INT RE\. APP PSYCHOL VOL. 27. NO. I Towards a taxonomy of responses to the built environment 1 ANKE OOSTENDORP, SUSAN McMASTER, MARTIN ROSEN, and PETER WAIND University of Toronto, Canada One of the major areas in environmental psychology deals with the perception and assessment of environments (see Craik. 1973; Kaye, 1975). Researchers in this area concern themselves with the determination ofwhat aspects of environments are attended to and what psychological responses can be related to these salient aspects. In other words, they ask: How do we react to design characteristics of a building, a street, or a city? The importance of this question in terms of psychological contributions to environmental design has been recognized by several researchers. Peopleâs reactions to residential areas (for example, Franke and Bortz, 1972; Acking and Sorte, 1972), to city districts (Lowenthal and Riel, 1972), to exteriors or interiors of private or public buildings (Canter, 1969; Ekambi-Schmidt, 1972; Hershberger, 1972; Kuller, 1972), and to buildings representing a variety of architectural styles (Oostendorp and Berlvne, 1978) have been studied. Researchers in this area have primarily utilized the semantic differential technique (see Osgood, Suci, and Tannenbaum, 1957) to determine the underlying variables in psychological responses to the built environment. Limitations of this technique have been discussed by Heise (1969) and Miron (1972). Bechtel (1975) pointed to several disadvantages of the technique when used in environmental design research. The major problem of the semantic differential is that i t can point only to those underlying variables that are related to sometimes arbitrarily selected rating scales. Thus, one tests the salience of those attributes that were judged to be relevant to the stimulus material on an a priori basis, rather than determining which characteristics were most salient to the subjects (cf. Garling, 1973). Furthermore, our evaluative and other emotional reactions may be a matter of early conditioning (Berlyne, 197 1 ). Hence we may not always be aware of these reactions, let alone be able to express them in self-reports. Therefore, in order to develop a taxonomy of environments in terms of psychological reactions to them, researchers have turned to techniques of multidimensional scaling which have more recently become available (for example, Blasdel, 1973; Garling, 1973; Oostendorp and Berlyne, 1978). Any data that reflect some form of distance between stimuli can serve as inputs for these techniques. The most common input is provided by requiring subjects to indicate the degree of similarity of all possible pairs of a set of stimuli. Similarity judgements are generally assumed to reflect the internal âcodingâ (perception, copitionj of the stimuli in terms ofone or more continuously varying dimensions. The computerized treatments of such data will discover the number of dimensions relevant to the perception of the set of stimuli, and will locate the stimuli along each of the dimensions. In making his similarity ratings a subject may attend to any attribute or collection of attributes that are important in the internal representation of a pair of stimuli. When judging the following pair a subject may use the same or another criterion to categorize the stimuli. Moreover, subjects may give differentia1 attention to dimensions. So, Carroll and Chang ( 1970) developed an individual differences scaling model (INDSCAL) which allows for the discovering of those dimensions that subjects have in common. The data input for an INDSCAL procedure consists of one similarity matrix for each subject. The output gives the locations of the stimuli as well as the subjects along the dimensions. The location of a subject along a dimension indicates 10 TAXONOMY OF RESPONSES T O THE ENVIRONMENT how much weight the subject attributes to the dimension. The stimulus locations indicates how much of the attributes relevant to the dimension was possessed by that stimulus. I t might thus be concluded that such results provide the basis for a taxonomy of psychological responses to a set of stimuli. Hence, multidimensional scaling techniques are free from the researcherâs preconceptions about relevant attributes. Unfortunately, the techniques bring up their own problems, namely, those of