7 Subjective measures

The subjective sensation of warmth, or thermal comfort, of the subject has traditionally been measured using a seven-point scale. The subject is asked to rate his or her feelings on a descriptive scale such as the ASHRAE or the Bedford scales:

ASHRAE Bedford
Hot +3 Much too warm
Warm +2 Too warm
Slightly warm +1 Comfortably warm
Neutral 0 Comfortable neither warm nor cool
Slightly cool -1 Comfortably cool
Cool -2 Too cool
Cold -3 Much too cool

The resulting number is called the Comfort Vote (C).

There is considerable controversy about these scales. However they have been proven by the very fact of their continued use. The main difference between the two scales is in the inclusion of the concept of comfort in the Bedford scale. Most authorities agree, however, that subjects use the two scales in very much the same way. There is no significant improvement in accuracy by adding more points to the scale.

One criticism of the Comfort Vote is that, by using a descriptive scale we run the danger of overlapping with cultural use of the words. Thus a person living in a cold climate might see `warm' as having a positive connotation ("nice and warm"), whilst the inhabitant of a hot climate would say the same of `cool'. This will tend to skew the use of the scale by a confusion between comfort and hotness. To get round this effect it is advisable to add a Preference Vote P to the Comfort vote, C. The preference vote most commonly used is the three-point one suggested by McIntyre:

I would like to be

1: Warmer 0: No Change -1: Cooler

Using this scale McIntyre (1976) has demonstrateda distinct difference between the preferences of British subjects in winter and American subjects in the summer

One way to get round the problem of a descriptive scale is to use a semantic diffferential scal (say with seven boxes between too hot and too cold). This type of scale would be particularly useful where subjects use a variety of languages. However people interpret the desired or comfortable range differently, some thinking it includes only the central box whilst others think that the middle five represent comfort.

Use a descriptive scale with seven alternatives for people to tick and be sure to make it clear to the subject that you want to know how they feel at this moment. Many tend to interpret the question as “how hot is it (ie the weather or the room) rather than giving information about their own sensation. The preference scale should be presented in a similar manner.

Comfort and preference votes can be collected automatically, possibly using the data logger being used for the physical environment. A number of such automatic set-ups have been used and they have the advantage of allowing the subjects to be prompted, and to cast their votes, fairly impersonally. Otherwise issue the subjects with a pad of questionnaires, or administer the questions verbally. This allows a more complex survey to be undertaken, with more detailed answers on things such as activity and clothing. For longitudinal surveys in particular, keep your questionnaire as simple as possible.

Subjects must be briefed on the aims and methods of the survey. They will, after all, be contributing a considerable amount of effort to your project, and they need to be clear about what is expected of them and what pitfalls to avoid - such as rushing in from another room with a different environment just before recording their comfort vote.

As an addition to the instantaneous comfort votes it would be well worth asking for a more generalised opinion from you subjects as to the satisfaction of their thermal environment. This can be done before and after the survey, and the latter could be specifically aimed at getting a sort of integrated evaluation of the period of the survey. Such a questionnaire might be divided in to a number of semantic differentials with the overall question "compared to usual, did you find the period of the survey:"


Knowledge of the insulation of the clothing worn is not necessary for an estimation of the comfort temperature in a given situation. The clothing is assumed to be a function of the climatic and social milieu of the subjects and one of the factors that make up the desired conditions. However, if the results of your survey are to be used to compare observed comfort vote and predicted PMV, or if some generalisations are to made from them a record of clothing worn will be useful.

If comparison is needed then you will need to use the clothing descriptions used in the development ofthe predictive model your are making a comparison with. A copy of the list of clothing from ISO 7730 is appended. Comprehensive lists of clothing descriptions are available which can be used to estimate clothing insulation. The result is in clo-units (one clo-unit is the insulation of a suit with normal underclothes = 0.155 Km²/W) which is the unit usually used in comfort equations.

Metabolic rate

The measurement of metabolic rate is also not really necessary in field studies, also being a function of the social and climatic milieu and of the task for which a comfort temperature is being found. There is reason to think, however, that the method used to predict metabolic rates in models based heat balance is a source of error in predictions of comfort based on heat balance assumptions. This makes it useful to retain a description of the tasks being performed by your subjects for comparison with comfort predictions.

A standard list of metabolic rates from ISO 7730 is given in Appendix 2. notice that the metabolic rates are assumed to be determined solely by the task being undertaken by the subject. It is also worth remembering that the measurements of metabolic rate upon which these tables are based were done in steady-state conditions, as were the determinations of their affect on comfort in the heat balance model. You will need to determine the best available description for the activity of your subject each time you get a comfort vote. A complication which may arise here is that people vary their activity from time to time, and a particular activity can affect the thermal balance for some time after it has finished. The recommendation in ISO 7730 is to use the average metabolic rate for the activities undertaken in the last hour. This may be difficult for a longitudinal survey.