, air speed, the concentration of carbon dioxide in the air, the horizontal illuminance, and the sound pressure level at the work stations. Outdoor tempera- tures were obtained from nearby meteorological stations. The respondents were interviewed and asked to evaluate the envi- ronment at thei
39、r workstation. The questionnaire included rating scales for thermal sensation (ASHRAE scale), thermal preference (a five-point scale), as well as rating scales for other aspects of the environment: air movement, humidity, light, sound, and air quality. The scales, in the language of each country, we
40、re administered using a flip-chart. The duration of the project, the use of identical methods across five countries, and the seasonal variation of their climates provided a large database of unique value. A full description of the data collec- tion, including details of the instrumentation, may be f
41、ound in Nicol and McCartney (2001). The de Dear Data De Dear et al. (1 997) assembled the raw data from a large number of field studies to form a database from which to develop for ASHRAE a quantitative adaptive model of ther- mal comfort and preference. (That is to say, to offer empirical evidence
42、for suitable temperatures in buildings in relation to the climate.) Some 15 independent research teams contributed their data, which had been obtained in various countries (Australia, Canada, Greece, Indonesia, Pakistan, Singapore, Thailand, U.K., and U.S.) and at various seasons. The data described
43、 the environments in some 160 buildings. Nearly all were offices, but a few were light industrial buildings, and a few studies included some observations from the respondents dwellings. Three types of building were distinguished: centrally air-conditioned, mixed mode, and naturally venti- lated. The
44、 data comprise over 20,000 individual comfort votes on the ASHRAE scale or on some other seven-point scale. Not all research teams had obtained thermal preference votes, and those who did sometimes differed in the scales they used. There were corresponding measurements of the thermal envi- ronment (
45、usually air temperature, globe temperature, humid- ity, air speed), together with records of clothing and activity. A full description of the data, including details of the instru- mentation, may be found in de Dear et al. (1997). The wide range of climates, the large number of buildings, and the di
46、verse cultural backgrounds combined to make these data useful for this exploration of the relation between thermal sensation and thermal preference. They cover a wider range of climate than do the SCATs data, but the various field studies used were not so uniform in their methodology. THE STATISTICA
47、L ANALYSIS The de Dear data are published as a set of files in Microsofi Excel format. We produced from them a single file in the format of the SPSS (statistical package for the social sciences) statistical package, at the same time incorporating a number of minor corrections (Humphreys and Nicol 20
48、02). In some of the Excel files, the McIntyre scales had been coded in reverse, and this we corrected. The SCATs data also were put into SPSS format, and SPSS (version 10) was used for the following analysis. Cross-Tabulations of the ASHRAE Scale and Preference Scale First we looked at the relation
49、between the votes cast on the two types of scale without reference to indoor temperature or to the climate. We selected those occasions where the respondents had provided an assessment on the ASHRAE scale, together with a corresponding assessment on the pref- erence scale. (This excluded the data from Pakistan because a seven-point semantic differential scale had been used, where the center was labeled “comfort” micol et al. 19941; the data from Greece, because they had used the Bedford scale ASHRAE Transactions: Symposia 571 Table 4. Cross-Tabulation of the ASHRAE and Mc
