Tobacco Documents Online

I'rarocidings of Indoor ;1ir '9,1. Vcil. I 785 A Cl1ARAC1'I;RIZA'TInN nF MI;TIInI)nLnG1l;S FOR ASSI;SSING HUMAN RI;SI'ONSI:S 'I'n 1'IIN INI)nOR ENVIRnNMI?N'C Nisha P. Sensharrna, Patricia K. Edwards; James E.Woods, Julie Seelen Indoor Environment Program, College of Architecture and Urban Studies, Virginia Tech, U.S.A. AI1S'fRA(:T Divergent results reported in indoor environmental studies are exacerbated by the variety of methodologies used for assessing human responses. This paper identifies and characterizes several extraneous factors and methodological effects that may account for much of this divergence. First, huntan responses are classified into a typology consisting of four cloinains: environmental-perceptual, personal-perceptual, environmental-affective, and personal-affective. Second, extraneous factors associated with each domain are arranged in a hierarchy. Third, the methods most frequently reported in the literature are organized into three types of methodologies and characterized in terms of methodological effects. A process of systematically matching human response domains with extraneous factors and available methodologies with methodological effects is then proposed to reduce false positive or false negative errors in the diagnoses of sick or healthy buildings. IN"CRUI)t1(:'I'iON Measures of human responses are especially important in detecting problems such as Sick Building Syndrome (S[3S) symptoms which are associated with combinations of low levels of environmental stressors that may not be amenable to physical measurement. Stressors in the indoor environment inciucle physical environmental factors such as thermal, air quarty, acoustics and lighting. Several methodologies are currently used to assess the response of building occupants to these stressors. They often yield varied results. This paper identifies and characterizes two types of confounding factors, narnely, extraneous factors and methodological effects that can affect the internal validity of human response methodologies. The analysis is conducted in three stages. First, human responses are classified into four domains. Second, each of these doniains is associated with a set of extraneous factors (i.e., physical factors other than those under study, that explain in whole or part the observed human responses). Third, three types of methodologies are identified and characterized in terms of potential methodological effects. This paper is based on a litcrature review of the results of past empirical research and on critiques of standard procedures used in incloor environmental evaluations. It concludes with a proposed process of systematically matching human response domains with extraneous factors and available methodologies with methodological effects to improve the conu il of confounding variables that are present during building diagnostic procedures. "fY1'OLO(;Y OF IIUDiAN RI;SI'ONSE The most freyuently used measures of human responses identified from the literature

786. t`rocccdingt (if lndoor Air'9t, Vol. I include sensory responses, physiological symp(orns, comfort, preference, and acceptability. One of the major problems when comparing results from different studies lies in the fact that human responses are not always defined and interpreted in a consistent manner. A typology of human response to the indoor environment is proposed here to distinguish between these different responses. lt is based on two attributes of judgement: a) the aspect of evaluation, i.e., perceptual and affective; and b) the object of evaluation, i.e., personal .ttate and environment (1). The resulting matrix, shown in Figure I, consists of four domains of human response, each of which are then related to a set nf extraneous factors. In this context, perceptual responses,are defined as those in which respondent.t are used as sensing devices and affective responses are those in which respondents are used as judging devices. Object of Response Aspects of Judgment Perceptual Affective Environment  External Sensory  Acceptability, Response  Preference Personal st.ite  lnternal Sensory  Comfort Response,  General Physiological Symptoms Figure I. Typology of Human Response Human responses, designated as envrronnrental-nercepwnl in the typology are external sensory responses; responses that can be directly linked to a sensory receptor and that refer to responses made about the environment (eg., the room is hot). Personal-hercehttral responses include: a) internal sensory responses; responses that can be directly linked to a sensory receptor and refer to the personal state of the respondent (eg., I am hot): and b) general physiological sytnptotns; responses related to the personal state of the respondent and that are likely to be linked to a combination of sensory receptors (eg., I have a headache). All the responses included in the perceptual response coiumn of Figure 1 are sensory responses based on existing definitions (2, 3): they are not conscious judgements of the personal state or environment. Envirorrrnerrtnl-aJj<'ective responses such as acceptability and preference are predicated on an evaluation of the environment (eg., the thermal environment of the room is not acceptable). Personal-affective responses such as comfort are associated with an individual's personal state (eg.. I am thermally uncomfortable). People "construct their own definitions of what comfort rneans to them, and who, in so doing, create their own criteria by which to evaluate the acceptability of interior environments (4)." Thus, in addition to acceptahle perceptual responses, criteria for comfort also include other psychological or attitudinal criteria. EXTRANEOUS FACTORS Results of past research indicate that human responses can be associated with several

Procccdings of Indoor Air '93. Vol. I c,xtraneous factors. Additional factors not addressed by past studies are also proposed,, here, to clarify the distinctions between different human response domains. We hypothesize that extraneous factors for each domain can be arranged in a hierarchy as shown in Figure 2. Extraneous Factors Physical Environmental Factors Adaptive Factors ~ r- i ~ ~--i i ~ Intrinsic Factors I I Social Environmental Factors Psychological Environmental Factors Secular Drift I Risk Perception ~ I I luman Response Catcgorics Environmcntal- Perceptual Response Personal- Perceptual Response Environmental- Affective Response Personal- Affective Response Figure 2. Extraneous Factors Corresponding to liuman Response Categories 787 Three factors directly influence environmental-perceptual responses. These have been adapted from Rohles (5): 1 1 Physical enrnironmental fcictors include perceived physical environmental variables other than those being studied. For instance, if the evaluation is concerned with the influence of the mechanical system on thermal sensation, it is also necessary to consider the interaction of lighting factors on thermal response (6). 21 Adaptive factors are those that influence the physiological characteristics of the response being assessed, and may differ for different responses. They include factors such as ingestion, clothing, length of exposure, time between exposures, and activity levels that may increase or decrease the sensitivity of responses (i.e., adaptation) (5,7). For instance, in the case of odor responses, a very short exposure time may be sufficient to cause adaptation to a particular odor. Social factors have been excluded from this category because they only indirectly affect environmental-perceptual responses. 31 Intrinsic Factors are individual characteristics that affect human response. `ihey comprise factors such as age, gender, health status, lifestyle, predispositions, experience and sensitivity (5,7). Extraneous factors influencing personal-perceptual responses include thc three listed above as well as two others found in SBS literature (6.8). These may also indirectly affect environmental-perceptual responses as shown in Figure 2: 4) Social environnrenual fcictors such as job category, complaint handling by the management, control, choice and motivation may influence human response. 5) Psychological environmenlas fcrctors address the respondent's perception of J

788 t'rocccding5 c+f Indoor Air '9t, Vol. I activities carried out in the space such as job difficulty and complexity. Extraneous factors influencing environrnental-affective and personal-affective response domains have not yet been adequately identified in indoor environmental literature. In addition to the factors that influence perceptual responses, two other categories of factors are proposed. 61 Secular driJ't (9) refers to variables such as the public awareness of hazards and their impacts, the visibility of specific issues, social norms, past environmental influences, and other long or short term political influences (6). 71 Risk perception refers to attitudinal vaiiables such as: a) denial of environmental risks due to feelings of helplessness to control hazards or the belief (hat management is in control; b) acceptable levels of risk associated with the job as influenced by job satisfaction, quality of life and wage issues; c) ideas of fairness; and d) oriPntation to the future (10). M!F'1'lIUIUUILU(:ICA IL EFFECTS Methodologies used in current research to assess human response to indoor environmentss can be classified with respect to the length of time that respondents spend in thee environment being evaluated and the amount of control over the environment and respondents. Three types of methodologies are identified: I) SrrhjPCt rnrrlinclc,loRies in which respondents may spend a few minutes or hours in a controlled 'laboratory' environment and in which intrinsic factors may be controlled or uncontrolled: 2) 1'nrrel rnethodoloRiec in which respondents are briefly exposed to the environment being evaluated and in which intrinsic and environmental factors may be controlled or uncontrolled;' 3) [lrcrqynnt methodologies in which the respondents' occupancy of the environment being evaluated extends beyond the time that the investigaticm is carried out, and in which both intrinsic and environmental factors are uncontrolled. Each of these methodologies can be associated with factors such as reliability and validity of responses, sample design effects. the effects of missing data, stochastic effects, i.e., chance effects, and reproducibility and generalizability of the results (9). Subiect Methodologies: Control over physical and social environmental factors in subject methodologies increases the reliability of responses, especially if intrinsic factors are also controlled. f lowever, saniple bias effects are likely because the group of subjects may not comprise a representative sample (2). A large sample size may serve to reduce the impact of nris.cirrg clcltu and stoclra.ctic effects on test results. Test effects such as the effects of using intrusive apparatus like sensors may also significantly influence responses. Predictive irnlidity of responses is likely to be high because the relevant parameters can be controlled. Genernlizahility of results is likely to be low because laboratory conditions are difficult to replicate in the field. Panel Methodologies: The reliability of measures used in panel methodologies is likely to be high. This is achieved by: a) rigorous selection and training procedures, which also permit the use of stnaller samples; and b) repeated observations in the case of untrained panels, as each panelist serves as his/her own control. However, this methodology introduces sample hias because panelists generally consist of volunteers who may differ in significant ways from the general population, thus excluding the most sensitive populations such as people in ill health. The small sample size can also be expected to magnify the

I'lucccitmgso! Iniluur :\u '14, Vvl. f inqmct of nlissing dutu or inter-individual variations resulting from weather conditions, illness, hormonal changes, adaptation, habituation, novelty and contrast (7). Small sample size may also increase the impact of stochastic effects on the results. Panel methodologies impose some test efj`~rrs such as the order of presentation of stimuli, design and administration of questionnaires, treatment of panelists, cornfort, rnotivation and interaction with panel leader (7). Predic•ri>>e validity may be problematic if environmental factors cannot be completely controlled. Finally, different tests are likely to yield different results depending on environmental and intrinsic factors, thus posing a threat to the gener'alizahility of results. Occupant Methodologies: Occupant methodologies are likely to have lower reliahility of responses because of the lack of control on extraneous factors and inter- and intra- individual variability. Surnhle bias is likely to be lower than the other two methodologies when a study or investigation uses the entire population within a building. Ilowever, even when samples are taken, they are more likely to represent the population than other methodologies because relevant population characteristics can be more easily identified. i'he influence of missing clutu and stochastic eipct,t can be expected to depend on the sample size. Test effects can be reduced by appropriate design and administration of yuestionnaires. Predic:tive validity may be low, as suggested by the lack of consistent correlations between human responses and 'objective' measures in SE3S studies (11). Generalintbility of results is likely to be higher than that of other methodologies if site characteristics are representative of other field conditions. PR()1'{)SF.i) !'R()LI;SS An important objec;tive of building diagnostics is to achieve true positive or tnle negative results, while minimizing either false positive or false negative errors (12). Ihe following is a proposed process that incorporates the use of the preceding characterization for improving diagnostics of sick and healthy buildings. By linking the objectives of an investigation with the relevant human response domain(s) and the available methodologies, extraneous factors and methocfological effects can be identified. Then, appropriate methodologies can be selected to control for the effects of confounding factors. 'The concept of continuous degradation of building perforniance has been introduced in which healthy buildings, which are not pro- actively operatecl and maintained, degrade to undetected problem buildings, then to SBS cases, and finally to E3uililing Related Illness (BRI) cases (13). Diagnostic techniques for each of the eiebracleel cuiulititms should maximize true positive results and minimize false negative errors. I towever, for healthy buildings, true negative results should be maximized while minimizing false positive errors. Appropriate human response domains and methodologies are proposed for each building condition with the exception of buildings with 13R1. I3R( cases are excluded because characterization of clinical signs requires more than an assessment of perceptual or affective responses. For SBS cases, the personal-perceptual domain should be selected as it includes physiological responses characteristic of SBS syinptoms. Occupant methodologies are most suitable here, us responses may require a relatively long exposure titne, and Inoxt cif the relevant methodological effects can be controlled. For undetected problem huilclings, environlicental-perceptual responses, evaluated by panel niethodologies are most suitable as repeated observations may be reyuired for trend analysis. Finally, for healthy buildings, personal -al fcc:tive responses, assessed by means of occupant methodologies are mos[ appropt t:1tC.

790 ('rocccdings of Indoor nir '93, Vol. I This basic process can be applied 'to research studies as well as to the comparative analysis of the results of different studies. As a result of the foregoing analysis, we conclude that use of the proposed process should lead to a better understanding of exposures as they relate to different domains of human response. A(:KN()Vt"LEUGEMENT We are pleased to acknowledge the partial support of Phillip Morris USA for this study. f3IBLlU(,RAPIIY 1. International Standards Organization, Assessment of the Influence of the Thermal Environment Using Subjective Judgement Scales. ISO/DIS 10551. 1992. 2. Molhave L. Principles of Experimental Studies of The Sick Building Syndrome. Presentation. at "Conference on Indoor Air Pollution." J.B. Pierce Labs. 1990. 3. Baird JC, Berglund B, Lindvall T. Sensory Reactions and Indoor Air Quality. In: Cheremisinoff PN. & Young RA, ed., Indoor Odor Technology Assessment. Ann Arbor Science Publishers Inc., Ann Arbor. 1975; ch.fi:71-R1. 4. Cooper 1. Comfort Theory and Practice: Baniers to the Conservation of Energy By Building Occupants. Applied Energy 1982; 11:243-288. (Martin Center Working Paper). 5. Rohles fa 1. The Ecosystem Complex: A New Approach in Specifying the Man- Environment Relationship. Journal of Environmental. Systems 1971; I(4):321-32R. 6. Woods Jf;, Teichman KY, Seppanen OA, Suter P. Relationships Between Building Energy Management and Indoor Air Quality: Perceptions of Conflict and Opportunity in the United States and Europe. Proceedings of the Third International. Congress on Building Energy Management: ICBEM'R7:49-7i1. 7. Meilgaard M, Civille GV, Carr BT. Sensory Evaluation Techniques (2nd ed.). CRC press lnc., Boca Raton 1991. 8. Cone JE, Ilodgson MJ, eds. Problem Buildings: Building-Associ7ted Illness and the Sick Building Syndrome. Occupational Medicine: State of the Art Reviews 1999; vol4/no4. 9. Rossi P11, Freeman IIE. Evaluation: A Systernatic Approach (41h ed). Sage Publications, Newbury Park. 1989. 10. Brown MS. Communicating Information About Workplace Nazarcls: Effects on Worker Attitudes Toward Risks. In: Johnson BB & Covello VT, ed., The Social and Cultural Construction of Risk. D.Reidel Publishing Co., Fiolland 1987: ch. 1 t):251-274. 11. Putnam VL, Woods, JE, Bosman, TA, Objective Measures and Perceived Responses of Air Quality in Two Hospitals. In: The Human Equation: llealth and Comfort, IAQ'R9. San Diego, CA 1989: 241-250. 12. Lane CA, Woods JE, Bosman TA. Indoor Air Quality Diagnostic Procedures for Sick and Ilealthy Buildings. In: The Human Equation: Ilealth and Comfort, IAQ'89. San Diego, CA 1989:237-240. 13. Woods JE. Continuous Accountability: A Means to Assure Acceptable Indoor Environmental Quality. In: Indoor Air '90. Toronto, Canada 19R9:R5-97.