Research methods

Research instruments Click here
Representing the landscape Click here
Influence of the observer on preferences Click here
References
Click here

In this theme. the characteristics of research of landscape quality are examined. How the landscape may be represented, such as by photographs, is discussed. It then examines the influence of the characteristics of survey participants on their ratings. Before these, however, we need to understand that numbers come in four forms.

Types of numbers

In measuring anything, including landscape quality, numbers are used. However, numbers come in a variety of capabilities – according to Stevens (1946) there are four types of numbers:

Nominal, the weakest form, merely take the form of labels or categories, such as the numbers on the backs of football players for identification. Numbers can be used for classes, e.g. males = 1, females = 2. Such classes or numbers have no order (i.e. they are random) and have no sense of relativity (i.e. none is better than the other).

Ordinal implies a relative ranking, e.g. one mineral is harder than another, an odour is more pleasant than another, this landscape is more appealing than that one. While relativity is apparent, the intervals between classes are not necessarily equal, nor is there a baseline – i.e. zero point.

Interval measures are quantitative in the usual sense of the word. It provides a ranking between classes and an equal spacing between them, however, a zero point is arbitrarily assigned. The temperature scale is an interval scale – the units are of equal size and the zero on the Celsius scale is based on the freezing point of water; it does not denote the absence of temperature in the way that 0° Kelvin does (i.e. absolute zero). An interval scale enables precision about the differences in magnitude of objects; one can state that one is twice that of the other. Interval scales are commonly used in psychology. Much landscape research assumes that ordinal numbers are interval measures.

Ratio measures are the strongest form of numbers. They are sometimes known as cardinal numbers. Relativity between objects is defined, the intervals between units are equal, and an absolute zero point is included, which means the complete absence of the characteristic being measured. Ratio measures have all the properties of interval measures plus an absolute basepoint. Measures of weight and distance are examples.

The four qualities of numbers are summarized in Table 1.

Table 1 Summary of Types of Measures

Table 9

Most landscape research assumes that the data is interval whereas in fact they are mostly ordinal. According to Schroeder (1984), The simplest scaling methods treat the non-interval scale problem by ignoring it, they assume that rating data already possess interval properties and analyze them accordingly.

Research Instruments

In this section the range of research instruments used to research landscape quality are summarized.

Psychophysics Click here
Law of Comparative Judgement Click here
Scenic Beauty Estimation Method
Click here
Rating of photographs Click here
Paired photographs Click here
Q sort of photographs Click here
Visitor Employed Photographs (participant photography) Click here
Semantic differential Click here
Adjective checklist Click here
Physiological tests Click here
Interviews and questionnaire Click here
Measurement of features on photographs Click here
Research Instruments – Conclusions
Click here

Psychophysics

Psychophysics originated in the 19th century through the work of a German physicist, Gustav Fechner (1801 – 87), who defined it as an exact science of the functional relations of dependency between body and mind (Torgenson, 1958), or the measurement of sensations and perception (Lindzey et al, 1988) – in other words, it measures how the brain interprets information provided by the senses (see Chapter 6).

A basic assumption of psychophysics is that people are reasonably consistent in making judgements or choices among options. In regard to landscapes, people are unlikely to switch their preferences markedly during a test. Although some variability is acknowledged, it is assumed to display a normal distribution with the true value being represented by the mean (Hull et al, 1984).

According to the Macaulay Institute (2013):

Of all landscape assessments, (psychophysical) methods have been subjected to the most rigorous and extensive evaluation. They have been shown to be very sensitive to subtle landscape variations and psychophysical functions have proven very robust to changes in landscapes and in observers. Relying on ordinal or interval scales of measurement, psychophysical methods have consistently been able to provide different landscape-quality assessments for landscapes that vary only subtly.

The basic methodological approach involves measuring the relationship between the landscape as perceived, the independent variable, by human observers stating their preferences for the landscape, usually by use of some rating scale (e.g. 1 – 10) or by other physiological and psychological measures. The landscape is known as the independent variable because it remains constant and invariable regardless of who views it, rates it, measures it or examines it. Human observers, on the other hand, are not constant but their preferences and reactions to a landscape can vary widely influenced by a range of personal, interpersonal, cultural and other factors. Figure 1 summarizes the three elements of the method.

Fig 9 Landscape ass components
Figure 1 Landscape assessment components

Most of the research instruments are used to evaluate the dependent variables, measuring the preferences of observers and how they vary with the quality of the landscape. This is achieved through statistical measures, including correlation, multiple regression and factor analysis. Other studies which compare the preferences of one population sample with another use no independent variable.

A range of research instruments for measuring landscape preferences have been developed, ranging from simple to the sophisticated. The sophisticated research methods produce interval-type data and these include the Law of Comparative Judgement (LCJ) method and the Scenic Beauty Estimation (SBE) method, developed from psychophysics.

