Journal of Environmental Psychology (1996) 16, 17–31 0272-4944/96/010017+15$18.00/0
© 1996 Academic Press Limited
ENVIRONMENTAL
PSYCHOLOGY
Journalof
DEMOGRAPHIC DIFFERENCES IN THE VISUAL PREFERENCES FOR
AGRARIAN LANDSCAPES IN WESTERN NORWAY
EINAR STRUMSE
Research Center for Health Promotion, University of Bergen, Øistens gt. 3, N-5007 Bergen, Norway
Abstract
This study examines the relationship between a number of demographic variables and the visual preferences
for a sample of colour slides depicting traditional and agrarian landscape scenes from Western Norway among
198 Norwegian students. In order to attain this purpose, eight demographic measures, tapping familiarity,
subculture and expertise, were developed. Serving as criterion variables, sumscores were constructed on the
basis of a nonmetric factorial analysis (SSA-III), identifying seven perceptual dimensions in the visual preferences
for agrarian landscapes. The results suggest areas of unanimity as well as areas of divergence with
respect to the visual perception of agrarian landscapes. Thus, an almost unanimous consensus with respect to
(a) the high preferences for traditional human-influenced settings and nature scenes, and (b) the relative dislike
for dominating human influence and many of the effects of modern farming practices was found. In contrast,
important divergencies were found for landscape categories in the moderate preference range, with the
highest occurrence of group differences for a category of scenes depicting farming activities. The most potent
demographic predictors of preference across landscape categories were present population density, gender,
organization membership and expertise. In addition, effects of age and present geographical region of residence
were observed. The findings support evidence from earlier analyses of the landscapes in question, indicating
strong preferences for traditional agrarian settings.
© 1996 Academic Press Limited
Introduction decisions. Such decisions will probably not be made
without support from both landscape experts and
the general public.Agrarian landscapes are threatened by the continuing
expansion of urban areas, the spreading of There is good reason to expect that support to
landscape preservation varies between segments ofindustrial agriculture and an increasing marginalization
and disuse of old agricultural land. In the population. This variation may, at least in part,
be associated with underlying divergencies in thethis situation, Norway occupies a special position
due to a topography generally not suited for mech- perception of the landscapes in question. The study
of visual preferences represents one way of sub-anized agriculture (Kaland, 1993). This is the case
especially in Western Norway, where peasants jecting environmental perception to systematic
inquiry. Moreover, an examination of the relation-often chose to maintain traditional practices. Here,
traditional agrarian landscapes of exceptional rec- ships between demographic variables and visual
landscape preferences might provide valuable infor-reational and aesthetical value, and an ecological
diversity not present in the natural landscape, can mation for decision-makers in the field of landscape
planning and management. For example, the identi-still be found (Austad et al., 1991). However, with
the last generation of farmers still knowledgeable of fication of similarities in landscape preferences
across groups would assist the development of gen-traditional techniques rapidly disappearing, these
landscapes are also threatened by disuse (Austad, eral guidelines for landscape design, whereas the
demonstration of group differences would help to1993). Thus, their preservation for the future
depends, at least in part, on national policy sort out cases or conditions in which the preferences
17
18 E. Strumse
of specific groups should be mapped. In contrast, group differences in landscape preferences according
to three main themes:the costs of not taking demographic variation into
account might be that well-meant efforts, to either
(1) familiarity, or experience, related to geopreserve
or develop an area, will end up mired in
graphical circumstances of residence and the effect
conflicts between opposing groups (cf. Kaplan &
of direct exposure to an environment;
Kaplan, 1983).
(2) cultural and ethnic variation, including the
An examination of the role of demographics in
question of age and other bases for belonging to a
visual landscape preferences also brings to the sur-
‘subculture’;
face the more general issue of the universality of
(3) the effects of formal knowledge and expertise.
human nature vs cultural variability (see, e.g.
Cosmides et al., 1992). Theoretical explanations of Evidence for the importance of familiarity can be
found in a number of studies. Landscape exposurevariations in visual landscape preferences have
commonly been taken to support either a con- as a child, childhood residence and place of residence
(Zube et al., 1974; Daniel & Boster, 1976;structivist (cf. Moore, 1979; Lyons, 1983) or a functionalist–evolutionary
position (Appleton, 1975; Lyons, 1983; Kellert, 1978) have all been reported
to influence landscape preferences. Moreover, in anKaplan & Kaplan, 1983, 1989). Somewhat simplified,
differences that can be attributed to the back- analysis of perceptual differences between residents
of and visitors to a specific area, Kaplan (1977)grounds of individuals or groups, i.e. to demographic
characteristics, would lend support to the found that long-time residents exhibited greater differentiation
of landscape features. However,constructivist assumption and, consequently, to the
notion of cultural variability in landscape prefer- urban/rural experience seems in most cases not to
influence landscape preferences (Gallagher, 1977;ences. In contrast, differences produced by the landscapes
in question (i.e. in the absence of individual Kaplan, 1977, 1985; Miller, 1984; Keane, 1990).
Generally, the relationship between preference andor group differences) would indicate the appropriateness
of a functionalist–evolutionary interpret- familiarity has been a positive one (cf. Hammitt,
1978, 1987; Keyes, 1984), but negative relationshipsation, thus supporting the notion of cross-cultural,
universal patterns in visual preference. However, have also been found (Penning-Rowsell et al., 1977).
Thus, familiarity does affect preference, but it is notfrom the position of evolutionary psychology, cultural
variability is not a challenge to claims of uni- always clear what the effect will be (Kaplan &
Kaplan, 1989).versality. Rather, the focus of this emerging field is
to identify the psychological mechanisms that come The potential effect of subculture is demonstrated
by findings suggesting that, in some cases, samplesbetween theories of selection pressures on the one
hand and socio-cultural behaviour on the other sharing the same landscape show substantial differences
in preference (Daniel & Boster, 1976;(Cosmides et al., 1992).
