Edited by: Isabella Pasqualini, Ecole Polytechnique Fédérale de Lausanne, Switzerland
Reviewed by: Timo Partonen, National Institute for Health and Welfare, Finland; Robert J. Lowe, University of Gothenburg & University of Skövde, Sweden
*Correspondence: Trevor P. Keeling
This article was submitted to Cognitive Science, a section of the journal Frontiers in Psychology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
The physical environment leads to a thermal sensation that is perceived and appraised by occupants. The present study focuses on the relationship between sensation and evaluation. We asked 166 people to recall a thermal event from their recent past. They were then asked how they evaluated this experience in terms of 10 different emotions (frustrated, resigned, dislike, indifferent, angry, anxious, liking, joyful, regretful, proud). We tested whether four psychological factors (appraisal dimensions) could be used to predict the ensuing emotions, as well as comfort, acceptability, and sensation. The four dimensions were: the Conduciveness of the event, who/what caused the event (Causality), who had control (Agency), and whether the event was expected (Expectations). These dimensions, except for Expectations, were good predictors of the reported emotions. Expectations, however, predicted the reported thermal sensation, its acceptability, and ensuing comfort. The more expected an event was, the more uncomfortable a person felt, and the less likely they reported a neutral thermal sensation. Together, these results support an embodied view of how subjective appraisals affect thermal experience. Overall, we show that appraisal dimensions mediate occupants' evaluation of their thermal sensation, which suggests an additional method for understanding psychological adaption.
Treating thermal comfort as a problem of
For the purpose of the present investigation, thermal experience is broken down into three components. First, physical environments, such as air temperature, air movement, etc., constitute the medium within which occupants operate. Secondly, thermal sensation is the interface between the occupant and the environment, which is predominately described using the ASHRAE thermal sensation scale, which runs from cold, through cool, neutral, warm, to hot (ASHRAE,
We look at the psychological factors that shape how thermal sensations are perceived and evaluated, by grounding our investigation in the field of emotion psychology. Particularly, we are interested in the way that four psychological factors (“appraisal dimensions”) may shape the criteria of acceptability, comfort, thermal sensation, and the ensuing emotional experience, which result from the exposure to a particular thermal environment, or thermal event. We aim to render explicit the relationship between the occupant's psychology and their thermal experience, and hope to inform building design practices by situating occupants at the center of the space they occupy.
Models of thermal comfort attempt to predict evaluation or sensation dependent upon the physical environment. For instance, both the
Physiological models of thermal experience describe the energy flows within the body. They split the body into several layered sections, each with different thermal properties, which are used to predict the energy balance and temperature throughout the body (Fiala et al.,
Alliesthesia provides one explanation why a neutral thermal sensation, or any other single thermal sensation, will not always lead to the same evaluation. As such, it provides a theoretical approach to understanding the relationship between sensation and evaluation. It suggests that when a person is overheated they will find a cold sensation pleasant, whilst when a person is overcooled they will find a hot sensation pleasant (Cabanac,
A final perspective pertains to the psychological effects that certain environments may have on individuals, yielding particular states (Farshchi and Fisher,
In the work presented here, we are interested in the overall experience of thermal comfort. Emotion psychology bridges the psychological antecedents of an event to the unfolding of psychological and physiological responses to that event. In the field of building design, adaptive comfort theory is the theoretical tradition that provides the most insight into psychological factors, and we therefore seek to enhance this understanding of occupants' experience with insight from psychology.
A fundamental question in the field of emotion psychology concerns the fact that two people may be presented with the same situation and yet have different subjective experiences. A growing body of results suggests that the appraisal of the situation mediates the sensation and the ensuing evaluative response (Arnold,
The example above exposes the relationship between appraisals and the evaluation that follows. In this example, a single appraisal dimension of conduciveness is used to evaluate a thermal situation. One person appraises the situation as conducive to their goal and experiences positive emotions, while the other appraises the same environment as obstructive and thus experiences negative emotions. These appraisal processes occur at a subconscious level, which influences the overall experience. Appraisal theorists attempt to characterize the quality of relevant appraisal dimensions and make predictions for the ensuing emotions (Arnold,
We propose to use appraisals as proxies for understanding how participants' past experience will affect their conceptualization of a given environmental stimulus or scenario. This experience is reduced to a limited number of fixed appraisals, and one of the simplest appraisals is whether a stimulus is consistent with a person's motives and desires or not; if it is, then the resulting emotion is likely to be positive, if not, then the emotion is likely to be negative.