Law of Comparative Judgement

During the 1920s, Louis Thurstone (1887 – 1955) developed psychophysical scaling laws that enabled the accurate measurement of those psychological attributes resulting from stimuli but which had no physical manifestation. His Law of Comparative Judgement (Thurstone, 1927) is one of the key foundations for research into landscape quality and has been used widely across a range of disciplines, including psychology, engineering, marketing, and ergonomics (Hull, 1986).

The problem that Thurstone addressed was that individuals making judgements about the same feature would give similar but not identical responses at different times. Furthermore, while some individuals are consistent in their reliability, others may be very inconsistent. If respondents use, say, a 10-point scale and are asked to rate some feature, they may regard the interval differences between 5 and 6 as different than between 8 and 9. Thus, while the researcher may treat each unit as equal, the respondents may not. Scaling methods provide a means of transforming the raw responses into accurate and reliable scale values that reflect the perceived properties of the features.

Thurstone defined five cases in the Law of Comparative Judgement (LCJ), but we will concern ourselves only with the third case which involves several observers, each judging each pair of scenes several times. This is widely used in landscape research with photographs of scenes providing the stimuli and assuming that each photograph may be judged independently of others. Two photographs are shown side by side. The observer judges between the two and chooses one based on whatever criterion is defined – such as the preference of one over the other. Each photograph is compared successively with every other photograph. No single photograph appears twice in succession, and the aim should be to space them as far apart as possible. No two photos should be judged equal.

Fig 10 paired cf
Figure 2 Numbers of Photographs for Paired Comparisons – LCJ Method

A limitation of the method is that the number of paired comparisons grows rapidly with the number of photographs. With n stimuli, there are n(n-1)/2 pairs of stimuli required (Figure 2). It is generally impractical to go beyond about 15 pairs (Guildford, 1954). This would require 105 paired comparisons, which are about as many as one could reasonably expect a survey participant to undertake. Increasing the number of photographs by one (to 16) increases the paired comparisons to 120, a sizeable increase for only one extra photograph. An experiment requiring observers to make 120 comparative judgements could suffer from observer fatigue (Hull et al, 1984). While most of the studies using the LCJ method had less than 15 pairs of photographs, one used 29 pairs, requiring 406 paired comparisons (Whitemore et al, 1995). Figure 10 indicates the number of paired comparisons that need to be made for n photographs. Thus for 20 scenes, 200 paired comparisons must be made.

The LCJ method requires a large number of comparisons of each stimulus to provide sufficient data for analysis, so a balance has to be struck between exceeding the patience of the observers and providing sufficient data for analysis.

Through a series of mathematical steps and data transformations, the LCJ method provides interval scaling between preferences. This enables the results to be analyzed using standard statistical methods. A key researcher who has used the LCJ method extensively is Greg Buhyoff, who has used the method in twelve of his studies.

Scenic Beauty Estimation Method

The second psychophysical method is the Scenic Beauty Estimation (SBE) method that was developed in the mid-1970s by Terry Daniel, a psychologist at the University of Arizona, and Ron Boster, a forester with the US Forest Service. The SBE method transforms ordinal ratings to an interval scale SBEs.

SBE method has its origins in both the Law of Categorical Judgement and the Theory of Signal Detection. The Theory of Signal Detection has close parallels with Thurstone’s law (Green & Swets, 1966) and grew out of research to detect a weak message over a noisy telephone (Lindzey et al, 1988). The theory is based on the research finding that the cognitive state of the person doing the detecting, their biases and expectations, influences the results they attain. Providing rewards or punishments for the detection changes the cognitive state and one’s willingness to make false alarms or misses. However, one’s true sensitivity remains the same. Signal Detection Theory allows the researcher to separate spurious and real influences so as to determine the observer’s true sensitivity, provided the observer is neither cautious nor reckless.

Applying SBE to landscapes, an observer may form a negative judgement about Landscape A that they do not like it. Landscape B, however, exceeds the implicit criterion the observer sets and results in a positive judgement – “I like it”. If, however, the observer’s standards were raised for some reason, their judgement:

“would be negative for both landscapes, even though their perceived beauty has not changed. Thus, scenic beauty judgements depend jointly on the perceived properties of the landscape and the judgemental criteria of the observer” (Daniel & Boster, 1976, the author’s emphasis).

Ideally, if each observer rated a landscape out of a possible top score of 10, a rating of, say, 7 would be the equivalent across observers. However, this would be unusual because each observer’s criterion is unique, and the same landscape will be rated differently by different observers, making the scoring difficult to interpret. For example, one observer may rate a landscape as 3 out of 10, applying a very high aesthetic criterion, while another having a low aesthetic standard may score it as 8.

These and other problems of observer differences are solved through the SBE method, in which a measure of landscape beauty “independent of observer judgemental criteria” is derived.

The SBE method involves three stages: 1. taking color photographs of the landscape, 2. presenting these to participants for evaluation and 3. evaluating their judgements. Daniel & Boster developed the method in ponderosa forest on flat topography, but it has been used across many landscapes.

Like the LCJ method, the SBE method assumes that all individuals will categorize each slide in essentially the same location on their respective scenic beauty continuum, and that differences among individuals are normally distributed. These categories can then represent a basis from which measurements of scenic beauty can be made. Each category is indicated by a distribution, reflecting both individual differences and variability in perceptual and judgemental processes. The mean of the distribution is assumed to represent the true category.

Hypothetical results of a survey are presented in Figure 3. The three graphs indicate the scenic beauty scales assigned the average ratings given by the three observers to each of the landscapes are as follows:

Daniel & Boster
Daniel & Boster, 1976
Fig 11 SBE

Table 2   Derivation of Scenic Beauty Estimation (SBE) for Three Scenes

Table 10
Brown & Daniel (1990) Note: Stimuli 1, 2, 3 are individual scenes and indicates the scores by three observers on the scale 1 – 10. cf = cumulative frequencies, cp = cumulative probabilities and Z is Standard Normal Distribution (Z tables).

While the order of these is the same, the scores given for each landscape differ between observers even though the perceived scenic beauty values are identical for all observers. Table 2 summarizes the derivation of the SBE score using the “by slide” or “by stimulus” method of Daniel & Boster. This method uses multiple observers to rate the scenes and from this, a distribution of ratings for each scene is obtained. The rating distributions are converted to frequencies for each rating category (e.g. 1 – 10) and Z values derived. Daniel and Boster’s alternative method is “by observer,” which uses few observers rating multiple scenes of a given landscape.

Daniel & Boster used the SBE method in relation to forest management and applications included deriving aesthetic contour maps for National Forest areas, examining the effects of timber harvesting on scenic beauty, and identifying the factors that affect the SBE scores. The amount and distribution of felled timber and stumps had a negative effect, while tree density, tree diameter and crown-canopy cover each contributed positively. It has also been used to estimate the scenic effect of beetle damage on trees, the forest factors that contribute to scenic beauty, the changes to scenic beauty over time with forest maturation, tradeoffs between scenic beauty and timber value, effect of distance on scenic beauty, Cultural differences in scenic beauty estimates, effect of scene composition on scenic beauty estimates, and alternative options for the landscape and their impact on scenic beauty.

Because the LCJ and SBE methods both produce interval scale metrics, they do not define absolute scenic values (i.e. no benchmark zero point is available), only relative differences. Thus, the scores derived from different studies of different landscapes cannot be compared as a SBE score of say 60 in one area bears no relation to the same score in another area. However, Hull considers it possible to compare two sets of landscapes that share some scenes (Hull, 1987).

From a user’s viewpoint, it is difficult to convey what a landscape with an SBE of -28 is compared with a second landscape of -7 and a third of +35. Using a 1 – 10 rating scale they can be described as 5, 6 and 8, which communicates far more clearly their relative quality. Used carefully, this scale also enables comparisons of landscapes from different surveys.

While there may be some appeal in the sophistication of the LCJ and SBE methods, it has been suggested that these complex methods do not produce any real gain over simpler methods. Schroeder (1984) tested both complex and simple methods for rating landscape quality and concluded that there was no discernible difference – even the most sophisticated of scaling methods seems to produce results that are equivalent to a simple mean rating.  He commented that rating scales that are often maligned as ‘merely ordinal’, may actually approximate interval scale data more closely than many people suspect. Thus a simple rating scale, such as 1 (low) to 10 (high) as used by the author suffices. Such scales are far more intuitively acceptable than the negative to positive integers that result from the SBE method.

Schroeder also makes the point in view of his research that small samples may be adequate as they produce acceptable intergroup reliability. Typically, researchers have a rule-of-thumb of 15 – 25 raters whereas Schroeder considers that 9 to 15 would be adequate. Prior to Internet-based surveys, the cost of the survey largely related to the number in the sample so this was a significant consideration but is less so now.

The following are research instruments that are generally simpler to use than the LCJ and SBE.

Rating of photographs

Rating of photographs is the single most prevalent method, and its attraction lies in its simplicity and effectiveness. Typically, photographs are presented as prints, as slides, or more commonly now, digital images are presented via the Internet, and respondents rate each scene. The advantage of the Internet is that randomization can be built into the order so that each viewer sees them in a different order. This overcomes issue of a rating of a scene being influenced by the previous scene. It also offsets the problem of fatigue affecting the ratings as the survey can be delayed for a period. Surveys should commence with a few scenes to “prime” the viewer of the range of images to be presented from low to high quality, thereby cueing their brain regarding their rating.

Online survey companies such as Survey Monkey are available with easy-to-use tools to construct the survey, input the images, and collect and even analyze the results. Other instruments include: Question Pro, eSurvey Pro, Zoomerang, Survey Gizmo, Free online survey, Fluid surveys, Qualtrics, Survey Expression, Goodle Consumer Surveys, and Smart-Survey. According to Roth (2006), the “scenic quality categories of visual variety, beauty, visual naturalness as well as overall scenic quality can be validly recorded on the Internet.”

Frome gibber
Frome gibber plain. Rating 3.8   An example of a flat landscape without land cover, land use or water, yet rates nearly 4

Rating scales vary among studies. Some use an odd number, say 1 – 9 so that the mid-point is an integer (5) whereas for the 1 – 10 scale the mid-point is 5.5. Some surveys use a 5 or 7 points. However, the scale needs to provide sufficient discrimination between scenes and scales of 5 and possibly 7 points fail in this. On the other hand, a 1 – 100 scale provides too much discrimination, and it can be difficult for a viewer to discriminate say between 65 and 66. The scale should not include the zero, 0, as there is logically no scene of zero rating.

Even scenes of a flat plain without any land form, land cover, land use, water or any other feature have been found to rate 3 – 4 on a 1 – 10 scale.

The author’s experience is that the 1 – 10 scale works very well; it provides sufficient discrimination and is most easily used by viewers. Most scales go from worst (1) to best (10) as the reverse (best 1 – worst 10), can be confusing for participants and they can revert to the worst-best scenario after a few scenes, which makes data analysis problematic.

Table 11
Table 3 Effect of Internet connections on survey times

Rating scales are used in landscape surveys to rate the aesthetic quality of the landscape as well as many other attributes, including Kaplan’s variables of coherence, complexity, legibility and mystery; tranquillity, naturalness and familiarity. The author’s surveys also assess the visual significance of components in the scenes, including land form, vegetation, land uses, water, diversity and naturalness. These components are rated on a 1 – 5 scale.

The number of photographs used in studies varies from 6 to 180 with the majority in the 40 – 60 range. However, with Internet surveys, longer surveys can be used as the person can pause the survey when fatigued and return to it later.  Table 3 indicates the average time taken to complete four of the author’s surveys. In the early days of the Internet, many respondents used dial-up connections but there was little difference in their times compared with broadband. Figure 4 shows the duration for the Barossa survey.

While rating of scenes enables comparisons of their ratings from a common base, rankings only provide an ordinal hierarchy of scenes, which is difficult to analyze.

Paired photographs

A group of photographs are allocated into pairs and the preference for each photograph in each pair is recorded. This differs from the LCJ method in that each photograph is different, so the sample is not limited to a relatively small number of photographs. Studies have used up to 120 photographs in 60 pairs. The method is particularly popular in Spain – 11 out of the 16 studies using this method were undertaken in Spain by researchers such as Abello, Bernaldez, De Lucio, Marcia and Rodenas.

Q-sort of photographs

The Q-sort procedure was originally developed for personality assessment (the ‘Q’ prefix has no special significance). Based on psychological research, which indicates that the human senses are not capable of discriminating sensory perceptions into more than nine categories (Pitt & Zube, 1979), participants are asked to sort a set of photographs into five or seven piles. Even numbers of piles are avoided so as to permit a central group.

The Q-sort allows a large number of stimuli to be evaluated. An advantage it has over a rating form is that the participant can shift items back and forth as they proceed (Cronbach, 1970). Thus, the photographs in a given pile at the end of the sorting can be regarded as approximately equivalent. Forcing the participant to allocate a pre-set number of photographs to each pile is regarded as being preferable to an unforced choice. Swaffield and Fairweather in New Zealand have used Q sort in several studies.

Visitor Employed Photographs (Participant Photography)

As the name suggests, this technique uses the selection of photographs taken by visitors as an indicator of landscape preferences. The method involves loaning park visitors an inexpensive and easily operated camera, and asking them to take photographs of given subjects. This may be as broad as “anything they wish” (Cherem & Driver, 1983), preferred scenes, anything of interest, or may be used to provide material for use in framing a questionnaire (Hammitt, 1979).

Gabriel Cherem developed the method in the early 1970s as a means of eliciting the public’s view of aesthetic objects. Cherem & Driver (1983) evaluated the method in landscape research, trialing it in three studies. They provided cameras to hikers on a trail (of 512 cameras loaned, only 6 were not returned) and from the hundreds of photographs taken, identified ‘consensus photographs’ (i.e. scenes photographed by 10% or more of the participants). They acknowledge that the 10% figure is arbitrary and could be set higher or lower, but it serves to provide a concrete representation of a scene which offers some common degree of perceptual interest.

Hull & Revell (1989) used the participant photography method in a study in Bali of cross-cultural landscape preferences among the Balinese and Western tourists. Based on the photographs taken, consensus scenes were identified based on 10% of the responses from each culture.

In an evaluation of the method, Chenoweth (1984) concluded that it is a tool that deserves serious consideration along with other tools for understanding people’s reaction to the landscape …

Semantic differential

The semantic differential (SD) is the first of the descriptive methods used in landscape research. Charles Osgood developed the SD technique in the mid-1950s as an objective method of measuring perception, meaning and attitudes. It has been used for assessing the reactions of observers to different types of environmental stimuli, including the architecture of buildings; rooms and interiors; snow, rain, fog and other meteorological phenomena; beach scenes; and roadside scenery

The SD technique is based on the following prerequisites (Osgood & Suci, 1955):

  • Judgement can be made in terms of a continuum, definable by polar terms (i.e. opposites such as like – dislike);
  • The many different ways in which meanings can vary can be represented by a single dimension (e.g. scenic quality as a term covering a landscape’s aesthetic qualities);
  • A limited number of continua can be used to define a quality within which the meaning can be specified.

The SD technique involves participants scoring photographs on a series of bipolar semantic scales, each of which has, say, a 7-point gauge. The scales might be expressed in terms of: common/ unusual, pleasant/untidy, obvious/mysterious, artificial/natural, weak/powerful and barren/fertile.

The SD technique is a mature method that has been used extensively in landscape preference assessment.

Adjective checklist

The second descriptive method is the adjective checklist that has been used for the evaluation of landscapes. In 1972, Kenneth Craik developed a Landscape Adjective Check List (LACL) based on descriptions of 50 natural landscape scenes by students. He asked students to list 10 adjective descriptors of each scene and while not all were able to achieve this, adjectives that were used six or more times comprised a list of 1196 distinct items. The LACL comprised 240 adjectives. He proposed that the list be used to:

  • derive impressions of landscapes quickly from large samples in the field
  • statistically compared descriptions of the same landscapes
  • record impressions of landscape
  • assess change in landscapes
  • evaluate the effectiveness of photographs, sketches and other surrogates of landscapes

Craik (1975) used the list in a field assessment of landscape in Marin County, California and identified 104 adjectives that were used by 10% or more of the participants. It identified the following attributes of the area: clean, hilly, tree-studded, grassy, pleasant, beautiful, natural, green, peaceful, and sunny. Factor analysis was then used to identify four descriptive landscape factors: serene/gentle, dry/barren, beautiful/ picturesque, and blooming/cultivated.

Kane (1976, 1981) in a study of South Australian landscape for the National Trust, developed and applied a bipolar list of 21 adjectives, of which 14 were significant to South Australians as descriptive of their landscape. The adjective pairs included wet/dry, cold/warm, private/public, unstimulating/stimulating, and disordered/order. Responses were transformed into a landscape rating score through application of a weighting factor derived from an earlier evaluation of 40 adjective pairs, and a selection of those which related most to beautiful/ugly and like/dislike. The scoring of adjective pairs was undertaken by ten respondents and applied to 46 scenes throughout South Australia. Checklist scores ranged from a high of 80 down to 29.

The adjective checklist method has not been widely used but can provide an effective and quick method of assessing impressions of a landscape.

Physiological tests  

While psychophysical tests and other preference rating methods assume that human cognitive and affective responses to landscapes can be expressed and measured, physiological tests aim to measure these responses more directly. Physiological effects are autonomic (i.e. self-governing) responses of the human body to environmental stimuli – the subject cannot intentionally create them.

Roger Ulrich who has postulated a landscape theory based on its affective effects has carried out a range of studies using physiological tests. In Ulrich, 1981, he used alpha wave amplitude and heart rate to compare reactions to photographs of scenes of natural and urban environments. In Ulrich et al, 1991, he used a battery of tests: electrocardiogram, pulse transit time (correlates with systolic blood pressure), spontaneous skin conductance responding, and frontalis muscle tension – to assess the rate of recovery from stress from a stressful movie during exposure to videos of natural and urban scenes. There have been many other studies by researchers using physiological tests.

Physiological tests are complex and require specialist equipment and expertise in their administration.

Interviews and questionnaires

Interviews and questionnaires play an important role in landscape research and are often used in association with other methods. About 50% are field administered surveys. They require a large number of respondents and therefore, are expensive in cost and resources required – one survey of 242 averaged only 4 interviews a day – 60 days total.

Measurement of features on photographs

The final method examined is the only one that provides an objective measure of the composition of landscapes as depicted in photographs. It is therefore, the only measure that has been developed of the independent variable (i.e. the landscape).

In 1969, Elwood Shafer, a researcher in the US Forest Service, published a unique approach to measuring landscape preferences by measuring areas and perimeters of features on 8” x 10” black-and-white photographs (Shafer et al, 1969). The photographs were of scenes across the US and included forests, mountains, meadows, water and various combinations. A total of 100 photographs were used. A 1/4” clear plastic grid was overlaid on each photograph, and the areas of landscape zones were then outlined by pen and measured. The 10 landscape zones were

  • sky;
  • vegetation in the foreground, mid-distance and distance;
  • non-vegetation (e.g. exposed ground, mountains, snowfields, grasslands) in the foreground, mid-distance and distance;
  • water – streams, waterfall and lakes.
Fig 13
Shafer, 1969
Figure 5   Example of landscape zones designated on photograph by grid squares

Figure 5 indicates the landscape zones identified in a scene of a lake and distant mountains, framed by trees. Each polygon is identified as set (Sn) that identifies the total squares it contains. Each set is identified by computer, using variables that describe its boundary; the interior number of squares; the area; and the horizontal end-squares. The tonal variations provided by sky, land and water are measured by a photometer and are included in the analysis. Each photograph is described by a total of 46 variables. This was subsequently reduced to 26 zones by removing redundancies.

The photographic evaluations that describe the elements in the landscape provide the independent variables in the research method. The rating of landscape quality provides the dependent variable and was assessed by asking participants to rank the landscape on a 1 – 5 scale. Factor analysis identified nine independent factors, and the model derived used ten terms and explained 65% of the variation in landscape preferences.

Using the model, the predicted scores of the 100 photographs ranged from 84 to 236 which approximated that derived from participants (Table 4).

Table 4   Shafer’s Predictive Model of Landscape Preferences

Table 12

Factors which had a positive influence on the landscape’s aesthetic appeal were the:

  • perimeters of near and middle distant vegetation
  • perimeter of distant vegetation multiplied by the area of water
  • area of middle distance vegetation multiplied by the area of distant non-vegetation
  • area of middle distant vegetation multiplied by the area of water.

The resultant scores are ordinal numbers that enable ranking of photographs.

Shafer has applied the method to several further studies. From an analysis, he suggested that farmland scenes could be improved by

  • eliminating tree cover in the middle distance and by replacing it with fields or pasture
  • establishing a lake
  • permitting vegetation to encroach in the distant zone

For each of these he was able to predict the change to the score that would result (e.g. establishing a lake would improve the score from 155 to 119 – the lower is better).

Shafer’s model was criticized as lacking intuitive appeal since some of the multiplicative independent variables, although mathematically proper, seem illogical (e.g. area of water X area of intermediate veg.) (Buhyoff & Leuschner, 1978). Whittow (1976) suggested that the method was like the well-known analogy of the computer attempting to describe Shakespeare, but he also recognized its worth. The philosopher, Alan Carlson (1977), issued a lengthy critique of the method in which he identifies three key assumptions in the model:

  • the aesthetic quality of the landscape is meaningfully correlated with certain preferences for that landscape;
  • the relevant preferences are those of the general public;
  • the presence of the formalist theme.

Of these, the third is possibly the most telling. Carlson notes that the methodology is “completely formalist” as the methodology measures only formal aspects of photographs – the shapes of the zones, not their contents, or the relationships between the shapes and lines. Formalism derives from the artistic tradition and identifies certain formal aspects of a scene such as shapes, lines, color, patterns, and the formal qualities which they produce such as balance, proportion, unity and diversity.

Bourassa (1991) also identified the formalist basis of Shafer’s approach and was critical of its lack of theoretical origin, stating that the choice of variables is completely without justification (and) do not even seem to make sense intuitively. Bourassa considered the results quite “spurious” as there is no causal link between the independent variables (i.e. the landscape’s formal qualities) and the dependent variable (i.e. preference scores). In another critique, Weinstein is also critical of Shafer’s use of regression analysis:

With enough independent variables a regression equation can be derived that will correlate perfectly with any dependent variable, no matter how meaningless and inappropriate the predictors actually are (Weinstein, 1976).

Despite its critics, Shafer’s approach has been used widely, simply because as it provides an objective basis for measuring the independent variable in landscape research.

Research instruments – Conclusions

The diversity of instruments used in the evaluation of landscape preferences is notable. Although the field of landscape research is relatively new, it has been characterized by considerable innovation and imagination in the application and modification of existing techniques and the development of new ones.

REPRESENTING THE LANDSCAPE

Use of photographs as surrogates Click here
Computer graphics and Geographical Information Systems Click here
Influence of labels on preferences Click here
Viewing Time
Click here

This section examines how landscape quality may be represented in questionnaires and other research instruments.

Use of photographs as surrogates

Most of the research instruments use photographs to represent the landscape as an alternative to taking participants out into the field. Photographs are a surrogate of the landscape, and the rating scale used to assess preferences is also a surrogate of these preferences. One of the earliest areas of inquiry in landscape research was the adequacy of photographs as surrogates. A photograph clearly differs from a field observation as summarized by Table 5.

Table 5 Landscape assessment: field observations vs photographs

Table 13

Overall, the advantages of using photographs easily outweigh those of field observations. Photographs offer rapid, comparatively inexpensive means of assessing landscapes, allowing scenes widely separated in space and time (seasonal) and also changes in landscapes to be assessed. Ratings better represent the disinterested view.

