was slight in the final tour. Four tours seemed to be enough to prevent performance impairment after a 1-month forgetting interval Reducing visual access to the.
Journal of Applied Psychology 1983, Vol 68, No 1, 177-186
Copyright 1983 by the American Psychological Association, lnc 0021-9010/83/6801-0177S00 75
Orientation in Buildings: Effects of Familiarity, Visual Access, and Orientation Aids Tommy Garhng, Erik Lmdberg, and Timo Mantyla University of Umea, Umea, Sweden
Spatial orientation when walking through a large building complex was investigated Familiarity with the building was found to facilitate orientation, as revealed by increased accuracy of estimates of targets made by subjects (64 undergraduates) after each tour through the building Subjects who had visited the building frequently before taking part m the experiment improved faster but their superiority was slight in thefinaltour. Four tours seemed to be enough to prevent performance impairment after a 1-month forgetting interval Reducing visual access to the building by restricting subjects' sight when they walked through it resulted in less improvement However, this effect was attenuated if before each tour subjects were shown a floor plan displaying the pathways m the building The method adopted may be useful in field evaluations of building complexes with regard to their ease of orientation and in establishing guidelines for how orientation should be facilitated
How accurate people are in remembering the spatial layout of cities and towns has been thoroughly investigated in recent years (e.g., Appleyard, 1969, 1970; Byrne, 1979; Canter & Tagg, 1975; Garhng, Book, & Ergezen, 1982; Golledge, Briggs, & Demko, 1969; Tversky, 1981; see Evans, 1980, and Garhng, Note 1, for reviews), but so far only a limited number of studies have been concerned with the related question of how accurately people manage to maintain orientation when moving (e.g., Book & Garhng, 1980a, 1980b, 1981a, 1981b; Book, Note 2) This type of knowledge is required to achieve the important practical goal of facilitating orientation in, for instance, public building complexes
(Weissman, 1981; Zimnng, 1981, Garhng, Note 1). Moreover, methods suitable for postoccupancy evaluations of buildings in this respect need to be developed (Friedman, Zimnng, & Zube, 1978). The aim of the present study is twofold. First, whereas our previous studies (Book & Garling, 1980a, 1980b, 1981a, 1981b; see Book, Note 2, for review) were concerned with how orientation is maintained during locomotion in general, the present study investigates factors that may facilitate orientation in buildings The second aim is to devise a method that can be adopted for field evaluations of buildings with regard to their eass of orientation. The method is a straightforward extension of the procedures developed m previous research (Lindberg & GarThis study wasfinanciallysupported by a grant from hng, 1981a, 1981b, 1982, Note 3) requiring the Swedish Council for Building Research to Tommy subjects to estimate the locations of targets Garhng The authors thank Ann Westerman, the Na- in environments that they repeatedly trational Swedish Board of Public Building, Stockholm, for her valuable comments in discussions about the choice verse. Not much appears to be known about of building and related matters of importance for the conception of the present research The Swedish au- what factors facilitate orientation in buildthorities are now through different agencies sponsoring ings, although common sense suggests a work which is expected to result in guidelines for how orientation should be facilitated in public buildings A number of possibilities. Familiarity with a final report of the project is available in Swedish (Gar- building is almost certainly a factor, but to hng, Note 6) and may be ordered from Svensk Byggtjanst, what degree orientation is facilitated and how Box 7853, S-103 99 Stockholm, Sweden important this factor is relative to other facRequests for reprints should be sent to Tommy Gar- tors are questions that need to be answered. hng, Environmental Psychology Research Unit, Department of Psychology, University of Umea, S-901 87 The present investigation, therefore, evaluated (mentation performance after each of Umea, Sweden 177
178
T GARLING, E LINDBERG, AND T. MANTYLA