Further appraisal dimensions can be used to predict which positive or negative emotions will be experienced. For instance, another common appraisal is what or who is responsible for the cause of the experience. If a person appraises that they are themselves responsible (for a negative outcome), the theory predicts they will experience regret. If someone else or unavoidable circumstances (e.g., the weather) are believed to be the cause, then anger, frustration, or resignation would be experienced. Together, these appraisal dimensions can help predict specific emotions (Scherer,
The value of appraisal theory for the field of building design is that it provides a framework to understand how people's conceptualization of a situation affects their experience. This sheds light on the mapping between sensation and evaluation. Four appraisals, which are implicit in adaptive comfort theory and explicit in appraisal theory, may be useful to our aim. These are goal conduciveness, causality for the situation, perceived control, and expectation (Smith and Ellsworth,
We focus our investigation on four psychological factors that are present in both Adaptive Comfort theory and the Appraisal theory of emotions, either implicitly or explicitly, and on related effects over thermal experience. The purpose of this comparison is to study the predictions from both sets of theories and highlight aspects that could be operationalized in design practice. The psychological factors of interest here relate to the information processing units that may serve in the evaluation of a given thermal environment by a given occupant. Although it is believed that such evaluations may be performed over continuous sets of criteria, we restrict our investigation to discrete, extreme situations to formulate working hypotheses.
Conduciveness—relates to the extent to which a given thermal event will serve or obstruct an occupant's goal. High conduciveness implies that the event supports the occupant's present goals, whereas low conduciveness implies that the event does not support or even hinders their goals in some ways.
Causality—relates to the extent to which a given thermal event has been caused by either unavoidable circumstances, the occupant themselves, or other occupants. By unavoidable circumstances, we mean natural conditions, e.g., a sunny day, or changes in the environment that affect occupants, e.g., a malfunctioning radiator. An example of a situation caused by the occupant themselves or others may be the opening of a window, or a voluntary change in the setting of the thermostat, in line with or against shared space policies (Leaman and Bordass,
Perceived control—relates to the extent to which the occupant perceives they have control over their environment (Brager and de Dear,
Expectations—relates to the extent to which the occupant was expecting a given thermal event (Brager and de Dear,
We thus formulate the following predictions, which drove the elaboration of our questionnaires and the ensuing analyses of the data. In our interpretation of the results, we compare the predictions from both sets of theories—see Table
(1) Conduciveness | High → Comfortable | High → Neutral | High → Joy, pleasure |
Low → Uncomfortable | Low → Cold, hot | Low → Displeasure | |
(2) Causality | Circumstances → Comfortable | Circumstances → Neutral | Circumstances → Resignation, anxiety |
Others/Self → Uncomfortable | Others/Self → Too hot, too cold | Others/Self → Dislike, anger | |
(3) Perceived control | High → Comfortable | High → Neutral | High → Anger, anxiety |
Low → Uncomfortable | Low → Too hot, too cold | Low → Resignation, frustration | |
(4) Expectations | High → Comfortable | High → Neutral | High → Resignation |
Low → Uncomfortable | Low → Too hot, too cold | Low → Frustration |
As part of a wider field study focusing on evaluating the relationship between environmental factors and psychological experience, occupants of seven office buildings responded to our survey (
166 | 105 | 57 | 4 | 84 | 77 | 5 |
A | 9 (18) | Design | Open plan | Shallow | MM |
B | 9 (69) | Academic | Open/Cell | Shallow | NV |
C | 46 (17) | Academic | Open/Cell | Shallow | NV |
D | 29 (15) | Academic | Open/Cell | Shallow | MM |
E | 9 (18) | Design | Open plan | Shallow | NV |
F | 25 (2) | Charity | Open plan | Deep | AC |
G | 39 (26) | Design | Open plan | Shallow | NV |
Tapping into the subjective experience of an individual is a major challenge, because the mere attempt to ask a question is likely to disrupt the unfolding experience altogether. To eliminate this disruption, we chose to use a recall survey, in which participants were asked to recall a salient event in their recent past and to answer a number of questions about that event. This also allows us to access a much greater range of experiences than if it was necessary to be present at the time of the event, measuring the thermal environment as the experience unfolded. The reliance solely on user reported data, with little or no measurement of the physical nature of the stimuli, is common in psychology (Fontaine et al.,
The recall survey started with a prompt for the participants to recall an event in detail. To do this they were asked to:
“
After this, a number of questions were asked about each of the four appraisal dimensions. Details of the questions and how they were combined can be found in the Appendix in Supplementary Material. These were used to understand:
Whether the participant felt the event was conducive to them (appraisal 1);
Who or what they thought caused the event (appraisal 2);
Who or what they thought controlled conditions in their office (appraisal 3);
How much they had expected the event to happen (appraisal 4).