Some of the early studies transported raters into the field: Dearden (1980) transported 12 observers on by mini-bus through the area over two days, Robinson et al (1976) used field methods in surveying the Manchester region of England, and Briggs & France (1981) transported observers through the study area in South Yorkshire. Bernáldez et al, (1988) transported subjects in a coach and invited them, at certain times, to mark on a form their preference of the landscape on the right or left; at the same time photographs of both scenes were taken. To overcome the transport issue, Brush & Shafer (1975) interviewed campers in the area being assessed.

Given the differences between photographs and field observations, it is not surprising that Carlson (1977) stated “It goes without saying that photographs are not landscapes and landscapes are not photographs” Do photographs provide a valid for landscape assessment? Is the assessment of photographs similar to that of field observations?

There have been many studies of this issue: Coughlin & Goldstein, 1970; Zube et al, 1975; Daniel & Boster, 1976; Dunn, 1976; Shuttleworth, 1980, Kellomaki & Savolainen, 1984; Stewart et al, 1984; Trent et al, 1987; Bernáldez et al, 1988, Brown et al, 1988; Hull & Stewart, 1992; Kroh & Gimblett, 1992 and Sevenant & Antrop, 2011.

 A definitive and widely quoted study on the use of photographs as a surrogate of field observations was undertaken by Shuttleworth (1980). He used the same group to assess landscapes in both the field and as photographs and randomized the assessments. He found no significant differences between groups in responses to landscapes in the field and to the photographs. However, he did detect greater differences between responses to black and white photographs and field views than between color photographs and field views. Shuttleworth concluded that the results “indicated that there were very few differences of significance between the reactions to and perceptions of the landscapes either when viewed in the field or as photographs” with any differences being explainable by content. He concluded that photographs can be used, providing they are in color and are wide-angled to provide a lateral and foreground context.

Stamps (1990) used meta-analysis of 1300 papers representing over 150 environments and 2400 respondents to determine a correlation of 0.86 between field and photograph assessments. Photographs have become the preferred means of representing the landscape.

Meitner (2004) compared ratings of individual photographs of the Grand Canyon with 360° panoramas separated into four separate images (orthogonal) which could be scrolled through by the participant before rating them as a set. An additional method was portraying the 360° panorama on the inside of a cylinder which was rotated at a fixed rate and then rated as the entire scene (non-interactive). The cylinder could also be rotated by the participant (interactive). Correlations between the different presentation methods were lowest were for the individual photographs but higher for the other methods (Table 6). The results indicate that using such methods may provide advantages in presenting landscapes, but at the cost of considerable complexity of methods.

Table 6 Correlations between viewing methods

Table 14
Meitner, 2004

In some instances, it may be necessary to draw from existing collections of photographs. Caution is needed in particular to avoid selecting photographs which are well composed, have appealing lighting or clouds, or have people or other extraneous features. The author’s experience is that about 95% of such collections are rejected as not meeting the criteria.

With Photoshop® and similar programs, photographs can be digitally altered, for example to remove extraneous objects. While this can be used to remove electricity poles and the like, such manipulation risks the photograph ceasing to accurately represent the landscape and should therefore be used minimally to edit out unnecessary objects and not to change colors or remove intrinsic features of the landscape. It may, however be used to lighten dark images and to improve their clarity.

Barroso et al, (2012) added features to photographs to assess their influence on preferences by various groups of participants. This ensured that the “variations shown to respondents are adequately controlled in the study and landscape features are easily recognized by the respondents.” Several of the author’s own studies covering the visual impact of wind farms, coastal developments and riverside developments have similarly manipulated the photographs digitally to insert or delete features (Lothian, 2005, 2008, 2009). Prior to digital photography, studies photographed a feature in one landscape and then selected a similar landscape without the feature (e.g. Hull & McCarthy, 1988) but the results were never perfect whereas with digital technology, the same scene can be used.

Computer graphics & Geographical Information Systems

Ian Bishop (2000, 2003) believes that advanced computer graphics is replacing photographs as a means of visual presentation and if this occurs the foregoing criteria remains relevant. Bishop (2011) examined the application of computer game technology – “virtual reality games and their use as collaborative virtual environments” (see also Orland, 2001 regarding the use of Virtual Reality technology). Bishop (2015) projected the use of a phone app to inform users of the landscape through which they traveled. Bishop has written much on the use of computer technology in landscape research (Bishop & Karadaglis, 1996, 1997; Bishop & Stock, 2002, 2006, Ghadirian & Bishop, 2008).  

In Switzerland and England, Lange and Hehl-Lange have used computer graphics to assess the visual changes in landscapes with different management regimes (Lange & Schmidt, 2000; Lange, 2001; Hehl-Lange, 2001; Lange & Hehl-Lange, 2010, Lange & Hehl-Lange, 2011). Lange & Schmidt (2000) used digital 3D virtual landscape images of proposed dams and urban development options in ecological planning.