several tours on which subjects were taken through the building that was under study. Furthermore, a group of subjects who had previously visited the building frequently during the course of their daily activities was also studied. Familiarity is of course of less importance if its facilitating effect does not extend in time. Public buildings are not likely to be visited frequently in a limited period of time Even if subjects in an experiment are found to improve their orientation performance very fast, this result is not easily generalized unless it is also shown that the improvement is retained for some nontrivial time span. No studies of forgetting of information about the spatial layout of environments pertinent to this question appear to have been done except for the study by Garhng, Book, Lindberg, and Nilsson (1981) In that study the results indicated that a 1-week forgetting interval had no appreciable effects on relearning tests whether they were administered after that acquisition had reached its asymptotic level or were interpolated during acquisition. A 1-week interval might, however, have been too short for effects of forgetting to show up The subjects participating in the study probably did not expect the relearning tests (of which they were not told in advance), but during such a short interval they might nevertheless have been engaged in rehearsal that effectively counteracted forgetting. Taking part in the experiment is likely to have been a unique experience to most subjects; thus they probably thought of it several times afterwards and in effect were engaged in rehearsal-like activities. The present study increased the forgetting interval to 1 month, which made it very unlikely that subjects would rehearse for more than a minor portion of the interval, presumably no more than part of the first week after the initial learning session. Orientation m a building is likely to be much easier if every part of the building can be seen from every other part, as may be the case in some airports or railway stations. The question posed here is whether a lesser degree of visual access facilitates orientation For instance, do windows that allow one to see from certain parts of the building to other parts facilitate learning of the building layout
and thereby orientation? Likewise, if one can see across hallways to other parts of the building, is learning and orientation facilitated7 Degree of visual access may not be the only factor, however Evans, Fellows, Zorn, and Doty (1980) found that color coding enhances learning Thus, even if visual access is necessary in order to perceive the colors, the effect of color coding suggests that recognition of the different parts is another necessary condition. That visual access facilitates learning and orientation needs to be demonstrated further, however The present study compares subjects with free viewing to subjects who walked through the building with their sight artificially restricted. The latter subjects were unable to see more than a few meters in front of them and to the sides; thus their visual access to the building layout was poor Orientation aids are not properly properties of a building as such, but in most cases they are probably indispensable as conveyors of information needed for orientation Whether the often-used floor plans are of any value as orientation aids may, however, be questioned. A floor plan presumably has several different functions. For one thing, it can be used to find the shortest path to any part of a building. The mam assumption behind floor plans (and similar devices such as scale models and three-dimensional drawings) is that they convey information about the layout of a building that cannot be acquired unless it is repeatedly traversed It is assumed that knowledge is provided byfloorplans that is possible to use in aflexiblemanner. A floor plan, however, must be read correctly and translated to the building and, since the plan is usually stationary, the information obtained must be remembered. Thus there appear to be several reasons why floor plans are less effective than they are intended to be. If floor plans are satisfactory in their primary function, they should facilitate improvement of orientation in a building. They should furthermore do so when it is most needed, that is, when the visual access to the building layout is poor. These implications were investigated in the present study. One group of subjects whose sight was artificially restricted when walking through the building were presented a floor plan before each tour.
179
ORIENTATION IN BUILDINGS
The performance of these subjects is com- buildings with regard to then" ease of orienpared to the performance of subjects with tation restricted sight who were not presented a floor plan. Method The objective of the present study was to investigate to what degree familiarity facili- Building tates orientation in a building, if the familThe building investigated is a large two-story complex iarity effect is enhanced by good visual access houses the humanities at the University of Umea, to the building layout, and, if the visual access that UmeS, Sweden When walking through the building, is poor, whether a floor plan providing an large windows and open hallways allow one to see outside overview of the layout may have an effect as well as from one part of the building to another part Only the main entrance floor (the second floor), decomparable to that of good visual access. The diagrammaticaly in Figure 1, was investigated method used to study these questions was picted The locations of eight targets (also shown in Figure 1) devised with the intention that it can be were to be learned None of the targets could be seen adopted in postoccupancy evaluations of from any other target or from the main entrance hall
I
I YARD
H
Rows UINHMS
A
TARGET LOCATIONS
——• PATH • *
STARTING POINT REFERENCE LOCATION
ion
GEOMETRICAL DESIGNS DEMOTING TARGET LOCATIONS
O OOO
oooo oooo oooo
m
Figure 1 Main entrance (second) floor of the building complex investigated. (The geometric designs denoting the target locations are shown below, and their locations in the building are indicated m the drawing)
180
T GARLING, E LINDBERG, AND T MANTYLA
where the subjects took the tests The hallway to the right was never used except for practice trials with two visible targets placed on the left wall In the entrance hall, a table and chair were placed so that the subjects could be seated there A visible black design on the wall 30 m straight in front of the subjects served as a reference location in the test procedure, a white design placed on the table indicated to the subjects the exact location of the starting point (also part of the test procedure)
Subjects The subjects, four groups with 6 men and 10 women each, were all undergraduates from the University of Umea taking part in return for payment Subjects recruited to one of the groups were students in the humanities who had frequently visited the building dunng the course of their daily activities for at least 6 months prior to the study, social sciences students, minimally familiar with the building, were randomly assigned to the remaining three groups
Procedure The subjects participated individually in the experiment for about 2 hours The sessions consisted of four guided tours through the building After each tour the subjects first took the orientation tests and then rated their impression of the building on four rating scales (pleasantness, enclosedness, coherence, and complexity) The tours through the building started with the experimenter (Timo Mantyla) meeting the subject at the main entrance hall The experiment was presented as a study of how much one learns about the spatial layout of a building that is repeatedly traversed and how one's impression of it changes The orientation tests were also thoroughly explained and the subject was requested to carry out two practice trials with the visible targets m the hallway to the right He or she received a set of copies of the geometric designs that were used to denote the targets (see Figure 1) and was asked to inspect them for a minute and to memorize each The subject was also informed about the descriptive ratings in advance The intention was to induce the subject to attend properly to the building and not only to the target locations The experimenter accompanied the subject through the building according to a prescribed path that passed each target once The subject was stopped for 10 sec at each target location dunng which time he or she inspected the geometric design placed on the wall to identify that particular location The experimenter walked at a moderately fast pace that he tried to adjust to each individual subject He avoided interfering in any way with the subject's attempt to learn the locations or, conversely, giving undue help Four different paths were used equally often across subjects in each group Moreover, the paths were chosen so that across subjects each location was passed equally often as the first or second, as the third or fourth, and so on The geometric designs appeared across subjects equally often in each location The second, third, and fourth tours were identical to the first except that half of the subjects in each group changed to one of the other paths Before each new tour the subjects were again shown the copies of the geometric
designs denoting the targets The tours took about 10 to 12 minutes depending on which path was traversed Two of the groups of subjects were run exactly as described above, one of them with the subjects familiar with the building (n = 16) and the other with subjects unfamiliar with it (« = 16) The subjects in the latter group were asked back for another identical session from 24 to 35 days later When recruited for the first session, these subjects had to agree to take part in two sessions but they were never told the purpose of the second session In one of the remaining groups (n = 16), the subjects' sight was restricted After the subject had read the instructions and carried out the practice trials, he or she put on a pair of goggles that restricted his or her sight to a 35°-wide sector extending 10 m away on the floor When walking through the building the subject could not see through the windows, across the hallways, or, since he or she walked just a few meters behind the experimenter, much of the hallway in front of him or her Before starting, the subject practiced walking to the targets placed in the hallway to the right of the entrance hall Two practice trials were always required but several were allowed if the experimenter judged it necessary The goggles were taken off each time the subject returned for the tests and ratings, then put on again before the next tour began The subjects m the remaining group (n = 16) also had their sight restricted A floor plan consisting of a scale drawing of the hallways (1 cm to 6 m, see Figure 2) was presented after the subject had read the instructions and earned out the practice tnals, then again before each new tour after he or she had completed the tests and ratings and inspected the geometnc designs denoting the targets The floor plan was inspected by the subject for a minute dunng which time he or she was asked to memorize it
Tests The onentation test reported here consisted of a set of location estimates ' The test was administered as soon as the subject returned to the main entrance hall Completing all the tests and the following desenptive ratings took about 15 minutes Location estimates were obtained of all targets by requmng the subject to mark a cross for each on separate response forms like the one shown in Figure 3 The crosses were to indicate the locations of the targets relative to the direction and distance to the black design on the wall in front of the subject The geometnc design in the upper nght corner of the response form denoted the target to be estimated, the center cross denoted the place where the subject was sitting (the white design on the table indicated the exact location), and the vertical 1 The subjects also estimated the crowflight distances to each target and, in connection with the location estimates and the distance estimates, 90% confidence intervals of direction and distance, respectively These estimates did not, however, yield any information over and above that provided by the location estimates (Garhng, Note 6) that are reported here
ORIENTATION IN BUILDINGS
Figure 2 Information abstracted from a floor plan presented to subjects before each tour (2-cm-long) arrow drawn from the center showed the direction and distance to the black design The subject had to put his or her cross within the outer circle with a 10-cm radius The order in which the targets were estimated was randomized for each subject and tour
181
Two accuracy measures of direction were calculated for each subject and tour (a) The linear regression coefficient (b) by means of least squares fitting a straight line to the plot of the estimated angles of direction (relative to the reference direction) against the corresponding actual angles across all eight targets, and (b) the productmoment correlations (r) between the angles of direction and the actual angles Statistical tests of the deviations from linear regression (Kirk, 1968) indicated that they were negligible The linear regression coefficients therefore express the degree of systematic errors, but since they are deflated if the linear fits are poor due to unsystematic errors, the correlations were calculated and are subjected to parallel analyses Accuracy measures of distance were likewise calculated However, the tests of deviation from linear regression in this case suggested that a log transformation of the variables was appropriate The linear regression coefficients and the correlations were therefore obtained for log-log plots The interpretation of these measures is slightly different The regression coefficients actually express degree of nonhneanty, whereas the correlations express the degree of unsystematic deviations from the nonlinear relationships
Measures The main dependent measure was the accuracy of the location estimates, which was divided into direction and distance accuracy as follows After transformation of the distance scale on the response forms to meters, all distances between the estimated locations and the actual locations were measured for each subject and tour A matrix of estimated locations by actual locations was thus formed in each case Then, because subjects may have confused the geometric designs denoting the target locations and in order not to deflate the accuracy of location measures because of this, the results were preprocessed in the following manner We began with the smallest distance in the mam diagonal (the estimates of the corresponding actual locations if no confusion had been made) If this distance was less than the distances to other actual locations, the target was assumed not to have been confused with any other target If, on the other hand, any of the off-diagonal distances were less than the smallest diagonal distance and also less than any other distance between the corresponding actual location and the remaining estimated locations, a confusion with that target was assumed to have taken place The estimate was accordingly assigned to the target location with which it was assumed to have been confused Assuming that the subjects never estimated the same actual location more than once, this pair was deleted and the steps described above were repeated until all estimates were assigned to one actual location each If ties occurred, which estimate to choose was randomly decided The algonthm was computerized and a random digit generator was used to make decisions about ties The accuracy of the location estimates was obtained by regressing the directions and distances of the estimated locations on the actual directions and distances