To finish the survey, there was an open response to describe feelings and a closed list of emotions to choose from: frustrated, resigned, dislike, indifferent, angry, anxious, liking, joyful, regretful, proud, or none of these. Then three questions were asked about the participant's thermal experience, using a thermal sensation scale, a comfort scale and an acceptability scale.
We examined whether appraisals have an effect upon emotions, acceptability, comfort, and sensation. The model used compares the likelihood of a particular evaluation, dependent upon the score on an appraisal dimension. The most appropriate statistical model for this is a logistic regression model. This allows prediction of the presence or absence of a given factor (a set of emotions or acceptance) dependent upon an ordered factor (the appraisal dimension). An extension to this model is the ordinal logistic model, which predicts the likelihood of achieving a given level of comfort or sensation depending on an appraisal dimension.
Equation (1) shows the logistic regression model. The model comprises a linear function and a link function. In the same way as standard linear models, the coefficients are derived so as to maximize the fit of the model. The link function
Participants were asked to report their thermal experience during the period that they recalled. Generally, they recalled periods of time when they were experiencing extreme thermal sensations, either too hot or too cold (Table
Cold | 30 |
Cool | 6 |
Slightly cool | 1 |
Neutral | 8 |
Slightly warm | 6 |
Warm | 30 |
Hot | 84 |
Undisclosed | 1 |
Very uncomfortable | 42 |
Uncomfortable | 84 |
Slightly uncomfortable | 38 |
Comfortable | 1 |
Undisclosed | 1 |
Not acceptable | 129 |
Acceptable | 33 |
Undisclosed | 4 |
Participants were asked to choose one of several emotions that best matched their feelings from a closed list. No one reported a positive emotion or an emotion associated with personal responsibility, i.e., regret (Table
Frustrated | 74 |
Resigned | 30 |
Dislike | 20 |
None of these | 16 |
Indifferent | 10 |
Angry | 8 |
Anxious | 8 |
Liking | 0 |
Joyful | 0 |
Regretful | 0 |
Proud | 0 |
Generally, participants reported that the event was unpleasant and worsened their ability to work. We also asked who they thought was responsible for the events leading up to their emotional experience (Figure
The absence of positive emotions and the absence of positive appraisals of conduciveness is in accordance with appraisal theory. However, the lack of positive emotions also means it is difficult to build a comprehensive statistical model for validation. For the remaining three appraisals, the emotions reported were partitioned into two groups according to the relevant hypothesis, i.e., for Causality, one group was aligned with the appraisal of caused by another (dislike and angry) and the other with appraisal of caused by circumstance (frustrated, resigned, indifferent, anxious). Figure
We tested several link functions to model these data, and report statistical tests of the best model in Table
Responsibility | Poisson | 7.1 ( |
−1.1 ( |
−0.28 ( |
|
Control | Cauchit | 3.8 ( |
−2.7 ( |
−0.65 ( |
|
Expectation | Cauchit | 2.70 ( |
−0.15 ( |
0.21 ( |
Figure
Responsibility | Cauchit | 0.82 ( |
−2.2 ( |
0.24 ( |
|
Perceived control | Probit | 2.46 ( |
−1.3 ( |
0.16 ( |
|
Expectation | Probit | 3.48 ( |
−0.27 ( |
−0.10 ( |
Responsibility | 0.18 | 0.10 | 1.76 | 0.08 |
Perceived control | –0.20 | 0.13 | –1.55 | 0.12 |
Expectation | –0.16 | 0.07 | –2.26 | 0.02 |
Figure
Table
Responsibility | 0.04 | 0.12 | 0.30 | 0.76 |
Perceived control | 0.07 | 0.15 | 0.51 | 0.61 |
Expectation | 0.24 | 0.87 | 2.75 | 0.006 |
Taken together, our results show that appraisal processes are important for shaping evaluation of the thermal environment. This approach complements adaptive comfort theory of thermal experience in building design. This work supports the notion that thermal experience is rich and complex, and requires understanding of how people conceptualize their thermal environment (Heschong,
The appraisals of Causality and Control were less useful for predicting the traditional thermal comfort evaluations of acceptability and comfort. This contradicts the extensive literature on perceived control and thermal comfort (Brager and de Dear,
The appraisal of Expectations was successful in predicting comfort evaluation. However, the correlation was opposite to that expected. The more an event was predictable or expected the more uncomfortable it was. Thermal comfort theory would predict that occupants acclimatize to events over time (Brager and de Dear,
When we asked about Expectations, participants may have focused on recurrent salient situations, whereas the classic expectation of thermal comfort refers to repeated and continuous exposure to a ubiquitous climatic experience. Given this observation, it appears that our results draw attention to a different type of expectation effect. Namely that when problematic conditions are recurrent, they become less and less acceptable.