interpretation of the dimensions and of choice of the number of dimensions. The latter problem can be solved without great difficulty, when using the INDSCAL technique. In this technique the squared subject weights produce an estimate of the percentage of variance accounted for by a dimensionality. Therefore, one should choose a smaller dimensionality, ifthe larger accounts for only slightly more variance than the smaller dimensional solution (cf. Shepard, 1972). Answers to the question of what stimulus characteristics correspond to the dimensions can be obtained in a number ofways. Shepard ( 1 972) suggested the study ofcommon characteristics ofclusters ofstimuli in the spatial configuration. This method was employed by Garling (1973), who compared stimulus configurations obtained through factor analytic and multidimcnsional scaling procedures. However, such an inspection should be regarded only as a tentative solution. The most fruitful procedure seems to be to find external measures. For example, objective measures such as year of construction of a building, or number of houses on a street might be helpful. For the purpose of developing a taxonomy of psychological responses to environments, other external measures such as ratings concerning the importance of certain design characteristics, or other descriptive or evaluative reactions seem to provide the best information (cf. Berlyne, 1976). The purpose of the present study is two-fold. First, it aims at the demonstration of the above discussed techniques, and it will evaluate their potential for the study of perception and assessments of environments. Secondly, it attempts to point to those aspects of building entrances that are predominant in peopleâs reactions to them. Entrances always play an enormous part in the design of buildings. They are extremely important in terms of their function and provide for the transition between the inside and the outside. However, the way its function is expressed, that is the physical design, may well determine peopleâs psychological reactions. Hence, the understanding of the relation between structural and technical characteristics and psychological reactions is the major contribution psychologists can make in terms of environmental decision-making. The present study hopes to make one step in this direction by studying building entrances. Stimulus Material Hierarchical cluster analysis (see Johnson, 1967) of a random set of slides of public building entrances taken in Metropolitan Toronto revealed three major clusters: entrances to traditional buildings, and entrances to modern buildings of larger scale, and of smaller scale designs. The three large clusters divided into twelve smaller âbranchesâ. One stimulus was chosen to represent each of these categories. Only entrances of libraries, schools, office buildings, and medical centres were included (see Table 1). All slides were taken from the point at which a person would turn from the public sidewalk to approach the entrance. Subjects Subjects were recruited at the University of Toronto from students taking an introductory psychology course (experiments 1 and 4), or taking a third-year architecture course (experiment 3 ) , and at the Ontario Science Centre, Don Mills, Ontario, from adult visitors (experiment 2). Approximately an equal number of males and females were employed in the experiments. Except for those in experiment 3 none of the subjects had expert knowledge of architecture. .\. OOSTESDORP, S. IlrJl.\STER. M. ROSEN. and P. IV.\ISD Experiment 1. Pairwise Similarity-Dissimilarity Ratings I 1 Procedure Eight subjects were individually presented with the 66 possible pairs of the 12 stimuli. A different random order of the pairs and position of stimuli within pairs was generated by computer for each subject. The stimulus pairs were projected side by side onto a screen, 1.5 metres in front of the subject. The image of each slide was 36 x 54 cm. Two Kodak Carousel (Model 750H) projectors were loaded manually and each pair was presented for 5 seconds. The subjects were instructed to examine the paired stimuli for as long as they were on the screen and then to rate each pair on a rating scale with numbers 1 to 7. I t was explained to the subject that 1 represented extreme similarity and 7 extreme dissimilarity between the stimuli. The subject was given three practice trials using the last three pairs of his random sequence. Then the subject went through all 66 pairs without any interruption. The interval between successive pairs was approximately 5 seconds. Results Analysis of variance of the similarity judgements for all 66 pairs showed a significant main effect for Pairs ( F = 3.06, with d f = 65,455 andp < 0.001). Thus, subjects demonstrated sufficient agreement in their judgements of similarity between the members of the pairs. The data were then subjected to an INDSCAL analysis. The percentage of variance accounted for by the different dimensional solutions increased from 23.7 per cent for the one dimensional solution, to 53.8 per cent for the four dimensional solution. However, since the three dimensional solution accounted for 51 per cent of the variance, little information was gained in the larger solution. In the three dimensional solution dimension 1 accounted for 18.9 per cent, 2 for 23.6 per cent, and 3 for 8.6 per cent. Thus, given these subjects and this sample of building entrances, dimension 2 was the most salient and 3 the least important dimension. The coordinates of the 12 stimuli on the three dimensions are shown in Table 1 . Following Shepardâs ( 1972) suggestion, the stimuli that were located closely together along a dimension were inspected. Some hypotheses concerning the attributes determining the similarity judgements could be formed. All buildings with positive loadings on dimension 1 had stairs towards the entrance. The attributes of the two traditional buildings figure most prominently in dimension 2. In the third dimension an extremely colourful entrance with TABLE 1 Stimulus co-ordinates on INDSCAL dimensions and estimated factor scores 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. Stimuli St CIair Ave, Office, 1975 Albert Campbell Lihrary, 1971 Toronto French School, 1971 Yorkview Medical Building, 1975 Kensington School, 1973 Leslie Medical Centre, 1971 SEF School, 1970 Massey College, 1963 Yorkville Public Library, 1907 Robaru Library, 1974 University College, 1856 Charles Sr Office, 1966 I 0.14 -0.35 -0.11 -0.14 -0.15 0 .09 -0.28 -0 -04 0 .42 0 .52 -0.29 -0.10 Dimemiom 2 -0.21 0 .oo -0.11 -0.19 -0.02 -0.23 -0.14 0.12 0 .58 -0.15 0 .62 -0.27 3 -0.07 0.51 - 0.34 -0.23 -0.54 0.29 -0.20 0 .24 -0.02 0 .29 -0.02 0 .09 I 0 .83 0 .57 -0.17 -0.98 -0.80 0 .64 -0.95 0 .29 -0.17 0 .63 0 .03 0 .08 Fnrtors 2 0.07 -0 6 7 -1.19 0 6 0 -0.24 -0.18 0 .20 0.73 0 .I6 0.57 -0 .oo - 0.05 3 0 .32 -0.93 -0 .I9 -0.31 -0.74 0 .55 0 .03 -0.18 0 .92 -0.53 0 .65 0 .41 B I ? TAXONOMY OF RESPONSES TO T H E ENVIRONMENT âoutspokenâ design features was contrasted with a colourless entrance that appeared to have no outstanding characteristics. In the following experiments some of the above-mentioned hypotheses concerning prominent attributes of the dimensions were tested in conjunction with other possible structural features and with descriptive and affective responses. Experiment 2. Ratings Associated with Architectural Analysis In this experiment the stimuli were rated on five scales. The scales were chosen to represent those collective aspects of a building entrance that receive the greatest attention of the designer o f a building. The architect has to decide whether he wants to give an entrance a monumental, symbolic, or human character. He also has to concern himself with the functional aspects, and with how accessible he wants the entrance to look. I t would be ofinterest to know whether the subjects in experiment 1 might have used coding procedures similar to those utilized by architects. Procedure Tu.el\.e subjects rated each stimulus on each of the characteristics described above. Each attribute was presented with a seven point rating scale ranging from N O T AT ALL to VERY MTCH. Each of the subjects saw the stimuli in a different order, corresponding to a randomly selectcd I ~ V of a 12 x 12 balanced Williamâs square. The subject rated a stimulus on all five scales beiore he would turn to the next. A different randomly selected permutation of the five scales was associated with each temporal position. The projection conditions were as those described for experiment 1 with the exception that only one stimulus at a time was presented on the screen. Furthermore, each stimulus was left on the screen until the subject had made all ratings for that stimulus. Results The F-values of the main effect for Entrances are presented in parentheses in Table 2 for each of the scales. With 11 and 12 1 degrees offreedom each ofthese values show significant inter-subject consistency. The mean ratings of the stimuli on these scales can be probed for their correlations with the INDSCAL dimensions. In other words, the stimulus ratings are treated as locations along a single variable, and these values are compared with the stimulus co-ordinates along the INDSCAL dimensions. I t might be observed in Table 2 that only the ratings on the MONUMENTAL and SYMBOLIC scales are significantly correlated with a dimension. So, the location of the building entrances on the MONUMENTAL scale was similar to their TABLE 2 Experiment 2. F-values for scales and Product-Moment correlations with INDSCAL dimensions Dimemiom 7 Scales F - d u e s I 3 1. ACCESSIBLE (8.481) -0.43 -0.10 -0.05 2. HUMAN (3.197) -0.28 0.30 0.1 7 3. MONUMENTAL ( 1 1.02f) 0.59* 0.38 0.55 4. FUNCTION.4L (3.99t) -0.56 -0.16 0.09 5. SYMBOLIC (7.51 $) 0.34 0.57 0.67* A. OOSTENDORP, S. MchlASTER, M. ROSEN, and P. \4'.4IND 13 position along the dimension that had stairs as an important feature. And, the ratings on the SYMBOLIC scale were related to the dimension that seemed characterized by 'outspoken' design. I t also seems that when perceiving building entrances, a t least on slides, FUNCTIONAL, HUMAN, and ACCESSIBLE aspects do not figure prominently in the internal coding processes. Experiment 3. Ratings of Technical and Structural Aspects The purpose of this experiment is to identify which physical properties were important in the perception of building entrances. Since these scales were dealing more with technical and structural characteristics of buildings, architecture students were solicited to do the ratings. Procedure The procedure was essentially the same as that used for experiment 2, but in this case only eight subjects were assigned randomly to a row of a 12 x 12 balanced William's square. The subjects were presented with 14 design characteristics (see Table 3 ) , and were required to judge how important each of these aspects was according to their perception of the building entrance. As in experiment 2, each aspect was presented in conjunction with a seven point scale, which ranged in this case from NOT A?' ALL IMPORTANT to EXTREMELY IMPORTANT. The scales were partly chosen after those used by Oostendorp and Berlyne (1976), and partly on the basis of aspects discussed in architectural literature concerning building design. Results Analysis ofvariance showed a significant effect of Entrances for 12 out of the 14 scales (see Table 3 for F-values). Since the means for all building entrances on the two non-significant scales were located closely together, correlations with the spatial distribution of the stimuli along the INDSCAL dimensions are not meaningful. Therefore, these are omitted in Table 3. It may be noted in Table 3 that six ofthe scales have a significant correlation with one ofthe dimensions. These ratings are the most important in terms of the purpose of this experiment. TABLE 3 Experiment 3. F-values f o r scales and Product-Moment corselations with IRDSCAL dimensions & a h 1. HORIZONTAL DIMENSION 2. VERTICAL DIMENSION 3. STRAIGHT LINES 4. ANGLES 5. CURVES 6. SHAPES 7. MASSING 8. T E X T U R E 9. COLOURS 10. LIGHT 11 .. ORX.Al1ENTATION 12. BALASCE 13. SCALE 14. R E P E T l T l O S * p 14 TAXONOMY O F RESPONSES TO T H E ENVIRONMENT Dimension 1 was significantly correlated with the aspects of VERTICAL DIMENSION and REPETITION. It may be suggested that the presence or absence ofstairs is related to these two characteristics. The most outstanding physical properties of the second dimension are CURVES and ORNAMENTATION. The traditional buildings clearly demonstrated more of these aspects. STRAIGHT LINES and SHAPES are the features that might partly explain the third categorization process. STRAIGHT LINES and SHAPES may be related to the degree of strikingness and interestingness of a building entrance. The latter hypotheses will be tested in the following experiment. Experiment 4. Descriptive and Affective Ratings In this experiment non-expert subjects were asked to give evaluative responses to the slides of the entrances. For this purpose Osgoodâs (see Osgood, Suci, and Tannenbaum, 1957) semantic differential technique was used. Thus, bipolar adjectives were presented on each side ofa seven point continuum. Following Berlyne (1972) three subclasses of scales were used. Some of them are descriptive, i.e. ORDERLY-DISORDERLY and CRAMPED-SPACIOUS, others are evaluative, HOSTILE-FRIENDLY and FORMAL-INFORMAL. The third category requires subjects to judge their internal state while perceiving the stimulus, i.e. whether they feel RELAXED or TENSE, APPREHENSIVE or CONFIDENT. TABLE 4 Experiment 4. F-calues f o r scales and Product-Moment correlations with INDSCAL dimensions Scales 1. STR.4IGHTFOR\\âARD-COSF~SlSG 2. ISFORMAL-FORMAL 3. COMhfOIGPLACF.-UNCSL;AL 4. HARMONIOUS-DISSONANT 5. CRAMPED-SPACIOUS 6. ORDERLY-DISORDERLY 7. RUGGED-DELlC.4TE 8. D4RK-BRIGH.1 9. A B R U P T G R A D U A L 10. h ? T R . 4 C T I V E - U N A â I T R h C ~ l ~ E 11. HOSTILE-FRIEND1.Y 12. VIVID-DRAB 13. \\âELC:OMING-URâ\CF.LCOMlNC 14. USASSUMI~Ci-PRETENTIOUS 15. CNINTERESTING-1NTERES~âIUG 16. UGLY-BEAUTIFUL 17. POWERFUL-WEAK 18. CSDICNIFIED-DIGNIFIED 19. *RELAXED-TENSE 20. âNO DISCOMFORT-EXTREME DISCOMFORT 21. *NO EXPECTATION-EXTREME EXPECTATION 22. *KO PLEASURE-EXTREME PLEASLâRE 23. â5AFE-UNSAFE 24. *.~PPREHENSIVEâ¬ONFIDENT 35. *DRO\SSY-ALERT Factor 1 (Arousal) I 0.42 0.67â 0.43 0.02 0.52 - __ -0.32 - 0.00 - 0.33 - 0.25 -0.37 0-1 5 0.56 0.43 0.36 - 0 6 3 t 0.46 - - - 0.26 0.3 1 - 0.43 0.47 Dimensions 2 - 0.26 0.29 0.18 - 0.20 -0.12 ~ - 0.1 7 0.28 0.02 0.08 0.19 -0.01 0.03 0.12 0.09 -0.13 0.54 - - - 0.1 1 0.12 -0.39 - 0.06 - 3 0.53 0.43 0,761 0.07 0.32 - - - 0.08 - 11-03 - 0.4 1 -0.19 - 0.677 - 0.09 0.8 I $ 0.72: 0.27 - 0.:9+ 0.43 -+ - ~ - 0.33 0-57 __ 0 4 8 0.75t Factor 2 (Repulsion) 0.34 0.0; 0.16 0.30 0.40 0.00 NOTE. The * in front of the scales indicates that they are dealing with the subjectâs internal state. Factor 3 (Elegance) t p .I\. OOSTENDORP, S. McMâGTER, M. ROSEN, and P. MâAIND 15 Procedure Twelve subjects rated each stimulus on 25 bipolar scales (see Table 4) in essentially the same way as that described for experiment 2. Results Analysis of variance showed the main effect of Entrances to be significant for 19 out of the 25 scales (see Table 4 for F-values). The correlations between these 19 scales and the INDSCAL dimensions are presented in Table 4. Two scales correlated significantly with the first dimension and five with the third dimension. The location of building entrances along dimension 1 was similar to their ratings on the powerful and formal scales. The attributes unusual, vivid, pretentious, interesting, and powerful were found to be related to dimension 3. These findings seem to confirm the hypotheses concerning important attributes of the dimensions. One might also study the interrelations among the rating scales and determine what the underlying variables of these verbal judgements are. So, one could also test whether such underlying variables are related to the dimensions found in experiment 1. For that purpose the correlation matrix of the 19 significant scales was sub,jected to factor analysis. The results of the principal component analysis followed by varimax rotation are presented in Table 5 . T.4BLE 5 Experiment 4. Descriptive and affective scales factor loadings and cornrnunalities 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. Scales STRAIGHTFORWARD-CONFLISING INFORMAL-FORMAL COMMONPL.4CE-USUSUAL HARMONIOUS-DISSONMT CRAMPED-SPACIOUS DARK-BRIGHT ABRUPTCRADUAL ATTRACTIVE-CNATTRACTIVE HOSTILE-FRIENDLY VIVID-DRAB \ 16 TAXONOMY OF RESPONSES TO THE ENVIRONMENT generally been found to be related to Osgoodâs Potency factor (see Berlyne, 1974). The mean factor scores of the stimuli (see Table 1 ) on Factor 1 were found to correlate significantly with INDSCAL dimension 3 (see Table 4). The second factor correlated highly with the HOSTILE-FRIENDLY and WELCOMING-UNWELCOMING scales. A considerable loading was also found for the SAFE-UNSAFE and HARMONIOUS-DISSONANT scales. This factor seems to be determined by evaluative judgements that have an underlying variable in common which might best be labelled as Repulsion. This interpretation seems to be strengthened by the finding that Massey College, which has a wired gate barring the entrance, has the highest estimated factor score (see Table 1) on this factor. The Repulsion factor seems to be the counterpart of Canterâs (1969) Friendliness factor. The interpretation of the third factor is somewhat more difficult. But, it seems that the label Elegance expresses the gradualness of approach as well as the formal, dignified, and beautiful character which were all found to be important attributes of this factor (see Table 5). I t can be noted in Table 4 that neither the Repulsion nor the Elegance factor are significantly correlated with any of the INDSCAL dimensions. Discussion The findings in this project seem to indicate that the demonstrated approach has great potential for a classificatory scheme ofbuilt environments and reactions to them. Although in the present study only a taxonomy of a small sample ofbuilding entrances was produced, it must be obvious that other parts of the environment could be studied in similar ways (see also Garling, 1973; Oostendorp and Berlyne, 1978). The individual differences multidimensional scaling model employed in conjunction with various types of rating scales seems to be extremely helpful in suggesting which physical features may be most prominent in our psychological reactions to the built environment. A tentative description of three dominant dimensions in the perception of and reactions to building entrances may be given. Dimension 1 seems to be related to the impression of Monumentality of a building entrance. The dimension is characterized by features such as stairs, repetition of elements, and emphasis on verticality. Judgements of powerfulness and formality were related to this dimension. A similar dimension ofMonumentality was found in the study by Oostendorp and Berlyne (1978). The second dimension, which appeared to be most prominent in this sample, may be determined by the degree of Detail expressed in the design. Curves and ornaments (as in the more traditionally designed buildings) appeared to be the most salient features. No significant correlations were found with the descriptive and affective scales. Dimension 3, although the least prominent for this set ofstimuli and these subjects, seems to be interpreted most readily. The most outstanding design features were straight lines and shapes. Building entrances for which these features were judged to be important were also rated as being more interesting, pretentious, vivid, and unusual. This dimension clearly points to the psychological impact of a stimulus condition and is very similar to the factor which had been labelled Arousal. Only one of the factors, revealed by the factor analysis of the semantic material in experiment 4, correlated significantly with one of the INDSCAL dimensions. This confirms the suggestion that too heavy reliance on a more or less arbitrary chosen battery of scales is undesirable. Those properties of the built environment that are most dominant in peopleâs psychological reactions may not be revealed. In conclusion, the results seem to point to a number ofimportant relations between design characteristics of building entrances and peopleâs reactions to them, thus providing a beginning for a taxonomy of responses to the built environment. However, since only a small sample of stimuli and population groups were used, the reader should be warned against making generalizations from the present study. The discovered dimensions need to be confirmed by A. OOSTENDORP, S. McMASTER, M. ROSEN, and P. MâAIND further experimentation. Nevertheless, the present project clearly demonstrates that careful psychological research has the potential of making contributions to the understanding of man-environment relations. 17 REFERENCES ACKING, C. A., and SORTE, G. J. (1972). Dynamisk uisuell arkitektur-perception (Lund: Teknisk Hogskolan). BECHTEL, R. B. (1975). âThe semantic differential and other paper-and-pencil testsâ, in Michelson, W. (ed.), Behauio7ol Research Methodr in Enuironmenta/ Design (Stroudsburg: Dowden, Hutchinson &.Ross). BERLYNE, D. E. (1971). Aesthetics and Psychobiology (New York Appleton-Century-Crofts). - (1972). âEnds and means of experimental aestheticsâ, Canadian Jo~l-nal of Psychology, 26, 303-25. ~ (1974). âConcluding observationsâ, in Berlyne, D. E. (ed.), Studies in the N e w Experimental Aesthetics (Washington: - (1976). âThe new experimental aesthetics and the problem ofclassiflring works ofartâ, Scientific Aesthetics, I, 85-106 BLASDEL, H. G . (1973). Identifying the Perceived Attributes of the Designed Eniironment Câtilizing Multidimensional Scaling: An CANTER, D. (1969). âAn intergroup comparison of connotative dimensions in architectureâ, Enuironment andBehauior, 1 , CARROLL, J. D., and CHANG, J. J. (1970). âAnalysis ofindividual differences in multidimensional scaling via an N-way CRAIK, K. H. (1973). âEnvironmental psychologyâ, Annual Re?iez of Psychology, 24, 403-22. EKAMBI-SCHMIDT, J. (1972). La Perception dc lâhabitat (Paris: Editions Universitaires). FRAXKE, J., and BORTZ, J. (1972). âBeitrage zur Anwendung der Psychologie auf den Stadtebau 1â, Zeitxrhriit fur expel-imentelle und angewandte Psychologie, 19, 76-1 08. GXRLING, T. (1973). âSome applications of multidimensional scaling methods to the structural analysis of environmental perception and cognitionâ, in Kiiller, R. (ed.), hchitertural Psychology (Lund: Studentenliteratur). HEISE, D. (1969). âSome methodological issues in semantic differential researchâ, Psychological Bulletin, 72, 406-22. HERSHBERGER, R. C. (1972). âToward a set ofsemantic scales to measure the meaning of architectural environmentsâ, in Mitchell, W. (ed.), Proceedings of the Third Annual Environmental Design Research Association Confoenre (Los Angeles: University of California). Hemisphere). Architectural Eualuafion (Berkeley: Univesity of California). 3 7 4 8 . generalization of âEckart-Youngâ decompositionâ, Psychomet?ika, 35, 283-3 19. JOHNSON, S. C. (1967). âHierarchical clustering schemesâ, Psychometrika, 32, 241 -54. KAYE, s. M. (1975). âPsychology in relation to designâ, Canadian Psychological Reuiew, 16, 104-10. K ~ L L E R , R. (1972). A Semantic Model f o r drsmâbing Perceived Enuironment (Stockholm: State Institute for Building LOMâENTHAL, D., and RIEL, M. (1972). âThe nature ofperceived and imagined environmentsâ, Enuz~unment andBehavior, MIRON, M. (1972). âUniversal semantic differential shell gameâ, Journal of Personalip and Social Psychology, 24, 313-20. OOSTENDORP, A., and BERLYNE, D. E. (1976). âDimensions in the perception of architectureâ I. Identification and OSGOOD, C. E., SUCI, G. J., and TANNENBAUM, P. H. (1957). The Measurement of Meaning (Urbana: University of SHEPARD, R. N. (1972). âIntroduction to volume 1â, in Shepard, R. N., Romney, A. K., and Nerlove, S. B. ieds.), VIELHAUER KASMAR, J. (1970). âThe development of a usable lexicon of environmental descriptionsâ, Environment and Research). 4, 189-207. interpretation of dimensions of similarity. SCQndinQViQn Journal of Psychology, 19, in press. Illinois Press). Multidimensional Scaling. Volume I, Thew? (New York: Seminar Press). Behavior, 2, 15369. NOTE 1. This investigation was supported by research grant S74-1650 to D. E. Berlyne from the Canada Council. The authors wish to express deep gratitude to the late Dr D. E. Berlyne for his support and encouragement. They also thank Mn Phyllis Mayzda of the Ontario Science Centre for her helpfulness. POUR UNE TAXONOMIE DES REPONSES A LâENVIRONNEMENT BAT1 Quatre exptriences ont utilist des photographies de lâentrte de bztiments publics. Leur objet Ctait, respectivement, (1) Iâttude des evaluations par paires de similarite-dissimilarite, (2) les Cvaluations assocites B lâanalyse architecturale, (3) les tvaluations des aspects techniques et 18 TAXONOMY OF RESPONSES TO T H E ENVIRONMENT structuraux, (4) et 1âCtude des evaluations descriptives et affectives. Une analyse multi- dimensionnelle des evaluations de similaritt:-dissimilaritC a permis de dtgager trois dimensions dominantes. Les correlations entre ces trois dimensions et les jugements verbaux recueillis dans les autres experiences permettent dâinterpreter ces dimensions. La premikre correspond a lâimpression de monumentaliti, la caracteristique la plus nette etant lâabsence ou la presence dâescaliers. La seconde dimension est determinee par le taux de dktazls, combes et ornernents. La troisieme dimension, nommee arousal semble caractkriste par lâeffet psychologique, lâkveil de motivation ou dâenergie quâexercent les caracteristiques architecturales sur lâindividu. Ces rksultats montrent lâutilitk de cette mCthode pour klaborer une taxonomie des rkponses a lâenvironnemen t biti.
Comments
Report "Towards a taxonomy of responses to the built environment"