Figure 3 Response form used by subjects for estimating the target locations (Note that an example of an estimate of a target location labeled by the geometric design in the upper-nght corner is given in the figure, d and v denote the distance and angle values, respectively, which were subjected to analyses as described in the text)
182
T GARLING, E LINDBERG, AND T. MANTYLA
Table 1 Percent Confusion of Target Locations as a Function of Number of Tours and Conditions Subjects not familiar with the building Subjects familiar with the building Tour 1
2 3 4
Restricted sight
Unrestricted sight (n = 16)
Unrestricted sight (n = 16)
Session 1
Session 2
No floor plan (n = 16)
Floor plan (n = 16)
34 7 6 6
42 23 15 9
16 11 6 5
35 34 28 18
50 29 16 13
Results The algorithm adopted in preprocessing the results revealed that the number of confusions of the target locations (expressed as percentages of the possible number of confusions) decreased across tours, mainly from the first to the second tour (see Table 1). There were also differences between the several conditions paralleling those observed for the accuracy measures. These differences were not expected, and they probably reflect that the algorithm overestimates the number of confusions when subjects are inaccurate in estimating the locations of the targets. The results for the accuracy measures of direction and distance averaged across subjects are given in Table 2 as functions of the number of tours and the conditions.2 They suggest a strong effect of familiarity, since performance improves quickly across tours, but they also suggest that there are differences due to the conditions. Analyses of variance (ANOVAS) including all four groups of subjects substantiate this by yielding significant effects of tours, F(3, 168) = 4.53 and 4.21, p < .01, for b; and F(3, 168) = 23.11, p < .001, and F(3, 168) = 5.74, p < .01, for r transformed to Fisher's z, direction and distance respectively. Significant or nearly significant effects of conditions were also found, F(3, 56) = 6.70, p < .001, and F(3, 56) = 2.28, p < .10, for b; and F(3, 56) = 5.81, p < .01, and F(3, 56) = 2.81,/? < .05, for r. The Tour X Condition interactions did not quite reach significance, however, F(9, 168) < 1.00 and F(9, 168) = 1.58, p < .25, for b, and F{9, 168) < 1.00 and F{9, 168) = 1.41, p< .25, for r. The subjects initially familiar with the
building improved reliably across tours (p < .05 or less). In further support of a familiarity effect, they did so much faster than any other group of subjects. As far as direction accuracy is concerned, they were almost perfectly accurate by the second tour. Distance tends generally to be less accurate than direction, m part probably because the actual distances differed marginally and therefore were difficult to discriminate. The other groups of subjects, who were initially unfamiliar with the building, performed with one exception at a comparable level in thefinaltour, suggesting that they attained an asymptotic level during the course of the four tours. But a difference remains in that these subjects persisted in making larger systematic errors, as indicated by the fact that the linear regression coefficients differ reliably from 1.0. The familiarity effect may extend across a substantial time span since, as Table 2 also shows, there appear to be marginal effects of the 1-month forgetting interval. The subjects asked back for another session actually performed better in that session than they did in the first tours of the first session, ANOVAS with forgetting interval as a within-subjects factor revealed that performance tended to be significantly more accurate in the second sessional, 14) = 31.78,/? < .001, and F(l, 14) = 3.81, p < .10, for b; and F(l, 14) = 23.84 and F(l, 14) = 16.06, p < .001, for r. Because there are reliable improvements 2 Analyses of the results of the ratings are presented in Garling (Note 6) They yielded effects of tours indicating that the subjects did attend to the building as was intended by including the rating procedure
183
ORIENTATION IN BUILDINGS Table 2 Accuracy of Estimates of Target Locations Measured by Mean Regression Coefficients (5) and Correlation Coefficients (f) Converted from Fisher's z as Functions of Number of Tours and Conditions Subjects not familiar with the building Subjects familiar with the building
Unrestricted sight (n- = 16)
Unrestricted sight (« = 16) Tour
5
f
1 2 3 4
86" 95 102 102
89 96 97 .98
Session 1
No floor plan (>i = 16)
Session 2
B
r
64* 70* 78" 78"
79 88 .92 94
•
Restricted sight
B
Floor plan (« = 16)
r
B
f
B
r
93 94 96 96
.75" 69" 78" 77"
84 82 90 93
71" .82" 80* 90"
81 91 .92 95
76 77 78 80
65" 52" 57" 62"
57 48 .56 61
51" 76a 85 a 79a
40 71 76 .78
Direction (degrees) 82" 84" 88" 87 a
Distance (log units) 1 2 3 4
78" 97 92 87"
56 79 79 80
39" 61" 97 82a
46 60 76 67
.89a 78 a 85" 78"
»Differs from 1 0 (p < 05)
across tours only in the first session (p < .05 or less), the main effects of tours were slight, F(3, 42) < 1.00 and F(3, 42) = 5.29, p < .01, for b; and F(3, 42) = 1.15, p < 05, and F(3, 42) = 2.09, p < .25, for r. Visual access to the building appears to be important. When the subjects' sight was restricted, the results suggest that they improved less across tours. Performance did not differ reliably from that of the other groups in thefirsttour but there were, with exception for the unsystematic direction errors, no significant improvements across tours (p = .05) Finally, presenting the floor plan before the tour attenuated the effect of restricting the subjects' sight. Performance of the subjects in this group improved reliably across tours {p < .05 or less), and these subjects improved as fast as did the subjects with no restricted sight. Discussion The hypotheses investigated in the present study were substantiated, although not all hypothesized effects reached statistical significance. Familiarity with the building fa-
cilitated orientation, and the facilitating effect remained after the 1-month forgetting interval. Decreasing visual access to the building by restricting the subjects' sight reduced the effect of familiarity, but presenting the subjects with afloorplan tended to counteract the negative effect of restricting the sight. The effect of decreasing visual access was, however, slight for the accuracy measures, in part probably because the algorithm used in preprocessing the results resulted in overestimates of degree of accuracy. Number of confusions was higher in the conditions in which the subjects' sight was restricted, and, since that was taken into account when pairing the estimates with the actual target locations before performing the regression analyses, the latter analyses probably produced spuriously high estimates of accuracy. The effect of familiarity was found to be strong, as indicated by the fact that performance increased rapidly across tours. However, the algorithm probably overestimated number of confusions in the first tours, and consequently accuracy was overestimated in these tours. Nevertheless, the subjects initially unfamiliar with the building were quite
184
T GARLING, E LINDBERG, AND T MANTYLA
accurate in the final tour and did not differ provement was less rapid for the subjects usmuch from those subjects initially familiar ing a floor plan than for those familiar with with the building. Thus four tours through the building in advance because it takes some the building appear to have had a substantial time to learn the pathways and/or other saeffect on orientation—in fact almost equal lient locations to that of frequent visits to the building over One implication of the finding that rean extended period. Of course, it must not stricted sight impaired improvement can be be implied that there were no differences at pointed out. Previous studies using blindall between these groups of subjects with re- folded or blind subjects (Lindberg & Garling, spect to more general knowledge about the 1981a; Rieser, Lockman, & Pick, 1980; Garbuilding ling, Book, & Lindberg, Note 4, Garling, That improvement was so rapid may be Mantyla, & Saisa, Note 5) found small difexplained by the hypothesis about the process ferences as compared to sighted subjects. of environmental learning proposed by Siegel However, as the present results imply, there and White (1975), which is partly supported may be differences in rate of improvement by the results of a number of recent studies Further studies of this question are needed (Evans, Marrero, & Butler, 1981; Garling et considering the importance of providing al., 1981; Garhng et al., 1982; Herman, Kail, measures that facilitate orientation for the & Siegel, 1979; see also Evans, 1980, for re- blind view). Siegel and White assumed that newApart from extending the familiarity effect comers to an environment first learn a num- in time, the present results m the relearmng ber of salient locations, then learn the paths session suggest that there was almost no forin between, and finally organize the acquired getting. The results of the previous study by knowledge spatially m a system. In the pres- Garling et al. (1981) warranted the same conent experiment, the designated target loca- clusion. An interesting implication of the tions may have become salient locations, hypothesis by Siegel and White (1975), retherefore orientation relative to them im- lated to the views expressed long ago by Gibproved rapidly This may hold even though son (1958, 1979), should be noted here. If the procedure employed should to some ex- visual access is good, it should be possible for tent have directed the subjects' attention away the subjects entering the building, literally from the particular target locations. The sub- speaking, to perceive the building, even those jects who were initially familiar with the parts not immediately in view. This could be building did not perform well m thefirsttour, accomplished by chains of spatial inferences possibly because the target locations did not (e.g., if C can be seen from B and B can be constitute salient locations to them before the seen from A, the location of C should be posexperiment, and therefore they had to be sible to infer when standing at A). If so, there learned in the first tour. However, since those may be no demand to store information in subjects improved faster than all the other memory for any longer than a few seconds. groups of subjects, they may have been ap- We are not suggesting that the subjects do plying knowledge they already possessed. not acquire a long-term memory represenOne possibility then, is that they knew other tation of the spatial layout, since there is an salient locations and/or the path system, and abundance of evidence to substantiate that that they were able to use this m learning the they do (Evans, 1980). However, relearmng target locations. may be facilitated because excellent perforThe facilitating effects of visual access and mance does not require that much infora floor plan may be interpreted in a similar mation is remembered. One prediction that vein. Both visual access and a floor plan (and can be made is that if forgetting had been similar devices such as scale models and measured before the subjects entered the three-dimensional drawings) provide infor- building, much larger effects would have been mation about the pathways and/or other sa- found. lient locations. This information in turn may The present results need to be replicated be remembered and used in learning the tar- in large-scale studies. In particular, whereas get locations. It is understandable that im- the present series of studies (Garling et al.,