The work on psychological adaption and embodied cognition hints at two different mechanisms through which psychological factors could affect thermal experience. The first mechanism suggests that psychological factors change the mapping between thermal sensation and thermal evaluation. These theories suggest that the benefit of personal control is that it reduces stress from mildly unfavorable conditions and effective control provides pleasure (Hellwig,
Our results support theories of embodiment because where appraisals have an effect on comfort, they also have an effect on reported sensation. However, this can only be taken as weak support for embodiment because our field study did not assess specific thermal environments. The inclusion of synchronous temperature measurements would provide conclusive evidence that the appraisal caused a sensation change, as opposed to thermal sensations causing both comfort and appraisal.
The lack of positive emotions supports work that suggests that temperature is a hygiene or basic factor responsible only for dissatisfaction (Herzberg,
To improve the method and repeatability, the survey could also be made easier to analyze. We focused here on the comparison of sets of theories and assessment of the best explanatory models. Further studies may choose to simplify the design by focusing on particular aspects. First, appraisal dimension could be specified to ease coding. The current system of combining many ordinal responses is convoluted and builds in uncertainty, which was reflected in our analyses. Second, continuous response for variables could be used. This would mean that analysis could be done with genuine ratio scale numbers rather than an ordinal scale that was transformed into a ratio scale.
Appraisal theory provides a simplified way to encapsulate people's thoughts about their thermal experience. These thoughts cover not only a person's core temperature and peripheral thermal stimulus but also their past experiences and future desires. The theory does not try to predict why people make certain appraisals but it identifies which appraisals are key. Overall, our results show that it is the combination of these appraisals that shapes a thermal experience. Multidimensional appraisals require multidimensional evaluations, and in this case we have successfully used ten emotions to describe thermal experience.
Our analysis suggests a new aspect to how expectation affects psychological adaption. We observe that recurrent problems (those that happened often and were predictable) resulted in greatest discomfort. People did not appear to adapt to them. This suggests an alternative way to conceptualize expectation.
With further modifications, the survey developed here could be used as a diagnostic tool where discomfort and dissatisfaction are caused because of psychological factors (as opposed to poor thermal conditions). From this it may be possible to design a program of measures that tackle those psychological causes. This would be in contrast to current industry approaches that focus on costly technical fixes and chase ever more control over the physical environment.
Interpretation, meaning, and other psychological approaches have been shown to play a part in subjective experience across a range of indoor environmental quality indices (Kwon et al.,
The study is based on TK's EngD research work. TK was responsible for the research design, data collection, and analyses. DC and ER provided invaluable guidance and supervision respectively, throughout the research. All authors contributed to the production of this manuscript.
This work would not have been possible without the Institute of Technology for a Sustainable Built Environment (TSBE), Reading University, the Engineering and Physical Science Research Centre (EPSRC) Doctoral Training Scheme (grant number EP/G037787/1) and TK received additional funding from BuroHappold Engineering for the duration of his engineering doctorate.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
A version of this manuscript has been submitted to the 9th Windsor Conference.
The Supplementary Material for this article can be found online at: