The psychological factors that affect the mapping of...

Proceedingsof9thWindsorConference:MakingComfortRelevantCumberlandLodge,Windsor,UK,7-10April2016.NetworkforComfortandEnergyUseinBuildings,http://nceub.org.uk

The psychological factors that affect the mapping of thermal sensation tothermalevaluation

TrevorKeeling1,EtienneRoesch2,DerekClements-Croome3

1BuroHappoldEngineering,LondonW1T1PD,andSchoolofBuiltEnvironment,UniversityofReading,[email protected];2CentreforIntegrativeNeuroscienceandNeurodynamics,UniversityofReading,Whiteknights,ReadingRG66AH,[email protected],UniversityofReading,Whiteknights,ReadingRG66AY,[email protected]

AbstractThephysical environment leads to a thermal sensation,which is thenevaluatedbybuildingoccupants. Thepresentstudyfocussesontherelationshipbetweensensationandevaluation.Weasked166peopletorecallathermaleventfromtheirrecentpast.Theywerethenaskedhowtheyevaluatedthisexperience intermsoftendifferentemotions(frustrated,resigned,dislike,indifferent,angry,anxious,liking,joyful,regretful,proud).We tested whether four psychological factors (appraisal dimensions) could be used to predict the ensuingemotions,comfort,acceptabilityandsensation.Thefourdimensionswere:conducivenessoftheevent,whowas responsible for the event, who had control and whether the event was expected. These dimensions,except for expectation, were good predictors of the reported emotions. Expectation was also useful forpredicting thermal sensation, acceptability and comfort: the more expected an event was the moreuncomfortable a person felt, and the less likely they reported a neutral thermal sensation. These resultssuggest a new way of conceptualising expectation and supports an embodied view of how subjectiveappraisals affect experience. Overall we show that appraisal dimensions mediate occupants’ evaluation oftheirthermalsensation,thissuggestsanadditionalmethodforunderstandingpsychologicaladaption.

Keywords:Emotion,appraisaldimensions,psychologicaladaption,thermalcomfort

1 Introduction1.1 Thermalenvironment,thermalsensation,andevaluativeresponseTreating thermal comfort as a problem of energy balances lends itself to a physiologicalapproach, which has been very successful. However, adaptive comfort theory providesscope fora rangeofpsychological factors tobeconsidered. It indeedseems intuitive thatthere issomepartofthermalcomfortforwhichthinkingandconceptualising is importantandconstituteoftheoverallexperience(Clements-Croome,2013).Theaimofthepresentstudyistorevealamechanismwherebythethermalenvironmentisconceptualisedandthisconceptualisationshapesthermalexperience.

For the purpose of this investigation, thermal experience can be brokendown into threecomponents. Firstly, physical environments, of temperature and airmovement constitutethemediumwithinwhichoccupantsoperate.Secondly, thermal sensation is the interfacebetween the occupant and the environment,which is predominately described using theASHRAE thermal sensation scale that runs fromcold, through cool, neutral,warm tohot.

Windsor Conference 2016 - Making Comfort Relevant - Proceedings 168 of 1332

Thirdly,anoccupant’sevaluationoftheirthermalenvironment,conventionallysatisfaction,comfortandacceptabilityareusedforevaluation(ASHRAE,2010).

This study looks at the psychological factors that shape how thermal sensations areevaluated.Todothiswelookathowfourappraisaldimensions(thepsychologicalfactors)shape acceptability, comfort, thermal sensation and emotions (the evaluations) resultingfrom the thermal environment. By doing this we show that the way people think about(appraise)theirenvironmentisimportantfortheirthermalexperience.

1.2 Appraisaltheory:factorsthataffecttheevaluationofsensationsAfundamentalquestionofemotionpsychologyishowcantwopeopleexperiencethesamesituationandhavedifferentsubjectiveexperiences.Fiftyyearsofworkonthetopicsuggeststhat there is an appraisal step between sensation and evaluative response (Figure 1)(Arnold,1960;Schereretal.,2001).Itisthissubjectiveappraisalprocessthatquiteliterallyshapestheuniqueemotionalresponsetoaparticularstimulus,givingrisetoanyoneofanumberofemotionsfromprideorjoytoangerorfrustration.Inthislight,twopeoplecanfeelthesamecoldstimulus,butappraisethesituationdifferentlydependingonwhetheritisconduciveorobstructivetotheirrespectiveneed.Forinstance,twopeoplecouldenteracoldoffice,andonefeelhappybecausetheyareabletowearafavouritejumper,whereastheothercouldfeelregretbecausetheydonothaveappropriateclothing.

Figure1:Appraisalsmediatehowsensationsareevaluated.

The example above exposes the relationship between appraisals and the subjectiveexperience that ensues. In this example, a single appraisal dimensionof conduciveness isusedtoevaluatethethermalsituation.Onepersonappraisesthesituationasconducivetotheir goal (wearing their favourite jumper) and experiences positive emotions, while theotherappraisesthesameenvironmentasnotconducive(itisunavoidablyunpleasant)andsoexperiencesnegativeemotions.Theseappraisalprocessesoccuronasubconsciouslevel;itistheroleoftheorytoexposewhattheyareandtheexperiencestheyleadto.

Weproposetouseappraisalsasproxiesforunderstandinghowparticipants’ lifehistorywillaffecttheirconceptualisationofagivenenvironmentalstimulusorscenario.This lifehistorygetsreducedtoalimitednumberoffixedappraisals.Appraisaltheoriesofemotionsattempttodefinethefinitesetofappraisalsthatarerelevanttoagivensituation(Schereretal.,2001).Oneofthesimplestappraisalsiswhetherastimulusisconsistentwithaperson’smotivesanddesiresornot,ifitisthentheresultantemotionislikelytobepositive,ifnotthentheemotionislikelytobenegative.

Thermalenvironment:i.e.19°C

Sensation:i.e.Cold

HappyRegretDislikeFrustration

Evaluation:

Appraisal:

Whodidtheythinkwasresponsiblefortheconditions?

Weretheconditionsconducivetotheindividualsneeds?Didtheyexpecttheconditions?

Whodidtheythinkhascontroloftheconditions?

Uncomfortable


Further appraisal dimensions can be used to predictwhich positive or negative emotionswillbeexperienced.Forexample,anothercommonappraisaliswhatorwhoisresponsiblefor the cause of the experience. If a person appraises themselves as responsible (for anegativeoutcome)thenregretwillbefelt.Ifsomeoneelseisappraisedasresponsibleangerordislikewouldbeexperienced.Finally,ifcircumstancewasappraisedascausingtheeventthen frustrationor resignationwouldbeexperienced.When several appraisal dimensionsareusedinconjunctiontheycanpredictspecificemotions(Scherer,2001).

Thevalueofappraisal theory is that itprovidesa framework tounderstandhowpeoples’conceptualisation of a situation affects their experience. This sheds light on themappingbetweensensationandevaluation.Fourcommonlyusedappraisalsaregoalconduciveness,responsibility for the situation, control, and expectation (Roseman, 1996; Scherer, 2001;Smith & Ellsworth, 1985). Used together they predict a range of positive and negativeemotions(Table1).Wesuggestthatthesefourappraisalsaresimilartoconceptsthathavebeenfoundtobeimportanttotheadaptivetheoryofthermalcomfort.Thereforeweaimtousethemtounderstandthepsychologicalcausesofthermalexperiences.Table1:Emotionsmappedtodifferentappraisalcombinations(derivedfromRoseman,1996;Scherer,1999).

1.3 ModelsofthermalcomfortBoth the heat balance and adaptive comfort approaches relate the indoor thermalenvironmenttotheevaluationofsatisfactionandcomfort(ASHRAE,2010;deDear&Brager,2001; Fanger, 1970). The universal thermal climate index (UTCI) relates outdoor thermalenvironment to thermal sensation (Fiala et al., 2012). These theories focus on therelationship between thermal environment and either thermal sensation or thermalevaluation. In their basic usage, they overlook processes that map a person’s thermalsensationtotheirthermalevaluation(Figure2).

Goalconducive Unconducive

Expected Unexpected Expected Unexpected

Circum

stance C

JoyfulResigned Frustrated

O Anxious Anxious

S Indifferent Indifferent

Other

C

LikingDislike Dislike

O Anger Anger

S Dislike Dislike

Self

C

Proud RegretOS

Respon

sibility

Control


Figure2:Thermalmodelstendtofocusonthethermalenvironmentandeithersensationorevaluation.Theytendtooverlooktherelationshipbetweensensationandevaluation.

Physiologicalmodelsofthermalexperienceaimtounderstandtheenergyflowswithinthebody. They split the body into several layered sections, each with different thermalproperties,whichareusedtopredicttheenergybalanceandtemperaturethroughoutthebody (Fiala et al., 2012; Schellen et al., 2013). Then by understanding these bodytemperatures and their rates of change, thermal sensation can be predicted (Fiala et al.,2012; Kingma et al., 2012). This still leaves the problem of relating a giventhermophysiological state to an evaluation of the thermal environment.Most often it isassumedthatthermalneutralityisdesiredandequatestomaximumcomfort(Fanger,1970).

Alliesthesiapredictswhyaneutralthermalsensation,oranyothersinglethermalsensation,willnotalways leadto thesameevaluation.Assuch itprovidesa theoreticalapproachtounderstandingtherelationshipbetweensensationandevaluation. Itsuggeststhatwhenapersonisoverheatedtheywillfindacoldsensationpleasant,whileifapersonisovercooledthey will find a hot sensation pleasant (Cabanac, 2006; Parkinson & de Dear, 2015).However Alliesthesia relies on a physiological approach to explain themapping betweensensationandevaluation.Incontrast,weaimtodemonstratepsychologicalreasonsthatthesamethermalsensationcanleadtodifferentevaluations.

Theoriesofpsychologicaladaptionsuggest thatexpectationandperceivedcontrolshiftorbroaden the comfort band. Previous thermal experience shapes habituation andexpectation,theseshiftthemappingbetweenphysicalconditionsandevaluation(Brager&deDear,1998;OleFanger&Toftum,2002).Thisleadsustohypothesisethatthegreaterthelevel of expectation themore likely someone is to be comfortable or accept the thermalconditions(H4TC).

Incontrast,occupants’ levelof controlbroadens their comfortband (rather thanshifts it)(Brager & de Dear, 1998). This leads us to hypothesise that the greater the level ofperceived control the more likely someone is to be comfortable or accept the thermalconditions(H3TC).

There is little inthethermalcomfort literatureabouttheappraisalofresponsibility foranevent. However, Leaman and Bordass (2007) do talk about naturally ventilated buildingshavinga forgivenessfactor.This leadsustotentativelyhypothesisethatwhenanevent isappraisedascausedbycircumstanceitwillbemoreacceptableandcomfortablethanwhenitisappraisedascausedbyanotherperson(H2TC).

Therehasalsobeenworkthatisnotsomuchconcernedwithpsychologicaladaptionbuthowcertainenvironmentscanengenderspecificpsychologicalstates(Farshchi&Fisher,2006);thisis known as embodied cognition (i.e. the similarity and interrelation of physical and


Sensation:i.e.Cold


Evaluation:i.e.Satisfaction

Adaptiveandheatbalanceapproaches

UTCIapproach


psychologicalexperience).Firstly,parallelshavebeendrawnbetweenthefeelingofphysicaland social warmth. It has been shown that experiencing physical warmth can promoteinterpersonalrelations(Williams&Bargh,2008)andexperiencingsocial inclusioncanaffectthejudgementoftemperatureanddesireforhotandcoldexperiences(Zhong&Leonardelli,2008).Thesefindingssuggestpsychologicalfactorscanaffectbodilysensationsdirectlyratherthanbychangingthenatureofevaluations.This leadsustohypothesisethattheeffectsoncomfortandacceptancewillbeaccompaniedwithaneffectonsensation(H1EtoH4E).

In this study we focus solely on thermal sensation and thermal evaluation. We test theutility of psychological factors (appraisal dimensions) for predicting thermal evaluations(comfort,acceptabilityandemotion).Wealsotestwhetherthepsychologicalfactorshaveasystematic effect on thermal sensations. This provides a methodology for understandingmultiplepsychologicalfactorsatonetime.Italsoelucidateswhetherpsychologicaladaptionchangesthemappingfromsensationtoevaluationorchangessensationitself.

1.4 HypothesesWerelatedtheabovefourappraisalsofinteresttothermalcomforttheory.Thisprovidesuswithasetofhypothesesthatusetheappraisalstoexplorepsychologicaladaption.

H1TC:Peoplewhoconsidertheeventtobeconducivetotheirneedsaremorelikelytofeel:comfortableandfindtheirthermalsensationacceptable.

H2TC:Peoplewhoconsidercircumstances,ratherthananotherperson,responsiblefor theeventaremore likely to feel: comfortableand find their thermal sensationacceptable.

H3TC:Peoplewhohavehighperceivedcontroloftheirofficeenvironmentaremorelikelytofeel:comfortableandfindtheirthermalsensationacceptable.

H4TC:Peoplewhoexpectedtheeventaremorelikelytofeel:comfortableandfindtheirthermalsensationacceptable.

Wefurtherproposethatappraisalsaffect themappingofsensationtoevaluationandarenotembodiedinthethermalsensationthatpeopleexperience.Totestthiswereframethefirstsetofhypothesisintermsofthermalsensation.GivenalackofembodimentweexpectonlyH1Etohold.

H1E:Peoplewhoconsidertheeventtobeconducivetotheirneedsaremorelikelytofeel:neutralratherthanhotorcold.

H2E: Peoplewho consider circumstances, rather than another person, responsiblefortheeventaremorelikelytofeel:neutralratherthanhotorcold.

H3E:Peoplewhohavehighperceivedcontroloftheirofficeenvironmentaremorelikelytofeel:neutralratherthanhotorcold.

H4E:Peoplewhoexpectedtheeventaremorelikelytofeel:neutralratherthanhotorcold.

Thefourappraisalsofinterestcanalsoberelatedtocommonemotions(

Table1).Thisprovidesuswithasetofhypothesesthattestappraisaltheory.

H1A:Peoplewhoconsidertheeventtobeconducivetotheirneedsaremorelikelytofeel:joyful,likingorproud.


H2A: Peoplewho consider circumstances, rather than anotherperson, responsiblefor the event are more likely to feel: resigned, anxious, frustrated or indifferentratherthandislikeoranger.

H3A:Peoplewhoconsidercircumstances,ratherthananotherperson,tocontroltheenvironmentaremorelikelytofeel:resigned,frustratedordislikeratherthanangeroranxiety.

H4A: Peoplewhoexpectedtheeventaremore likely to feel: resignedratherthanfrustrated.

2 Methods2.1 ParticipantsandbuildingsAs part of a wider field study focusing on evaluating the relationship betweenenvironmental factors and psychological experience, occupants of seven office buildingsresponded to our survey (N=166). The sample size is similar to other appraisal studies(Folkman & Lazarus, 1985, N=136-189; Roseman, 1996, N=182; Scherer & Ceschi, 1997,N=112).Respondentswerearangeofagesandgenders(Table2)andfromsevendifferentbuildings(Table3).Participantswererewardedwithasnackoftheirchoice.

Table2:Summaryofparticipants.

N Female Male Undisclosed 18-34yrs

35+yrs

Undisclosed

166 105 57 4 84 77 5

Table3:Overviewofbuildings.NV=naturallyventilated,MM=Mixedmode,AC=fullyairconditioned.

Building N(resp.) Occupier Typology Plan HVAC

A 9(18%) Design Openplan 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 Openplan Shallow NV

F 25(2%) Charity Openplan Deep AC

G 39(26%) Design Openplan Shallow NV

2.2 QuestionnairedevelopmentTapping into the subjective experience of an individual is a major challenge, the mereattempt to ask a question is likely to disrupt the unfolding experience altogether. Toeliminate this disruptionwe chose to use a recall survey. This also allows us to access amuchgreater rangeof experiences than if itwasnecessary tobepresent,measuring thethermalenvironmentastheexperienceunfolded.Thereliancesolelyonuserreporteddata,withlittleornomeasurementofthephysicalnatureofthestimuli,iscommoninpsychology(Fontaine et al., 2007) and is appropriate here because of this study’s focus on therelationshipbetweenparticipants’sensationandtheirevaluation.


Therecallsurveystartedwithapromptfortheparticipantstorecallaneventindetail.Todothistheywereaskedto:

“Imagineaspecifictimewhenyouhavebeenawareofthetemperatureinyourofficeand it has given rise to strong feelings. Describe what happened leading up to theeventandhowyoufelt.”

Afterthis,anumberofquestionswereaskedabouteachofthefourappraisaldimensions.Detailsofthequestionsandhowtheywerecombinedcanbefoundintheappendix.Thesewereusedtounderstand:

• Whethertheparticipantfelttheeventwasconducivetothem(appraisal1);• Whoorwhattheythoughtcausedtheevent(appraisal2);• Whoorwhattheythoughtcontrolledconditionsintheiroffice(appraisal3);• Howmuchtheyhadexpectedtheeventtohappen(appraisal4).

To finish the survey, therewasanopen response todescribe feelingsanda closed listofemotions to choose from: frustrated, resigned, dislike, indifferent, angry, anxious, liking,joyful, regretful, proud, or, none of these. Then three questions were asked about theparticipant’s thermal experience, using a thermal sensation scale, a comfort scale and anacceptabilityscale.

2.3 AnalysistechniqueTo test this theory, we examine whether appraisals have an effect upon emotions,acceptability, comfort and sensation. The model used compares the likelihood of aparticular evaluation, dependent upon the score on an appraisal dimension. The mostappropriate statistical model for this is a logistic regression model. This predicts thepresence or absence of a factor (a set of emotions or acceptance) dependent upon anorderedfactor(theappraisaldimension).Anextensiontothismodel istheordinal logisticmodel, which predicts the likelihood of achieving a given level of comfort or sensationdependingonanappraisaldimension.

Equation1showsthelogisticalregressionmodel.Themodelcomprisesalinearfunctionanda link function. Just as in a standard linear model the coefficients are derived so as tomaximise the fit of the model. The link function𝑚()transforms the linear model to aprobabilityofsuccess,𝜋% boundedbetweenoneandzero.Thereareseveralfunctionsthatfitthiscriteria,themostcommonlyusedarethe“logit","probit","cauchit","log",andthe"complementaryloglog”(McCullagh&Nelder,1989).Inthisstudywecompareallpossiblelinkfunctionsandselectedthebestfittingmodel.

𝜋% = m 𝛽) + 𝛽+𝑥% + 𝛽-𝑥%+. . . Equation1

Tocomparethelogisticmodelsweusedachisquaretestofthedevianceaccountedforbytheregressionmodel.Forboththelogisticandordinallogisticmodelwealsocharacterisedthemodelbythelikelihoodthattheregressioncoefficients(βi)arenon-zero.

3 Results3.1 Theexperiencesreported3.1.1 Sensation,comfortandacceptabilityParticipant were asked to report their thermal experience during the period that theyrecalled. Generally, they recalled times when they were at the extreme of thermal


sensations,eitherhotorcold (Table4).Mostparticipants found this tobeuncomfortableratherthanveryuncomfortable(Table5).Overwhelminglytheseconditionswerefoundtobeunacceptable(Table6).

Table4:Thermalsensationcounts.

Thermalsensation CountCold 30Cool 6Slightlycool 1Neutral 8Slightlywarm 6Warm 30Hot 84Undisclosed 1

Table5:Comfortcounts.

Comfortrating CountVeryuncomfortable 42Uncomfortable 84Slightlyuncomfortable 38Comfortable 1Undisclosed 1

Table6:Acceptabilitycounts.

Acceptabilityrating CountNotacceptable 129Acceptable 33Undisclosed 4

3.1.2 EmotionsrecalledParticipants were asked to choose an emotion that best matched their feelings from aclosed list. No one reported a positive emotion or an emotion associated with personalresponsibility,i.e.regret(Table7).Mostlyparticipantsreportedfeelingfrustrated,resignedor dislike.A smaller number of participants felt indifferent, angry or anxious. Therewerealsosixteenwhofelt thatnoneof the tenemotions fittedwellwithhowthey felt.Acrossbuildings the trendwas generally the same, except Building A and Bwhere peopleweremorelikelytofeeldislikeandbuildingFweretheyweremorelikelytofeelangry(Figure3).

Table7:Theemotionsreportedacrossallbuildings.

Emotion CountFrustrated 74Resigned 30Dislike 20Noneofthese 16Indifferent 10Angry 8Anxious 8Liking 0Joyful 0Regretful 0Proud 0

Figure3:Theemotionsreportedacrossallbuildings.

Frust. DislikeResign. None.

All

Bld.A

Bld.B

Bld.C

Bld.D

Bld.E

Bld.F

Bld.GIndiff. Angry Anxious


3.2 TheappraisalsGenerally,participantsreportedthattheeventwasunpleasantandworsenedtheirabilitytowork.Wealsoaskedwhotheythoughtwasresponsible fortheevents leadinguptotheiremotionalexperience(Figure4).Theyrarelythoughttheythemselveswereresponsible.Weaskedtheparticipantswhotheythoughtwasgenerallyincontrolofthetemperatureintheiroffice(Figure5).OccupantsofbuildingFfelttheyhadespeciallylittlecontrol.OccupantsofbuildingCandDthoughtnopersonwasincontrol.Acrossmostbuildingscircumstanceswasthoughttocontrolconditions.Overall,therewasamixtureofwhetherpeoplethoughttheevent they reported could have been expected. However, there is a lot of differencebetweenbuildings(Figure6).OccupantsfrombuildingEandFtendedtoreporteventsthatwereunexpected.Elsewhereeventsreportedhadbeenexpected.

Figure4:Whoisappraisedasresponsiblefortheevent,acrossthedifferentbuildings.

Notatall Notatall NotatallExtremely Extremely Extremely

Oneself Anotherperson CircumstanceAll

Bld.A

Bld.B

Bld.C

Bld.D

Bld.E

Bld.F

Bld.G


Figure5:Whoisappraisedasincontrolingeneral,acrossthedifferentbuildings.

Figure6:Appraisalofexpectednessoftheevent,acrossdifferentbuildings.

3.3 UsingappraisalstopredictemotionsTheabsenceofpositiveemotionsandtheabsenceofpositiveappraisalsofconducivenessisinaccordancewithappraisaltheory.However,thelackofpositiveemotionsalsomeansitisdifficulttobuildastatisticalmodelforvalidation.ThispartiallysupportH1A.

Notatall Notatall NotatallExtremely Extremely Extremely

Oneself Anotherperson CircumstanceAll

Bld.A

Bld.B

Bld.C

Bld.D

Bld.E

Bld.F

Bld.G


Figure7:Appraisalsofresponsibilityandcontrolhaveaneffectontheemotionreported.

Foreachappraisaltheemotionsreportedwerepartitionedintwogroupsaccordingtotherelevant hypothesis, i.e. for responsibility, one group was aligned with the appraisal ofcaused by another (dislike and angry) and the other with appraisal of caused bycircumstance(frustrated,resigned,indifferent,anxious).Whereemotionswerenotrelevantto the hypothesis they were discarded from the analysis (i.e. for H3A indifference isdiscarded). Figure 7 showshow the likelihoodof feeling one set of emotions rather thananothervarieswithparticipants’appraisal.

Table8:Characteristicsforemotionsmodels.

Appraisal Bestlinkfunctionχ2goodnessoffit Modelcoefficients

χ2 Singletailed β0 β1

Responsibility Poisson 7.1(df=1) P=0.01 -1.1(p<0.001)

-0.28(p=0.003)

Control Cauchit 3.8(df=1) P=0.05 -2.7(p<0.001)

-0.65(p=0.03)

Expectation Cauchit 2.70(df=1) P=0.10 -0.15

(p=0.81)0.21

(p=0.11)

Event appraised as caused by others

Event appraised as caused by circumstance

0

1

Likl

ihoo

d of

feel

ing

disl

ike

or a

nger

Event appraised as controlled by other

Event appraised as controlled by circumstance

0

1

Likl

ihoo

d of

feel

ing

angr

y or

anx

ious

Event appraised as unexpected

Event appraised as expected

0

1

Likl

ihoo

d of

feel

ing

frust

rate

d

Event appraised ascaused by others

Event appraised ascaused by circumstance

0

1

Likl

ihoo

d of

feel

ing

disl

ike

oran

ger

Event appraised ascontrolled by other

Event appraised ascontrolled by circumstance

0

1

Likl

ihoo

d of

feel

ing

angr

yor

anxi

ous

Event appraised asunexpected

Event appraised asexpected

0

1

Likl

ihoo

d of

feel

ing

frust

rate

dArea of bubble is proportional tonumberofpeoplemakingthischoice

xmarkstheexpectedvaluefromoursurvey

95%CIsforexpectedvalue


Severaldifferent link functionswere testedtomodel thedata, statistical testsof thebestmodel are reported in Table 8. These suggest that there is a tendency to feel angry ordislike when another person is deemed responsible for the thermal experience, thissupports H2A. The results also show a tendency to feel angry or anxious when anotherperson is appraised as in control of the thermal experience, this support H3A. For theappraisalofexpectednessthereisnotsuchanobviouspatternasfortheotherappraisals,thisiscountertoH4A.

3.4 UsingappraisalstopredictcomfortandacceptabilityTheabsenceofpositiveappraisalsofconducivenessmeansH1TCwasnot tested.Figure8shows how the likelihood of finding a thermal experience acceptable varies withparticipants’appraisal.Figure9showstheeffectofthesameappraisalsoncomfortrating.

Figure8:Theappraisalsofexpectationhasasmalleffectonacceptability.

Severaldifferentlinkfunctionsweretestedtomodelthedata,statisticaltestsofthebestmodelsarereportedinTable9.Thesesuggestthattheappraisalshavelittleeffectontheacceptabilityoftheexperience(countertoH2TCandH3TC).Thereisaweaklinkthatsuggeststhatthemoreasituationisexpectedthelessacceptableitis(oppositeofH4TC).Similarresultsarefoundforcomfort(Table10).



0

1

Likl

ihoo

d of

feel

ing

disl

ike

oran

ger



0

1

Likl

ihoo

d of

feel

ing

angr

yor

anxi

ous



0

1

Likl

ihoo

d of

feel

ing

frust

rate

d

Area of bubble is proportional tonumberofpeoplemakingthischoice


95%Cisforexpectedvalue



Unacceptable

Acceptable

Exp

ecte

d ac

cept

ance

sco

re

Low pereived control High pereived control

Unacceptable

Acceptable

Exp

ecte

d ac

cept

ance

sco

re



Unacceptable

Acceptable

Exp

ecte

d ac

cept

ance

sco

re


Table9:Characteristicsforacceptabilitymodels.

Appraisal Bestfittinglinkfunction

χ2goodnessoffit Modelcoefficientsχ2 Singletailed β0 β1

Responsibility Cauchit 0.82(df=1) P=0.37 -2.2(p<0.01)

0.24(p=0.36)

Perceivedcontrol Probit 2.46(df=1) P=0.12 -1.3

(p<0.001)0.16

(p=0.11)

Expectation Probit 3.48(df=1) P=0.06 -0.27(p=0.39)

-0.10(p=0.06)

Figure9:Theappraisalsofcontrolandexpectationhaveasmalleffectoncomfort.

Table10:Characteristicsforcomfortmodels.

Appraisal β1 SE t-value p-valueResponsibility 0.18 0.10 1.76 0.08

Perceivedcontrol -0.20 0.13 -1.55 0.12Expectation -0.16 0.07 -2.26 0.02

3.5 UsingappraisalstopredictdeviationfromneutralsensationTheabsenceofpositiveappraisalsofconducivenessmeansH1Ewasnot tested.Figure10showshowthelikelihoodofreportinganeutralthermalsensationchangeswithparticipants’appraisal.



0

1

Likl

ihoo

d of

feel

ing

disl

ike

oran

ger



0

1

Likl

ihoo

d of

feel

ing

angr

yor

anxi

ous



0

1

Likl

ihoo

d of

feel

ing

frust

rate

d

Area of bubble is proportional tonumberofpeoplemakingthischoice





Veryuncomfortable

Comfortable

Exp

ecte

d co

mfo

rt sc

ore


Veryuncomfortable

Comfortable

Exp

ecte

d co

mfo

rt sc

ore



Veryuncomfortable

Comfortable

Exp

ecte

d co

mfo

rt sc

ore


Figure10:Theappraisalsofexpectationhasaneffectonthermalsensation.

Table11showsthedifferentmodelcharacteristics.TheresultsconfirmthenullhypothesisforH2EandH3E.TheyruncountertoH4Ebecausethemoreexpectedaneventisthelesslikelythepersonistoreportbeingthermallyneutral.

Table11:Characteristicsforsensationmodels.

Appraisal β1 SE t-value p-valueResponsibility 0.04 0.12 0.30 0.76

Perceivedcontrol 0.07 0.15 0.51 0.61Expectation 0.24 0.87 2.75 0.006

4 DiscussionWe have presented evidence that supports hypothesis HA1, HA2 and HA3. These resultsshow that appraisal processes are important for shaping evaluation of the thermalenvironment.Thiswork supports thenotion that thermalexperience is richandcomplex,and requires understanding of how people conceptualise their thermal environment(Heschong,1979).Itispossiblethatappraisals,especiallyconduciveness,couldbedrivenbythethermophysiologicalstateoftheparticipant,thoughthereisnoneedforthistobethecase.

The appraisals of responsibility and controlwere less useful for predicting the traditionalthermalcomfortevaluationsofacceptabilityandcomfort(H2TCandH3TC).Thiscontradicts



0

1

Likl

ihoo

d of

feel

ing

disl

ike

oran

ger



0

1

Likl

ihoo

d of

feel

ing

angr

yor

anxi

ous



0

1

Likl

ihoo

d of

feel

ing

frust

rate

dArea of bubble is proportional tonumberofpeoplemakingthischoice





Neutral

Hot or coldE

xpec

ted

devi

atio

n of

sens

atio

n fro

m n

eutra

l


Neutral

Hot or cold

Exp

ecte

d de

viat

ion

ofse

nsat

ion

from

neu

tral



Neutral

Hot or cold

Exp

ecte

d de

viat

ion

ofse

nsat

ion

from

neu

tral


theextensiveliteratureonperceivedcontrolandthermalcomfort(Brager&deDear,1998;Hellwig,2015).Thisanomalousresultcouldbeasideeffectoftherecallmethod.Therecallmethod provides access to thermal events that are of high saliency. It is possible thatperceivedcontroldoesnotaffecttheseverityofthemostextremebadevents.

The appraisal of expectation was successful in predicting comfort evaluation (H4TC).However,thecorrelationwasoppositetothatexpected.Themoreaneventwaspredictableor expected themore uncomfortable it was. Thermal comfort theory would predict thatoccupantsacclimatisetoeventsovertime(Brager&deDear,1998).Ourresults,however,suggestthateventsthatarenovelandfleetingmaycauselessdiscomfortthanrecurringandpredictableproblems.

When we asked about expectation it seems that we were asking about whether aproblematicoreventfulsituationwasrecurrent.Whereastheclassicexpectationofthermalcomfortreferstorepeatedexposuretoaubiquitousclimaticexperience.Consideringthisitappears that our results drawattention to a different typeof expectation effect.Namelythatwhenproblematicconditionsarerecurrenttheybecomeslessandlessacceptable.

The work on psychological adaption and embodied cognition suggest two differentmechanismsthroughwhichpsychologicalfactorscouldaffectthermalexperience.Thefirstsuggests that psychological factors change the mapping between thermal sensation andthermal evaluation. These theories suggest that the benefit of personal control is that itreducesstressfrommildlyunfavourableconditionsandeffectivecontrolprovidespleasure(Hellwig,2015).Incontrastembodiedcognitionsuggestthatthepsychologicalfactorwouldchange thermal sensation itself. Interestingly a study carriedout in a climate chamberbyZhouetal.(2014)suggestthatperceivedcontrolactuallychangesbothbodilysensationaswellasreducingstress.

Our results support theories of embodiment becausewhere appraisals have an effect oncomforttheyalsohaveaneffectonsensation(H2-4TCandH2-4E).However,thiscanonlybetakenasweaksupport forembodimentbecausewecannottraceourresultsbacktoaspecific thermal environment. The inclusion of synchronous temperature measurementswouldprovideconclusiveevidencethattheappraisalcausedasensationchangeasopposedtothermalsensationscausingbothcomfortandappraisal.

Thelackofpositiveemotionssupportsworkthatsuggeststemperatureisahygieneorbasicfactorresponsibleonlyfordissatisfaction(Herzberg,1964;Kim&deDear,2012).However,it may be that people just chose to focus on negative events from their past. Futureinvestigations could be contrived to test this by asking participants to describe twoexperiences and stipulate that one had to be positive. This approach would provide agreater range of experiences and hopefully contribute positive emotions to improve theanalysisofH1A.

Toimprovethemethodsapplicationandrepeatabilitythesurveycouldalsobemadeeasiertoanalyse.First,itshouldbemadeeasiertoobtainscoresforeachappraisaldimension.Thecurrentsystemofcombiningmanyordinalresponsesisconvolutedandbuildsinuncertaintytothemethod.Second,continuousresponseforvariablescouldbeused.Thiswouldmeanthatanalysiscouldbedonewithgenuineratioscalenumbersratherthananordinalscalethatwastransformedintoaratioscale.


5 ConclusionsAppraisal theory provides a simplified way to encapsulate peoples’ thoughts about theirthermal experience. These thoughts not only cover a person’s core temperature andperipheralthermalstimulusbutalsotheirpastexperiencesandfuturedesires.Thetheorydoesnottrytopredictwhypeoplemakecertainappraisalsbutitidentifieswhichappraisalsarekey.Overall,ourresultsshowthatitisthecombinationoftheseappraisalsthatshapeathermalexperience.Multidimensionalappraisalsrequiremultidimensionalevaluations,andinthiscasewehavesuccessfullyusedtenemotionstodescribethermalexperience.

Ouranalysissuggestsanewaspecttohowexpectationaffectspsychologicaladaption.Weobservethatrecurrentproblems(thosethathappenedoftenandwerepredictable)resultedingreatestdiscomfort.Peopledidnotappeartoadapttothem.Thissuggestsanalternativewaytoconceptualiseexpectation.

With furthermodifications, the surveydevelopedhere couldbeusedas adiagnostic toolwherediscomfortanddissatisfactioniscausedbecauseofpsychologicalfactors(asopposedto poor thermal conditions). From this it may be possible to design a programme ofmeasures that tackled those psychological causes. This would be in contrast to currentindustryapproachesthat focusoncostly technical fixesandchaseevermorecontroloverthephysicalenvironment.

Interpretation,meaningandotherpsychologicalapproacheshavebeenshowntoplayapartinsubjectiveexperienceacrossarangeofindoorenvironmentalqualityindices(Kwonetal.,2011;Lehman,2011).Thisleadsustosuspectthatthismethodcouldbeusedtounderstandemotionsandtheirappraisalscausedbymultisensoryexperiencesofbuildings,beyondonlythermalcomfort.

AcknowledgementsThis work would not have been possible without the Institute of Technology for aSustainable Built Environment (TSBE), Reading University, the Engineering and PhysicalScienceResearchCentre (EPSRC)DoctoralTrainingScheme (grantnumberEP/G037787/1)andBuroHappoldEngineering.Queriesregardingunderlyingresearchmaterialsrelatedtothisstudy(forexampledata,samplesormodels)shouldbeaddressedtothecorrespondingauthor.

ReferencesArnold,M.B.1960.Emotionandpersonality:Cassell.ASHRAE.2010.Standard55-2010Thermalenvironmentalconditionsforhumanoccupancy.Atlanta:ASHRAE.Brager,G.S.,&deDear,R.J.1998.Thermaladaptationinthebuiltenvironment:aliteraturereview.EnergyandBuildings,27(1),83-96.Cabanac,M.2006.Pleasureand joy, and their role inhuman life. InD.Clements-Croome(Ed.),Creatingtheproductiveworkplace(2ed.,pp.3-13).London:Taylor&Francis.Clements-Croome, D. 2013. Sustainable healthy intelligent buildings for people. In D.Clements-Croome(Ed.),Intelligentbuildings:design,managementandoperation(2nd.ed.).London:ThomasTelford.de Dear, R., & Brager, G. 2001. The adaptive model of thermal comfort and energyconservationinthebuiltenvironment.InternationalJournalofBiometeorology,45(2),100-108.


Fanger,P.O.1970.Thermalcomfort:analysisandapplicationsinenvironmentalengineering.Copenhagen:DanishTechnicalPress.Farshchi,M.A.,&Fisher,N.2006.Emotionandtheenvironment:theforgottendimension.InD.Clements-Croome(Ed.),Creatingtheproductiveworkplace(2ed.,pp.55-74).London:Taylor&Francis.Fiala, D., Havenith, G., Bröde, P., Kampmann, B., & Jendritzky, G. 2012. UTCI-Fialamulti-nodemodel of human heat transfer and temperature regulation. International Journal ofBiometeorology,56(3),429-441.Folkman,S.,&Lazarus,R.S.1985.Ifitchangesitmustbeaprocess:studyofemotionandcoping during three stages of a college examination. Journal of Personality and SocialPsychology,48(1),150.Fontaine,J.R.J.,Scherer,K.R.,Roesch,E.B.,&Ellsworth,P.C.2007.TheWorldofEmotionsisnotTwo-Dimensional.PsychologicalScience,18(12),1050-1057.Hellwig, R. T. 2015. Perceived control in indoor environments: a conceptual approach.BuildingResearch&Information,43(3),302-315.Herzberg,F.1964.Themotivation-hygieneconceptandproblemsofmanpower.PersonnelAdministration.Heschong,L.1979.Thermaldelightinarchitecture.Cambridge,Massachusetts:M.I.T.Press.Kim, J., & de Dear, R. 2012. Nonlinear relationships between individual IEQ factors andoverallworkspacesatisfaction.BuildingandEnvironment,49(0),33-40.Kingma,B.R.M.,Schellen,L.,Frijns,A.J.H.,&vanMarkenLichtenbelt,W.D.2012.Thermalsensation:amathematicalmodelbasedonneurophysiology.IndoorAir,22(3),253-262.Kwon, S.-H., Chun, C., & Kwak, R.-Y. 2011. Relationship between quality of buildingmaintenance management services for indoor environmental quality and occupantsatisfaction.BuildingandEnvironment,46(11),2179-2185.Leaman, A., & Bordass, B. 2007. Are users more tolerant of ‘green’ buildings? BuildingResearch&Information,35(6),662-673.Lehman, M. L. 2011. How sensory design brings value to buildings and their occupants.IntelligentBuildingsInternational,3(1),46-54.McCullagh,P.,&Nelder,J.A.1989.Generalizedlinearmodels(2ed.):ChapmanandHall.Ole Fanger, P., & Toftum, J. 2002. Extension of the PMV model to non-air-conditionedbuildingsinwarmclimates.EnergyandBuildings,34(6),533-536.Parkinson, T., & de Dear, R. 2015. Thermal pleasure in built environments: physiology ofalliesthesia.BuildingResearch&Information,43(3),288-301.Roseman,I.J.1996.AppraisalDeterminantsofEmotions:ConstructingaMoreAccurateandComprehensiveTheory.Cognition&Emotion,10(3),241-278.Schellen, L., Loomans,M.G. L. C., Kingma, B. R.M., deWit,M.H., Frijns, A. J. H.,& vanMarken Lichtenbelt, W. D. 2013. The use of a thermophysiological model in the builtenvironment to predict thermal sensation: Coupling with the indoor environment andthermalsensation.BuildingandEnvironment,59(0),10-22.Scherer, K. R. 1999. Appraisal Theory. In T. Dalgleish &M. J. Power (Eds.),Handbook ofcognitionandemotion.Chichester:Wiley.Scherer,K.R.2001.Appraisalconsideredasaprocessofmultilevelsequentialchecking.InK.R. Scherer, A. Schorr, & T. Johnstone (Eds.), Appraisal processes in emotion : theory,methods,research(pp.92-120).Oxford:OxfordUniversityPress.Scherer, K. R., & Ceschi, G. 1997. Lost luggage: A field study of emotion–antecedentappraisal.MotivationandEmotion,21(3),211-235.


Scherer, K. R., Schorr, A., & Johnstone, T. 2001. Appraisal processes in emotion: theory,methods,research.Oxford:OxfordUniversityPress.Smith,C.A.,&Ellsworth,P.C.1985.Patternsofcognitiveappraisal inemotion. JournalofPersonalityandSocialPsychology,48(4),813.Williams,L.E.,&Bargh, J.A.2008.ExperiencingPhysicalWarmthPromotes InterpersonalWarmth.Science,322(5901),606-607.Zhong,C.-B.,&Leonardelli,G. J.2008.ColdandLonely.Psychological Science,19(9),838-842.Zhou,X.,Ouyang,Q.,Zhu,Y.,Feng,C.,&Zhang,X.2014.Experimentalstudyoftheinfluenceofanticipatedcontrolonhumanthermalsensationandthermalcomfort. IndoorAir,24(2),171-177.


AppendixI:Surveyquestions


AppendixII:CombinationofindependentvariablesThis section summarises how individual survey questions were combined into scores foreach appraisal dimension. This was not done for conduciveness because of the resultsobtained.

Fortheappraisalofresponsibilitytheparticipant’ssevenrelevantanswersweregroupedinto those that suggested a circumstantial cause (i.e. building or weather) and those thatsuggestacausebyanotherperson(i.e.colleagueorbuilding/facilitiesmanager).Thosethatsuggestedacausebyselfwerediscarded.Eachofthetwogroupingswerecombinedintoasinglescorebytakingthemaximumratingacross thegroup(i.e.extremelyequalshighestpossiblescore).Wetookthemaximumacrossthequestionsinsteadofanaveragebecausethe emotional response is contingent on the relative importance of the highest scoringitemsineachgroup(i.e.overalldidthisparticipantthinktheeventwasmorecausedbyaperson,regardlessofthespecificperson,ormorecausedbycircumstance,regardlessofthespecific circumstance). These two groupings were combined into a single score bysubtractingthepersonappraisalfromthecircumstanceappraisal.

𝐴- = max 𝑄10, 𝑄11, 𝑄12, 𝑄13 − max(𝑄8, 𝑄9)For the appraisal of control the nine relevant answers were combined into three scores(circumstance, selfandother)by taking themaximumratingacross the relevantanswers.Thecircumstanceandselfgroupingswerecombinedintoasinglescorebysubtractingtheperson appraisal from the circumstance appraisal, this was used to test H3A. Those thatsuggestedacausebyselfwereseparatedandusedfortestingH3TCandH3E.

𝐴< = max 𝑄22, 𝑄23, 𝑄24 − max(𝑄19, 𝑄20, 𝑄21)

𝐴<>?@A?%B?CADEF@DG = max 𝑄18, 𝑄25, 𝑄26

Producing a combined expectation score, from the four relevant questions, was morecomplicated than theother appraisal scales. Firstly the twoquestions “Howoftendoes ithappeninwinter?”and“Howoftendoesithappeninsummer?”weremergedbytakingthehighestscoreasperbefore.Thisproducesascoreequivalentto,doesithappenofteninanyseason.Thiswasthencombinedwiththelevelsforthetworemainingquestionsbyaddingscores that were suggestive of predictability and subtracting scores associated withsuddenness. This produced a scale running from -3 (most unexpected) to 13 (mostexpected).

𝐴J = max 𝑄16, 𝑄17 + 𝑄15 − 𝑄14


The psychological factors that affect the mapping of...

Documents

Transcript of The psychological factors that affect the mapping of...