185
ORIENTATION IN BUILDINGS
1981; Garling et al., 1982; Garhng, Note 6) have compared different types of environments such as towns, neighborhoods, and building complexes, a comparison between different types of building complexes would be pertinent to the general aim of the present study. The method adopted is potentially useful in efforts to establish guidelines for how orientation can be facilitated in buildings However, if the purpose is valid assessment of general performance level, some changes in the method may be called for. First, in addition to measuring orientation performance, as we have done, wayfinding may need to be studied. It is, of course, our assumption that wayfinding is not possible unless orientation is maintained, but this assumption should be tested (some buildings may, for instance, be designed in a way that makes accurate wayfinding trivial). Second, if orientation is measured, there are several options concerning the selection of the target locations They may, for instance, be self-selected but since this poses serious methodological problems it is probably better to select targets in pilot studies. Salient locations may be selected, but important destinations, such as exits and rest rooms, could also be included. Furthermore, recognition of the parts of a building that constitute target locations is of importance for orientation and should therefore be studied as well. Subjects could, for instance, be presented with photos of the target locations rather than geometric designs used in the present study to denote the target locations. The scoring algorithm we have developed makes it feasible to analyze recognition errors, although it has some shortcomings, as have been pointed out. Third, orientation in a building should be measured in different parts, not only in the mam entrance hall. It may be equally important to remain oriented on one's way back to the exit. Fourth, in any learning study factors such as motivation are important. We explicitly instructed the subjects to attempt to improve their performance. Other possibilities, more similar to real-life conditions, are conceivable and should be explored in a field evaluation. The distraction task technique used by Lindberg and Garhng (198 la, 1981b, 1982) is in this connection another attractive possibility.
Reference Notes 1 Garhng, T Environmental orientation during locomotion Experimental studies of human processing of information about the spatial layout of the environment (Document D24) Stockholm. Swedish Council for Building Research, 1980 2 Book, A Maintenance of environmental orientation during locomotion (Umea Psychological Reports Supplement Series, No 8) Umea, Sweden University of Umea, Department of Psychology, 1981 3 Lindberg, E, & Garhng, T Acquisition of locatwnal information about reference points during locomotion Effects of amount of central information processing. Manuscript submitted for publication, 1982 4 Garling, T, Book, A, & Lmdberg, E Orientation performance in two-segment and three-segment route tasks during blindfolded and sighted walking (Umea Psychological Reports No 94) Umea, Sweden University of Umea, Department of Psychology, 1975 5 Garling, T, Mantyla, T., & Saisa, J The importance of vision during locomotion for the acquisition of an internal representation of the spatial layout of largescale environments Blindfolded and sighted car passengers with and without a distracting task learning to localize invisible targets during a town route (Umea Psychological Reports No 147) Umea, Sweden University of Umea, Department of Psychology, 1978 6 Garling, T Perception av miljoers spatiala organisation underfbrflyttning [Perception of the spatial layout of the environment dunng locomotion] (Report R159) Stockholm Swedish Council for Building Research, 1980
References Appleyard, D Why buildings are known A predictive tool for architects Environment and Behavior, 1969, 1, 131-156 Appleyard, D Styles and methods of structuring a city Environment and Behavior, 1970, 2, 100-117 Book, A , & Garhng, T Processing of information about location dunng locomotion Effects of a concurrent task and locomotion patterns Scandinavian Journal of Psychology, 1980, 21, 185-192 (a) Book, A , & Garhng, T Processing of information about location dunng locomotion Effects of amount of visual information about the locomotor pattern Perceptual and Motor Skills, 1980,57,231-238 (b) Book, A., & Garhng, T Maintenance of onentation during locomotion in unfamiliar environments Journal of Experimental Psychology Human Perception and Performance, 1981, 7, 995-1006 (a) Book, A , & Garling, T Maintenance of environmental onentation dunng body rotation Perceptual and Motor Skills, 1981, 52, 583-589 (b) Byrne, R W Memory for urban geography Quarterly Journal of Experimental Psychology, 1979, 31, 147154 Canter, D , & Tagg, S K Distance estimation in cities. Environment and Behavior, 1975, 7, 59-80. Evans, G W Environmental cognition. Psychological Bulletin, 1980, 88, 259-287.
186
T GARLING, E. LINDBERG, AND T MANTYLA
Evans, G W, Fellows, J , Zorn, M , & Doty, K. Cognitive Lindberg, E , & Garling, T. Acquisition of locational inmapping and architecture Journal of Applied Psyformation about reference points during blindfolded chology, 1980, 65, 474-478 and sighted locomotion' Effects of a concurrent task and locomotion paths Scandinavian Journal of PsyEvans, G W, Marrero, D G , & Butler, P A Environchology, 1981,22, 101-108 (a) mental learning and cognitive mapping Environment and Behavior, 1981, 13, 83-104 Lindberg, E , & Garling, T Acquisition of locational information about reference points during locomotion Friedman, A , Zimnng, C , & Zube, E Environmental with and without a concurrent task Effects of number design evaluation New York Plenum Press, 1978 of reference points Scandinavian Journal of PsycholGarling, T, Book, A, & Ergezen, N Memory for the ogy, 1981,22, 109-115 (b) spatial layout of the everyday physical environment Differential rates of acquisition of different types of Lindberg, E , & Garling, T Acquisition of locational ininformation Scandinavian Journal oj Psychology, formation about reference points during locomotion 1982, 23, 23-35 The role of central information processing Scandinavian Journal of Psychology, 1982, 23, 207-218 Garling, T., Book, A , Lmdberg, E , & Nilsson, T Memory for the spatial layout of the everyday physical en- Rieser, J J , Lockman, J j f & Pick Jr, H L The role vironment Factors affecting rate of acquisition Jourof visual experience in knowledge of spatial layout nal of Environmental Psychology, 1981, 1, 263-277 Perception and Psychophysics, 1980,25, 185-190 Gibson, J J Visually controlled locomotion and visual Siegel, A W, & White, S H The development of spatial orientation in animals British Journal of Psychology, representations of large-scale environments In H W 1958,49, 182-194 Reese (Ed ), Advances in child development and beGibson, J J The ecological approach to visual percep- havior (Vol 10) New York Academic Press, 1975 tion Boston Houghton-Miffiin, 1979 Tversky, B Distortions in memory for maps Cognitive Psychology, 1981, 13, 407-433 Golledge, R G., Bnggs, R, & Demko, D The configuration of distances in mtra-urban space Proceedings Weissman, J Evaluating architectural legibility Wayof the Association ofAmerican Geographers, 1969, 1, finding in the built environment Environment and 60-65 Behavior, 1981, 13, 189-203 Herman, J F, Kail, R V, & Siegel, A W Cognitive Zimnng, C M Stress and the designed environment maps of a college campus A new look at freshman Journal of Social Issues, 1981, 37, 145-171 orientation Bulletin ofthe Psvchonomic Society, 1979, 13, 183-186 Received May 3, 1982 Kirk, R E Experimental design Procedures for the behavioral sciences Belmont, Calif Brooks/Cole, 1968 Revision received August 22, 1982 •