In an attempt to overcome the conflict between Kaplan & Herbert, 1987; Porter, 1987). Some
studies have found that preference tend to decreasebiological and cultural explanations as it is
reflected in the field of landscape aesthetics, with increasing age (Balling & Falk, 1982; Lyons,
1983), and vast preference differences have beenBourassa (1990) suggests a tripartite theory, making
a distinction between biological, cultural and found between teens and people involved in teaching
about the environment (Medina, 1983).personal modes of aesthetic experience. A particularly
interesting feature of this contribution is the An effect of gender has repeatedly been found in
studies of environmental perceptions (Kellert, 1978;proposal that natural landscapes should be experienced
primarily through a biological mode, thus Macia, 1979; Lyons, 1983; Dearden, 1984), but this
effect seems far from unambiguous. For example inimplying universal patterns of preference. On the
other hand, urban landscapes (and, we would add, a recent review of the more general research literature
on gender and environmental concern, Stern etother strongly human-influenced landscapes) would
probably be experienced through the cultural mode al. (1993) found that while some studies reported
women to be more concerned about the environ-and thus be subjected to variability. The proposed
framework seems useful because it urges research- ment, others reported the opposite findings. When
differences are found, women seem to be more nega-ers to make explicit the distinctions between the
three modes of aesthetic experience, and to clarify tive towards intrusions into natural environment
than are men (Levine & Langenau, 1979; Kardell &the way in which they apply them to their own
research. Ma˙rd, 1989). There also seems to be some support
for gender differences along evolutionary lines, withKaplan and Kaplan (1989) summarized studies of
19Demography and Preferences for Norwegian Agrarian Landscapes
a tendency for males to prefer more challenging seemed, in large part, to conform with these
findings.environments (Woodcock, 1982; Bernaldez & Abello,
1989) and a greater facility by females to remember The purpose of the present study was to examine
to what extent differences in visual preferences forspatial configurations of objects (Silverman & Eals,
1992). agrarian landscapes in Western Norway can be
explained by demographic factors. In order to attainMembership in environmental groups has been
shown to influence landscape preferences, members this goal, scenes from traditional and modern agrarian
landscapes were sampled. Modern scenesbeing more in favour of wilderness scenes (Dearden,
1984) and showing lower preferences for manipu- were included for two major reasons: (1) to assure a
degree of representativeness in the sample oflated landscapes and higher preferences for nature
scenes (Kaplan & Herbert, 1987). scenes, and (2) to make possible a comparison of the
effects on visual preference of changing agriculturalA distinction between expert evaluations and the
preferences of the general public has commonly techniques (i.e. traditional vs modern). The study
focused on the following questions:been made in studies of visual preference. Concluding
from the results of a number of studies (Kaplan, (1) Will consensus among demographic groups in
terms of visual preferences be found for any given1973; Anderson, 1978; Buhyoff et al., 1978; Medina,
1983), Kaplan and Kaplan (1989) note that experts category of landscape?
(2) Will subcultural variables, such as age, gen-seem to weigh the role of informational aspects of a
given setting differently from lay groups. Further- der and organization membership, influence preferences
for agrarian landscapes in Western Norway?more, these studies also suggest that expert judgements
do not correspond well to the public’s and (3) Will familiarity with types of landscape in
question result in higher preferences for them?that experts tend to be unaware of the difference. In
Scandinavia, Hultman (1983) found that Swedish (4) How does landscape expertise influence the
visual preferences for Western Norwegian agrarianforest managers exhibited preferences close to those
environmentalists, whereas Savolainen and Kello- landscapes?
ma¨ki (1984) found no differences attributable to
expertise.
Clearly, demographic characteristics can be a
Methodsource of variation of environmental preference.
However, it should be kept in mind that consensus,
Subjects were 198 volunteer male (n=72) andrather than divergence, seems to be the rule (cf.
female (n=126) Norwegian university and collegeDearden, 1984). A series of studies support this
students, taken from landscape-related disciplinesargument. First, the content of scenes has consist(n=94)
and introductory courses in psychology (n=ently emerged as a major contributor to preference.
104). The rationale for this composition of theHere, the most preferred scenes have repeatedly
sample was to make it consist of equally largebeen shown to be those in which human influence
groups of landscape ‘experts’ and ‘nonexperts’. Thedoes not dominate the natural elements or where
mean age of participants was 25.1 years (S.D.=5.96,nature dominates, whereas the least preferred
minimum=19 years, maximum=50 years). Thescenes often represent intrusions into the natural
study was carried out in several sessions in Bergen,environment (Gallagher, 1977; Anderson, 1978;
Oslo, Trondheim, Sogndal, and Bøi TelemarkHammitt, 1978; Herbert, 1981; Ellsworth, 1982;
between May and October 1993.Herzog et al., 1982; Hudspeth, 1982; Miller, 1984;
Kaplan, 1985; Strumse, 1994a).
However, not all nature scenes are highly pre- Visual stimuli
ferred. Consequently, similarities in preference cannot
be explained sufficiently on the basis of content Sixty colour slides, drawn from an initial collection
of 76 slides, were used. In so far as this was compat-alone. According to Kaplan and Kaplan (1989),
additional important factors are environmental ible with the composition of a fairly balanced
sample of settings, a major concern was that theattributes enhancing the processes of understanding
and exploration. Also spatial information, indi- settings chosen should represent areas well documented
with respect to earlier research (geographycating how well one could function in the space represented,
seems important. An earlier examination and ecology) as this would add to the usefulness of
the results. A more detailed description of theof preference predictors for the agrarian landscapes
employed in the present study (Strumse, 1994b) sampling procedure can be found in Strumse
20 E. Strumse
(1994a, b). The areas chosen for the sampling of Procedure
scenes were:
(1) traditional agrarian landscapes from various Projected in random order on a screen, 60 slides
intended for the analysis of preference ratings andlocations in Sogn og Fjordane county. Considerable
geographical and botanical ecological data exist on 10 ‘filler slides’, five at the beginning and five at the
end of each sessions in order to avoid start and endsome of these areas (Austad & Kaland, 1991; Austad
et al., 1991); effects, were rated by the participants. The slides
were rated on a scale ranging from 1 to 5 according(2) well-documented modern agrarian landscapes,
mostly from the surroundings of the lake to how much it was liked (1=does not like at all, 2=
likes a little, 3=likes somewhat, 4=likes quite a bitKalandsvatnet (Sandahl, 1989; Lundberg, 1990;
Østerbø, 1990) just outside Bergen, the second and 5=likes very much). Each slide was exposed for
10 seconds with a random interstimulus intervallargest city of Norway, but other areas (Sogndal,
Aurland, Laerdal and Vik) were also included. ranging from 1 to 9 seconds, the latter in order to
avoid response set (see e.g. Polit & Hungler, 1991).These landscapes have undoubtedly lost some of the
values that characterize the traditional landscapes Immediately following the visual preference
ratings, subjects filled in the questionnaire, includ-but are nevertheless important as they provide outdoor
recreation opportunities to the urban ing the items designed for the measurement of
demographic variables.population.
SPSS/PC+ V5.0 was used for descriptive statistics,
one-way analyses of variance and multipleThe questionnaire
classification analyses (MCA).
The visual preference ratings and the demographic
measures investigated in the present study were
parts of a larger questionnaire designed to explore a Results
wide range of potential predictors of visual preferences.
The questionnaire included: An analysis considering the influence of demography
on the preferences for each of the 60 scenes(1) a visual preference rating sheet, with 5-point
Likert type rating scales intended for visual prefer- included in the present study would be neither practical
nor very interesting. Instead, landscape cate-ence ratings;
(2) eight items designed to measure the follow- gory sumscores were created on the basis of perceptual
dimensions derived from the preference ratingsing demographic characteristics: age, gender, organization
membership, geographical region during of the scenes in the current data set. Such dimensions
were identified through the application of achildhood, present geographical region, population
density during childhood, present population den- nonmetric factor analytic procedure, the SSA-III, or
the Guttman–Lingoes Smallest Space Analysissity and expertise (i.e. type of university or college
study currently being undertaken). (Lingoes, 1972), with normalized varimax rotation.
The benefits of the SSA-III procedure comparedIn the questionnaire, some of these measures
were dichotomies (gender and organization with metric factor analysis, is that it increases the
stability of the solution (Kaplan & Kaplan, 1989)membership), whereas others were polytomies.
However, because of the relatively small sample (n= and decreases the number of dimensions (Herzog,
1992). Following established criteria (cf. Kaplan &198) in the present study, and in order to simplify
the use of statistical techniques, all variables were Kaplan, 1989, for example) a seven-dimensional
solution including ratings for 50 of the scenes gave adichotomized. For the variable ‘age’, dichotomization
was based on the fact that 1970 was the readily interpretable result, with no dimension
including fewer than three items (Table 1). Thusmedian for this variable. For all other originally
nondichotomous variables, the guiding principle for sumscores were computed by summing up the items
measuring preferences for the scenes included indichotomization was to end up with intuitively
meaningful categorizations, for example urban vs each landscape category. In this process, missing
values were substituted by item means, and outliersrural, and Western Norwegians vs others. It should
be noted that in answering the two items measur- were recorded to ±2S.D. from the mean of its sumscore
(cf. Aarø, 1986). A more detailed description ofing, respectively, geographical region and population
density during childhood, subjects were the analysis of these dimensions has been given
elsewhere (Strumse, 1994a).asked to think of the place they lived most of the
time before their sixteenth birthday. In the following presentation of results, mean
21Demography and Preferences for Norwegian Agrarian Landscapes
preference ratings at 3·7 and above were considered ferred of all was the Modern Farming Elements category.
For a thorough analysis of these preferenceas ‘high’, mean ratings between 3·0 and 3·7 as ‘moderate’,
and means below 3·0 as ‘low’. These cut-off levels, see Strumse (1994a).
One-way (bivariate) analyses of variance werepoints were chosen as they have repeatedly been
employed in visual preference studies based on the performed separately for each demographic variable
(independent) on each landscape category sumscoregeneral finding that ratings at 4·0 or above and at
2·0 or lower are highly unusual (cf. Kaplan & (dependent). Moreover, to assess further the effectiveness
among demographic variables as predictorsKaplan, 1989). In addition, these characterizations
of mean preference levels enhance both the review of differences in preference for each landscape category,
a special kind of multiple regression, the Mul-of the results and the subsequent interpretation of
divergencies between group means and the overall tiple Classification Analysis (MCA) (Andrews et al.,
1973) was employed. MCA is designed to handlemean preference for a given category.
The two most preferred categories, both at high categorical predictors.
A first look at the bivariate analyses (Table 2)mean rating levels, depicted traditional scenes
(Table 1). However, whereas the Flowers category revealed that for two of the landscape categories, no
significant group differences were found. Thus, forwas dominated by ‘nature’ scenes, the Old Structures
category was characterized by human influ- the Old Structures category, across all groups, the
mean ratings were consistently high, between 4·35ence. The moderately preferred Farming category
was, in contrast to all other categories, dominated and 4·48 (see examples of the settings in Fig. 1). The
Spruce Plantations category, however, was ratedby activity, but contained both traditional and modern
scenes. Also at a moderate level, but distinctly consistently low, between 1·78 and 1·96 (Fig. 2). No
further analysis was done for these two categories.lower than the Farming scenes, was the Green,
Grassy Fields category, depicting vegetation from For the remaining landscape categories, between
one and five demographic variables yielded signifi-predominantly modern agrarian landscapes. The
New Dominating Structures and the Spruce Plan- cant results. However, two familiarity variables,
geographical region during childhood and popu-tations categories, although very different in content,
were at essentially the same preference level, lation density during childhood, proved insignificant
across all landscape categories. These resultsi.e. they were generally not preferred. Least preTABLE
1
Sumscores computed on the basis of preference-derived (n=198) categories in Western Norwegian agrarian landscapes:
descriptive statistics
Label Description No. of Mean S.D. Minimum Maximum
items
Farming Persons doing manual labour or using machines, 12 3·63 0·62 1·83 4·92
products of such activities, or reflections
of activities associated with farming
Old Old buildings or mechanical equipment embedded 9 4·43 0·42 3·49 5·00
structures in a nature setting, but also a nature scene,
the ground covered with stone
Green, grassy Modern mowed fields or meadows 9 3·14 0·57 1·93 4·67
fields
Modern farming Silos, drainpipes and forest machines 6 1·66 0·51 1·00 2·80
elements
New dominating Modern buildings and constructions 7 1·95 0·51 1·00 3·04
structures
Flowers Flowers and colourful meadows, high in 4 4·34 0·54 3·09 5·00
biological diversity
Spruce Relatively dense spruce plantations 3 1·86 0·81 1·00 3·58
plantations
22 E. Strumse
TABLE 2
Mean preference ratings for seven landscape categories by demography: one-way analyses of variance (S.D.s in
parentheses)
Groups (n) Landscape category
Farming Old Green, Modern New Flowers Spruce
Structures Grassy Farming Dominating Plantations
Fields
Familiarity
Geographical region during
childhood
Western Norway (95) 3·62 4·42 3·10 1·71 2·01 4·37 1·88
(0·63) (0·42) (0·60) (0·55) (0·50) (0·52) (0.81)
Other (103) 3·64 4·45 3·17 1·60 1·90 4·32 1·84
(0·60) (0·43) (0·54) (0·46) (0·51) (0·56) (0·82)
Present geographical region
Western Norway (146) 3·56** 4·42 3·16 1·67 1·98 4·34 1·85
(0·63) (0·42) (0·60) (0·53) (0·51) (0·55) (0·80)
Other (52) 3·83 4·46 3·09 1·63 1·89 4·34 1·88
(0·52) (0·44) (0·49) (0·45) (0·49) (0·51) (0·84)
Population density
during childhood
Urban (85) 3·66 4·49 3·18 1·65 2·01 4·37 1·83
(0·58) (0·43) (0·55) (0·51) (0·48) (0·56) (0·80)
Rural (113) 3·60 4·39 3·11 1·66 1·91 4·32 1·88
(0·64) (0·42) (0·59) (0·51) (0·52) (0·52) (0·82)
Present population density
Urban (136) 3·52*** 4·42 3·19 1·72** 2·01* 4·33 1·89
(0·60) (0·43) (0·54) (0·52) (0·51) (0·54) (0·79)
Rural (62) 3·87 4·47 3·03 1·51 1·82 4·37 1·78
(0·60) (0·41) (0·62) (0·46) (0·47) (0·55) (0·85)
Subculture
Age (year of birth)
Before 1970 (92) 3·82*** 4·47 3·12 1·72 2·02 4·37 1·85
(0·53) (0·45) (0·54) (0·53) (0·50) (0·54) (0·84)
After 1970 (106) 3·46 4·40 3·15 1·60 1·89 4·32 1·87
(0·64) (0·40) (0·60) (0·49) (0·50) (0·54) (0·79)
Gender
Men (72) 3·56 4·36 2·99** 1·71 2·01 4·20** 1·96
(0·61) (0·48) (0·52) (0·51) (0·51) (0·55) (0·80)
Women (126) 3·67 4·48 3·22 1·63 1·92 4·43 1·80
(0·62) (0·38) (0·58) (0·51) (0·50) (0·52) (0·82)
Organization membership
Yes (153) 3·70** 4·45 3·15 1·66 1·95 4·40** 1·84
(0·62) (0·41) (0·55) (0·50) (0·51) (0·52) (0·80)
No (45) 3·38 4·36 3·09 1·64 1·97 4·15 1·92
(0·52) (0·46) (0·62) (0·53) (0·51) (0·57) (0·86)
Expertise
Expert (94) 3·84*** 4·48 3·02** 1·64 1·96 4·40 1·82
(0·54) (0·43) (0·57) (0·49) (0·51) (0·51) (0·82)
Nonexpert (104) 3·44 4·39 3·24 1·67 1·95 4·29 1·89
(0·62) (0·42) (0·55) (0·53) (0·50) (0·56) (0·80)
*p<0·05; **p<0·01; ***p<0·001.
23Demography and Preferences for Norwegian Agrarian Landscapes
FIGURE 1. No group differences were found for the scenes included in the Old Structures category, which received consistently high
preferences (overall mean=4·43).
Thus, while subjects presently living in Western
Norway were found at a moderate preference level
(mean=3·56), those living in other regions had high
preferences (mean=3·83) for these scenes. Likewise,
in contrast to urban residents (mean=3·52), subjects
living in rural areas were found to have high preferences
(mean=3·87) for the Farming scenes. Among
subcultural variables, age and organization membership
both yielded highly significant results (both
at p<0·001). Here, those born before 1970 (mean=
3·82) and members (mean=3·70) were highly in favour
of the Farming scenes, whereas younger subjects
(mean=3·46) and nonmembers (mean=3·83)
were found within the moderate preference range.FIGURE 2. The preferences for scenes like this one from the
Space Plantations category, were consistently low (mean=1·85). Expertise also yielded clear group differences (p<
0·001), with high preferences (mean=3·84) in the
expert group and moderate preferences (mean=are somewhat surprising as landscape exposure as
3·44) among nonexperts.a child has repeatedly been found to influence prefThe
MCA for the Farming category (Table 3)erences. All other demographic variables yielded
yielded a significant (p<0·001) over all multiple Rsignificant results for one or more of the landscape
square of 0·195, with age as the most powerful pre-categories. With an exception for the Farming catedictor
variable (β=0·21, p<0·01). Not significant ingory, the MCAs did not provide additional inforbivariate
analysis, gender obtained a highly signifi-mation, as they only confirmed the bivariate
cant result in MCA (β=0.18, p<0·01). Present popu-results. For this landscape category, additional
lation density also proved to be a significant predic-analyses of variance were performed to explore the
tor, with a β of 0·16 (p<0.05). However, in contrastnature of these divergencies. In the following secto
the bivariate results, both present geographicaltions, results of the bivariate and multivariate
region and expertise turned out to be insignificantanalyses (MCA) will be presented in greater detail.
in the MCA. Thus further analyses had to be made.The Farming category yielded the highest degree
of diverging preferences in bivariate analyses, with In bivariate analysis of expertise, controlling for
present population density, it was found thatsignificant results for five of the eight demographic
variables (Table 2, first column from the left). More- expertise remained significant only for experts and
nonexperts living in urban areas (Table 4, upperover, all observed differences went beyond the overall,
moderate preference level of the category. First, section). With the rural nonexpert group being too
small (n=9), the difference between experts andamong measures of familiarity, present geographical
region (p<0·01) and present population density nonexperts living in rural areas could not be tested
statistically. Further bivariate analysis also indi-(p<0·001), yielded clear differences in preference.
24 E. Strumse
TABLE 3.
The Farming category sumscore by demographic variables: Multiple Classification Analysis
Unadjusted Adjusted
Variable n deviation η deviation β
Geographical region
during childhood
Western Norway 95 −0·01 0·02
Other regions 103 0·01 −0·01
0·02 0·02†
Present geographical region
Western Norway 146 −0·07 −0·02
Other regions 52 0·20 0·06
0·20 0·06†
Population density
during childhood
Urban 85 0·03 0·02
Rural 113 −0·03 −0·02
0·05 0·03†
Present population
density
Urban 136 −0·11 −0·07
Rural 62 0·24 0·15
0·26 0·16*
Age (year of birth)
Before 1970 92 0·19 0·14
After 1970 106 −0·17 −0·12
0·29 0·21**
Gender
Men 72 −0·07 −0·15
Women 126 0·04 0·08
0·09 0·18**
Organization membership
Yes 153 0·07 0·04
No 45 −0·25 −0·14
0·22 0·12†
Expertise
Experts 94 0·21 0·07
Nonexperts 104 −0·19 −0·06
0·33 0·11
Multiple R squared 0·195***
Multiple R 0·442
Grand mean=3·629. *p<0·05; **p<0·01; ***p<0·001; †not significant.
cated that expertise seemed to remove the effect of only for urban residents (Table 4, lower section).
Photographic samples of the Farming scenes arepresent geographical region. Also here, the differences
between two of the subgroups, nonexperts liv- found in Fig. 3.
For the Green, Grassy Fields category (see Tableing in Western Norway and nonexperts living in
other regions, could not be subjected to statistical 2, third column from the left), significant bivariate
results were found for gender and expertise (both atanalysis, as the latter group consisted of only one
individual (see Table 4, middle section). On the p<0·01): women were, well in keeping with the overall
preferences for this category, moderate (mean=other hand, when controlling for present population
density, the effect of present geographical region 3·22) in their liking of these scenes, whereas men
(mean=2·99) were found in the low preferencefound in bivariate analysis, was reproduced, but
25Demography and Preferences for Norwegian Agrarian Landscapes
TABLE 4
Further bivariate analyses for the Farming category: expertise controlled for present population density, present
geographical region controlled for expertise, and present geographical region controlled for present population density:
one-way analyses of variance
Variable n Mean S.D.
Expertise controlled
for present population density
Urban experts 41 3·81*** 0·49
Urban nonexperts 95 3·40 0·60
Rural experts 53 3·87‡ 0·59
Rural nonexperts 9 3·85 0·71
Present geographical
region controlled for expertise
Western Norwegian experts 43 3·84† 0·59
Experts from other regions 51 3·85 0·51
Western Norwegian nonexperts 103 3·44‡ 0·62
Nonexperts from other regions 1 3·00 −
Present geographical region
controlled for present population density
Urban Western Norwegians 115 3·48** 0·61
Urban, other regions 21 3·73 0·50
Rural Western Norwegians 31 3·83† 0·67
Rural, other regions 31 3·90 0·53
*p<0·05; **p<0·01; ***p<0·001; †not significant; ‡due to an insufficient group size, statistics could either not be
calculated, or results would be unreliable.
FIGURE 3. Although at an overall moderate preference level, the Farming scenes were highly preferred by the older age group in the
sample (mean=3·82), rural residents (mean=3·87), urban residents not living in Western Norway (mean=3·73), urban experts (mean=
3·81) and organization members (mean=3·70).
range. While both were moderate in their prefer- indicated that only familiarity seemed important
here, with present population density yielding a sig-ences for this category, experts (mean=3·02) were
found at a significantly lower level than nonexperts nificant (p<0·01) result. Here, urban residents
(mean=1·72) were more in favour of the category(mean=3·24). One of the scenes included in this category
is depicted in Fig. 4. than were those living in rural areas (mean=1·52),
but both groups were still found within the low pref-The bivariate results for the Modern Farming
Elements category (see Table 2, middle column) erence range (Fig. 5, left photo).
26 E. Strumse
For the New Dominating Structures category (see (mean=4·40) were more in favour of the Flowers category
than were, respectively, men (mean=4·20) andTable 2, third column from the right), the results of
bivariate analyses were similar to those found for nonmembers (mean=4·15). Settings from the Flowers
category are depicted in Fig. 6.the Modern Farming Elements, only somewhat
weaker, with p<0·05 for present population density. A brief summary of the findings presented above,
sharpening the focus upon demography by consider-While both groups were found in the low preference
range, subjects presently living in urban areas ing each demographic variable separately across
landscape categories, might add to their clarity.(mean=2·01) were somewhat more in favour of the
category than were rural residents (mean=1·82). Among the familiarity variables, two of these, geographical
region during childhood and populationOne of the Modern Farming Elements scenes is
depicted in Fig. 5, right-hand photo. density during childhood, did not affect the visual
preferences for any of the landscape categories. InPreferences for the Flowers category (see Table 2,
second column from the right) proved in bivariate contrast, the parallel circumstances presently did
seem to affect landscape preferences. Thus, subjectsanalyses to be significantly affected by two subcultural
variables: gender and organization member- presently living in urban areas scored higher than
rural residents in their visual preferences for theship (both at p<0·01). However, all group means
were within the overall high preference range. With New Dominating Structures and Modern Farming
Elements categories, whereas they were lower inthis reservation, women (mean=4·43) and members
their preferences for the Farming category. In the
latter case, further analysis revealed that high preference
was only found among urban landscape
experts, and that Western Norwegian urban residents
were lower in their visual preferences for
Farming scenes than were urban residents living
elsewhere.
Among subcultural variables, age was a strong
predictor of visual preference for the Farming category,
with older subjects found at the higher level of
preference. Furthermore, gender affected the visual
preferences for the Farming, Green, Grassy Fields
and Flowers categories, women in all cases being
more in favour of these scenes. In addition, organization
members exhibited higher preferences than
did nonmembers for the Farming and Green,FIGURE 4. The scenes included in the Green, Grassy Fields
Grassy Fields categories.category were more preferred by women (mean=3·22) and
nonexperts (mean=3·24) than by men (mean=2·99) and experts Finally, knowledge, in this case landscape expert-
(mean=3·02).
ise, led to higher preferences among experts comFIGURE
5. Urban residents were somewhat more (mean=1·72) in favour of scenes containing Modern Farming Elements and New
Dominating Structures than were rural residents (mean=1·51).
27Demography and Preferences for Norwegian Agrarian Landscapes
ability to respond with preference to environments
exhibiting such attributes has been selected for, as
they facilitate the efficient processing of environmental
information. However, for the Spruce Plantations
category, Mystery seems to have an
opposite, perhaps even fear-inducing effect, which
might be due to its combination with dense vegetation,
making the prediction of what might happen
next difficult (Strumse, 1994b).
The results for the Flowers and Green, Grassy
Fields categories could be understood as a relatively
clear example of the complementarity of constructivist
and evolutionary perspectives. While the
FIGURE 6. For the Flowers category, gender and organization impression of consensus dominated, significant
membership produced clear differences in preference. Women
group differences were found in the preferences for(mean=4·43) and members (mean=4·40) were significantly (in
both cases at p<0·01) higher in their visual preferences for these both categories. Unanimity seemed to be explained
scores than were men (mean=4·20) and nonmembers (mean= by these scenes representing environments provid-
4·15).
ing safe conditions for understanding and exploration
(cf. Kaplan & Kaplan, 1989). Nevertheless,
the subcultural factors gender and organizationpared with nonexperts for the Farming scenes,
whereas nonexperts were higher in their prefer- membership, and one knowledge-related factor,
expertise, seemed capable of producingences for the Green, Grassy Fields category.
divergencies.
The observed differences produced by gender support
the notion that this is a ‘subcultural’ variableDiscussion
capable of accounting for important differences with
The finding that no significant group differences respect to landscape preferences, here expressing
were found for the Old Structures and Spruce Plan- itself by women taking a more positive stance
tations categories, suggests that preferences for towards nature than men. It is also possible, howthese
landscapes could be explained, first, on the ever, to interpret the gender differences in evolbasis
of their content, and second, according to the utionary terms. For the Flowers category, the differpresence
or absence in them of preference-promot- ences seem compatible with findings indicating
ing environmental attributes (Strumse, 1994a, b). women to be superior in their perception and memSpecifically,
the high preferences for the Old Struc- ory of complex arrays of vegetation, whereas men
tures category seem to be related to those scenes are better on mental rotation, map-reading and
containing human influence in balance with the maze learning (Silverman & Eals, 1992). This is
natural elements (Strumse, 1994a). For the Spruce hypothesized to be so because, during hominid evolPlantations
category, however, dense vegetation, ution, men predominantly hunted and women preresulting
in a blocked view, seemed responsible for dominantly foraged (Tooby & DeVore, 1987). Morethe
low preference levels (cf. Strumse, 1994a). over, a higher preference in women could be due to
The role of preference-promoting environmental the fact that categories such as Flowers and Green,
features will here be interpreted within a func- Grassy Fields, which are open and well-defined settionalist–evolutionary
framework (cf. Kaplan & tings (Strumse, 1994b), would most probably induce
Kaplan, 1989; Kaplan, 1992). Thus, for the Old feelings of security, which is perhaps more importStructures
category, preference seems associated ant to women than to men.
with high degrees in it of one informational vari- The finding that experts exhibited somewhat
able, Mystery (Strumse, 1994b), which promotes lower preferences than nonexperts for the Green,
exploration by promising new information if one Grassy Fields category suggests that training
could enter into the setting. Mystery has reliably within landscape disciplines leads to a more critical
proved itself to predict visual preferences and is (or negative) attitude towards scenes of less than
assumed to be, together with factors promoting optimal ecological quality, in this case manifesting
understanding, a salient characteristic of environ- itself in a somewhat uniform, schematic layout of
ments preferred by humans (Kaplan & Kaplan, space and a low biological diversity (Strumse,
1989). Presumably, through evolutionary time, the 1994b). However, organization members were
28 E. Strumse
somewhat more in favour of the highly preferred directly, or at all, with the other categories
employed in this study, precisely because farmingFlowers category than were nonmembers. This
makes sense, as membership could be regarded as activity, rather than landscape type, emerged as the
common perceptual theme for this category. On thean indication of interest in a broad spectrum of
issues, including environmental protection. This other hand, the inclusion of this category could be
justified because such activities are intimately con-applies well to the Flowers category, containing as
it does aesthetically valuable traditional meadows nected with the shaping and maintenance of agrarian
landscapes, and that both visitors to and resi-with high biological diversity (Strumse, 1994a).
For both the Modern Farming Elements and the dents in such landscapes are likely to encounter
such situations.New Dominating Structures categories, present
population density resulted in one small group dif- The present study has demonstrated an almost
unanimous consensus with respect to (a) the highference, with urban residents slightly more in
favour of these scenes. These findings suggest a preferences for traditional human-influenced settings
and nature scenes, and (b) the relative dislikeslightly negative relationship between familiarity
and preference. Apart from these minor divergenc- for dominating human influence and many of the
effects of modern farming practices. This lendsies, for these settings too the overall (low) preference
pattern seemed readily explained both by their support, at least indirectly, to the relevance of a
functionalist–evolutionary interpretation of visualcontent and by the degree to which preferenceenhancing
environmental attributes are present in landscape preferences, and, at the same time, it
confirms the impression of strong support for thethem (Strumse, 1994a, b).
So far, a relatively unambiguous consensus in protection of traditional agrarian landscapes,
reported in earlier analyses (Strumse, 1994a, b).landscape preferences, favouring a functionalist–
evolutionary position, has been demonstrated for However, in spite of this strong tendency towards
consensus, diverging preferences for agrarian land-the categories discussed. However, the comparatively
large divergencies among groups found for scapes in Western Norway do exist. The observed
group differences suggest the importance of takingthe moderately preferred Farming category constituted
a sharp contrast to this overall picture and into account the preferences of those affected by the
management of particular landscapes, especiallythe strongest support for the constructivist position
in the present study. In the final analysis, the the moderately preferred landscapes, where divergencies
are more pronounced. Particularly import-strongest predictors of differences in visual preferences
for the Farming scenes were, in descending ant group differences seem to be associated with
gender, organization membership, present popu-order, age, gender and present population density.
However, expertise and present geographical region lation density, expertise and, to a lesser extent, age
and present geographical region. On the otheralso exerted important influences on visual preference
for this category of scenes. hand, in this study at least, geographical region and
population density during childhood did not seem toResults indicating that preferences for the Farming
category increase with increasing age could per- influence landscape preferences in adults.
haps be understood as a result of the rarifying of
direct experience of farming practices in our age of
rapidly increasing urbanization. Furthermore, the Conclusion
impact of expertise may be understood as resulting
from a higher level of knowledge about agrarian The present study permitted an estimation of the
impact of demographic variables on visual prefer-practices in experts than in nonexperts. Finally,
subjects with higher preferences for the Farming ences for a given sample of agrarian landscapes. For
highly preferred traditional, as well as for less-pre-category, who were indeed rural residents, and thus
supposedly more familiar with such scenes than ferred modern landscapes, the impression of consensus
is overwhelming, leading to the conclusion that,other groups, could be hypothesized to identify more
closely than others with the persons depicted and for these landscapes, group differences are relatively
small compared with the effects produced bythe activities they performed. Such identification
could, in turn, be understood as part of subjects’ the landscapes themselves. Nevertheless, some
clear effects of demography, notably for moderatelygeographical (or place) identity (cf. Proshansky et
al., 1983; Feldman, 1990; Williams & Roggenbuck, preferred landscapes, were identified. This
reinforces the argument that group differences in1989). This, however, also raises the question of
whether the Farming category can be compared landscape preferences should not be neglected by
29Demography and Preferences for Norwegian Agrarian Landscapes
perceptions of familiar natural environments. Unpub-planners, managers and other landscape experts. A
lished doctoral dissertation, University of Michigan,further accentuation of this recommendation comes
Ann Arbor.
from the fact that one of the more important vari- Andrews, F. M., Morgan, J. N., Sonquist, J. A. & Klem, L.
ables in producing such differences is expertise (1973). Multiple Classification Analysis. A Report on a
Computer Program for Multiple Regression Using Cat-itself. It might be argued that visual preference
egorical Predictors, 2nd Edn. Ann Arbor, MI: Instituteresults obtained for colour slides limits their genfor
Social Research, University of Michigan.eralizability into real world settings. However,
Appleton, J. (1975). The Experience of Landscape.
several studies have pointed out that the use of London: John Wiley & Sons.
photographic representations yields valid and Austad, I. (1993). Verdier i det tradisjonelle kulturlandskapet.
In E. Framstad & S. Rysstad, Eds, Jordbru-reliable results (cf Levine, 1977; Ulrich, 1979;
kets kulturlandskap. Forskerkonferansen 1992 26–27.Shuttleworth, 1980; Savolainen & Kelloma¨ki, 1984;
oktober—Sundvolden Hotell. A° s: Norges Forsknings-Zube et al., 1987).
ra˙d, Forskningsprogram om kulturlandskapet.
Some important weaknesses of the study should Austad, I. & Kaland, P. E. (1991). Endring i biologisk
be noted. First, virtually all nonexperts were living mangfold i tradisjonelle kulturmarkstyper pa˙ Vestlandet
ved gjengroing, tilplanting og skjo´tseltiltak. For-in urban areas, resulting in a near-absence of rural
skningsprosjekt i tilknytning til NMF/NLVF’s for-nonexperts. Also, the age span in the sample was
skningsprogram om kulturlandskapet. Sogn ogsmaller than desirable, suggesting that differences
Fjordane distriktshøgskule og Botanisk institutt,
due to age could have been underestimated. Fur- Universitetet i Bergen.
thermore, using students, the effect of direct experi- Austad, I., Skogen, A., Hauge, L., Helle, T. & Timberlid,
A. (1991). Human influenced vegetation types andence with the landscapes in question could not be
landscape elements in the cultural landscapes ofexamined. These specific weaknesses reflect the
inner Sogn, western Norway. Norsk. geogr. Tidsskr.,more general objection that the cultural and demo-
45, 35–58.
graphic homogeneity of the sample reduces the Balling, J. D. & Falk, J. H. (1982). Development of visual
probability that group differences will be found. preference for natural environments. Environment
and Behavior, 14, 5–28.Thus, a more heterogenous sample would certainly
Bernaldez, F. G. & Abello, R. P. (1989). Environmentalmake it easier to find support for the constructivist
challenge and environmental preference: age and sexposition. In an eventual follow-up, these considereffects.
Journal of Environmental Management, 28,
ations should be kept in mind. 53–70.
Bourassa, S. C. (1990). A paradigm for landscape aesthetics.
Environment and Behavior, 22(6), 787–812.
Acknowledgements Buhyoff, G. J., Wellman, J. D., Harvey, H. & Fraser, R. A.
(1978). Landscape architects’ interpretations of
people’s landscape preferences. Journal of Environ-This study was made possible by a research grant
mental Management, 6, 255–262.from the Norwegian Research Council, as part of a
Cosmides, L., Tooby, J. & Barkow, J. (1992). Introduction:
research programme on cultural landscapes. The Evolutionary psychology and conceptual integration.
author wishes to express his thanks to Leif Edvard In J. Barkow, L. Cosmides & J. Tooby, Eds, The
Aarø at the University of Bergen for excellent Adapted Mind. Evolutionary Psychology and the Generation
of Culture. New York: Oxford Universityadvice in methodological and statistical matters, to
Press, pp. 3–15.Leif Hauge at the Sogn og Fjordane College for
Daniel, T. C. & Boster, R. S. (1976). Measuring Landscape
taking all the photos employed in this study, and to Aesthetics: The SBE Method. Boulder, CO: USDA.
Rachel and Stephen Kaplan at the University of USDA Forest Service Research Paper RM-167.
Michigan for helping with the analysis of the prefer- Dearden, P. (1984). Factors influencing landscape preferences:
an empirical investigation. Landscape Plan-ence ratings.
ning, 11, 293–306.
Ellsworth, J. C. (1982). Visual assessment of rivers and
marshes: an examination of the relationship visualNotes
units, perceptual variables, and preference. Unpublished
Master’s thesis, Utah State University, Logan.
Correspondence should be addressed to: Einar Strumse, Feldman, R. M. (1990). Settlement-identity. Psychological
Research Center for Health Promotion, Øisteins gt. 3, N- bonds with home places in a mobile society. Environ-
5007 Bergen, Norway; Tel. +47 55 21 33 06; E-mail: einar- ment and Behavior, 22(2), 183–229.
.strumse@psych.uib.no. Gallagher, T. J. (1977). Visual preference for alternative
natural landscapes. Unpublished doctoral dissertation,
University of Michigan, Ann Arbor.
References Hammitt, W. E. (1978). Visual and user preference for a
bog environment. Unpublished doctoral dissertation,
Anderson, E. (1978). Visual resource assessment: local University of Michigan, Ann Arbor.
30 E. Strumse
Hammitt, W. E. (1987). Visual recognition capacity Kellert, S. R. (1978). Perceptions of animals in American
society. 41st North American Wildlife Conference Pro-during outdoor recreation experiences. Environment
and Behavior, 14, 478–493. ceedings, pp. 533–546.
Keyes, B. E. (1984). Visual preferences of a forest trailHerbert, E. J. (1981). Visual resource analysis: prediction
and preference in Oakland County, Michigan. Unpub- environment. Unpublished Master’s thesis, University
of Tennessee, Knoxville.lished Master’s thesis, University of Michigan, Ann
Arbor. Levin, J. (1977). Riverside preference: on-site and photographic
reactions. Unpublished master’s thesis, Uni-Herzog, T. R. (1992). A cognitive analysis of preferences
for urban spaces. Journal of Environmental Psy- versity of Michigan, Ann Arbor.
Lingoes, J. C. (1972). A general survey of the Guttman–-chology, 12, 237–248.
Herzog, T. R., Kaplan, S. & Kaplan, R. (1982). The predic- Lingoes nonmetric program series. In R. N. Shephard,
A. K. Romney & S. B. Nerlove, Eds, Multidimensionaltion of preference for unfamiliar urban landscapes.
Population and Environment: Behavioral and Social scaling, Vol. 1. New York: Seminar.
Lundberg, A. (1990). Vegetasjon og kulturlandskap vedIssues, 5, 43–59.
Hudspeth, T. R. (1982). Visual preference as a tool for Kalandsvatnet i Fana Bergen: Institutt for geografi,
Norges Handelshøyskole og Universitetet i Bergencitizen participation: a case study of urban waterfront
revitalization in Burlington, Vermont. Unpublished (Geografi i Bergen, nr. 140).
Lyons, E. (1983). Demographic correlates of landscapedoctoral dissertation, University of Michigan, Ann
Arbor. preference. Environment and Behavior, 15, 487–511.
Macia (1979). Visual perception of landscape: sex and per-Hultman, S. G. (1983). Allma¨nhetens bedo¨mning av
skogsmiljo¨ers la¨mplighet fo¨r friluftsliv. 2. En sonality differences. In G. H. Elsner & R. C. Smardon,
Eds, Our National Landscape. USDA Forest Serviceriksda¨ckande enka¨t. Rapport 28. Sveriges lantbruksuniversitet.
General Technical Report PSW-35. Berkeley, CA:
USDA.Kaland, P. E. (1993). Forskning pa˙ tradisjonelle kulturmarkstyper.
In E. Framstad, E. & S. Rysstad, Eds, Medina, A. Q. (1983). A visual assessment of children’s
and environmental educators’ urban residential pref-Jordbrukets kulturlandskap. Forskerkonferansen
1992 26–27. oktober—Sundvolden Hotell. A° s: Norges erence patterns. Unpublished doctoral dissertation,
University of Michigan, Ann Arbor.Forskningsra˙d, Forskningsprogram om kulturlandskapet.
Miller, P. A. (1984). Visual preference and implications
for coastal management: a perceptual study of theKaplan, R. (1973). Predictors of environmental preference:
designers and ‘clients’. In W. F. E. Preiser, Ed., British Columbia shoreline. Unpublished doctoral dissertation,
University of Michigan, Ann Arbor.Environmental Design Research. Stroudsburgh, PA:
Dowden, Hutchinson & Ross. Moore, G. T. (1979). Knowing about environmental knowing:
the current state of theory and research onKaplan, R. (1977). Preference and everyday nature:
method and application. In D. Stokols, Ed., Perspec- environmental cognition. Environment and Behavior,
11, 33–70.tives on Environment and Behavior: Theory, Research,
and Applications. New York: Plenum. Østerbø, V. (1990). Kulturlandskapet kring Kalandsvatnet,
Fana, Bergen kommune. Bergen: Universitetet iKaplan, R. (1985). Nature at the doorstep: residential satisfaction
and the nearby environment. Journal of Bergen, Institutt for geografi.
Penning-Rowsell, E., Gullett, G. H., Searle, G. H. &Architectural and Planning Research, 2, 115–127.
Kaplan, R. & Herbert, E. J. (1987). Cultural and subcul- Witham, S. A. (1977). Public Evaluation of Landscape
Quality. Planning Research Group Report 13. Enfield:tural comparisons in preference for natural settings.
Landscape and Urban Planning, 14, 281–293. Middlesex Polytechnic.
Polit, D. F. & Hungler, B. P. (1991). Nursing Research.Kaplan, R. & Kaplan, S. (1983). Cognition and Environment.
Functioning in an Uncertain World. Ann Arbor, Principles and Methods, 4th Edn. Philadelphia: J. B.
Lippincott.MI: Ulrichs Bookstore.
Kaplan, R. & Kaplan, S. (1989). The Experience of Nature. Porter, M. V. (1987). The role of spatial quality and familiarity
in determining countryside landscape prefer-A Psychological Perspective. Cambridge: Cambridge
University Press. ence. Unpublished Master’s thesis. University of
Washington, Seattle.Kaplan, S. (1992). Environmental preference in
a knowledge-seeking, knowledge-using organism. In Proshansky, H. M., Fabian, A. K. & Kaminoff, R. (1983).
Place-identity: physical world socialization of the self.J. H. Barkow, L. Cosmides & J. Tooby, Eds, The
Adapted Mind. Evolutionary Psychology and the Gen- Journal of Environmental Psychology, 3, 57–83.
Sandahl, T. J. (1989). Kalandsvatnet—Fanaelva Eneration of Culture. New York: Oxford University
Press, pp. 581–598. utredning av omra˙dets egnethet for allment friluftsliv,
basert pa˙ areal- og landskapsvurderinger, samtKardell, L. & Ma˙rd, H. (1989). Na˙gra gruppers attituder
till stubbrytning 1976 och 1988. Rapport 41. Avdeln- fysisk/biologiske registreringer langs strandkantene i
vassdraget. Bergen: Bergen kommune. Kommuna-ingen fo¨r landskapsva˙rd. Sveriges Lantbruksuniversitet.
lavdeling fritid, kultur og kirke. Fritidsavdelingen.
Savolainen, R. & Kelloma¨ki, S. (1984). Scenic value of theKeane, T. D. (1990). The role of familiarity in landscape
aesthetics: a study of tallgrass prairie landscapes. A forest as assessed in the field and the laboratory. In
O. Saastomoinen, S. G. Hultman, N. E. Koch & L.dissertation submitted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy Mattsson, Eds, Multiple-use Forestry in the Scandinavian
Countries. Helsinki: Finnish Forest Research(Landscape Architecture) in The University of
Michigan. Institute.
31Demography and Preferences for Norwegian Agrarian Landscapes
Shuttleworth, S. (1980). The use of photographs as an In W. Kinzey, Ed., Primate Models of Human
Behavior. New York: SUNY Press.environmental presentation medium in landscape
studies. Journal of Environmental Management, 11, Ulrich, R. S. (1979). Visual landscapes and psychological
61–76. well-being. Landscape Research, 4, 17–23.
Silverman, I. & Eals, M. (1992). Sex differences in spatial Williams, D. R. & Roggenbuck, J. W. (1989). Measuring
abilities: evolutionary theory and data. In J. H. Place Attachment: Some Preliminary Results. Paper
Barkow, L. Cosmides & J. Tooby, Eds, The Adapted presented at the session on outdoor planning and
Mind. Evolutionary Psychology and the Generation of management. NRPA symposium on leisure research,
Culture. New York: Oxford University Press, pp. San Antonio, Texas, October.
533–553. Woodcock, D. M. (1982). A functionalist approach to
Stern, R. C., Dietz, T. & Yaloff, L. (1993). Value orien- environmental preference. Doctoral dissertation,
tations, gender, and environmental concern. Environ- University of Michigan.
ment and Behavior, 25(3), 322–348. Zube, E. H., Pitt, D. G. & Anderson, T. W. (1974). PercepStrumse,
E. (1994a). Perceptual dimensions in the visual tion and Measurement of Scenic Resources in the
preferences for agrarian landscapes in western Connecticut River Valley. Amherst: University of
Norway. Journal of Environmental Psychology, 14, Massachusetts.
281–292.
Zube, E. H., Sell, J. L. & Taylor, J. G. (1982). Landscape
Strumse, E. (1994b). Environmental attributes and the
perception: research, application and theory. Landprediction
of visual preferences for agrarian landscape
Planning, 9, 1–33.
scapes in western Norway. Journal of Environmental
Zube, E. H., Simcox, D. E. & Law, C. S. (1987). PerceptualPsychology, 14, 293–303.
landscape simulations: history and prospect. Land-Tooby, J. & DeVore, I. (1987). The reconstruction of homiscape
Journal, 6, 62–80.nid behavioral evolution through strategic modeling.