Kellomäki & Pukkala (1989) used computer graphics comprising tree symbols whose species and size distributions corresponded to those of a forest. Bergen et al (1995) reported reasonable correlation between the mean ratings of photographs and images but low correlations at the individual level. Daniel & Meitner (2001) compared ratings of four images: full color, grayscale, 4-bit color, and black & white sketch. They found the ratings of the grayscale correlated reasonably well with the full color, 0.69, however the 4-bit color less so, 0.43, and the B/W sketch very poorly, 0.04.

The use of computer graphics is an emerging field and with the advances in computer and imaging technologies, the results are far better than in the early years. Further information can be gained from the following publications: de Vries et al, 2007, Lange, 2001.  Lange (2011) provided a review of the development of the Journal, Landscape & Urban Planning’s articles on landscape visualization over the period, 1974 – 2010. The October 2015 (vol. 142) edition of Landscape & Urban Planning was a special issue on Critical Approaches to Landscape Visualization.

Paar (2006) found from a survey in 2000 of over 1000 consultancies and authorities in Germany that while 28% of consultants used 3D visualization software, those who did not cited inadequate computers, lack of technical expertise and cost as their reasons. “Ease of learning” and “interoperability” were regarded as key issues.

Yu et al, (2005) has applied sophisticated Geographical Information Systems and three-dimensional landscape modeling which enables the manipulation of large amounts of physical data.  Smith et al (2009) has similarly employed GIS with scenario software to examine multiple options for forest management in Tasmania, and Brown & Brabyn (2012) used GIS in New Zealand to examine the relationship between multiple landscape values and the character of the physical landscape. Pukkala & Kellomäki (1988) used computer simulation of forest logging as an aid in management. Legge-Smith et al, (2012) used a web based interface, Scenario Chooser, to present a range of hypothetical future forested landscapes to the public with the aim of eliciting the most/least preferred forest management scenarios. Scolozzi et al, (2015) used visitor interviews and GIS to map landscape features in northern Italy according to the number of mentions by interviewees, different value categories and respondent types. The natural scenery of the Valle di Ledro was the topmost landscape value.

Fig 14 Yosemite
Dunkel, 2015        Figure 6 Visual sightlines for Half Dome in Yosemite Valley

Dunkel (2015) used crowd-sourced photos from Flickr to plot the images that people photograph, to examine landscape change, to identify key sightlines for Half Dome in Yosemite (Figure 6), and to identify country of origin of photographers in the Black Forest in southern Germany. This is clearly a technology with enormous potential.

The use of computer generated visual simulations of proposed developments such as wind farms and urban developments was examined by Downes & Lange (2015) who found gaps and deficiencies in the way the environment was portrayed, including the omission of much street furniture, unrealistic camera angles and inaccurate depiction of scale. Similarly, Lovett et al (2015) argued that the urge to use the latest simulation technology should be balanced by a “more comprehensive and critical evaluation of how they are used and the benefits that this may bring.”

Further studies are found in: Bergen et al, 1995, Orland et al, 2001, Germino et al, 2001 and Lim et al, 2006.

Influence of labels on preferences

Several studies have examined the influence that labelled photographs have on preferences and found that they can unduly influence preferences. Anderson (1981) found that labeling scenes as wilderness or a national park raised ratings while economic uses (forest, grazing) lowered them. Hodgson & Thayer (1980) labelled identical scenes with differing labels: lake – reservoir, forest growth – tree farm, pond – irrigation and stream bank – road cut. They also found that the labels implying human influence lowered ratings to 78% of the natural labels, a sizeable difference.

One group of students were told about harvesting practices of a forest including the term clearcut, while another group was not informed about this (Vodak et al, 1985). The ratings of both groups were however very similar. In contrast, a similar study (Simpson et al, 1976) where one group was informed about the forestry management practices and a second group was not, produced marked differences in responses for the clearcut, thinned and natural scenes. In another study, Yeiser & Shilling (1978) used a conservation group and forestry students as subjects plus a control group of non-forestry students and showed them scenes of forest management practices with selected stimulus terms displayed – cull tree, site preparation and charred slash piles. Using galvanic skin response to measure the intensity of emotion among viewers (as used in lie detector tests), surprisingly they found the control group displayed greater concern than the conservation group. It could be that the conservation group had greater knowledge of forestry practices which lessened their reaction.

Overall, these studies indicate that, for whatever reason, appellations given to scenes do affect the responses significantly. They indicate the importance of not coloring responses by suggesting or including anything that will constrain or direct the respondent towards a particular response.

Viewing time

In a significant paper, Feeling and thinking: preferences need no inferences, Robert B. Zanonc (1980) argued against the prevailing doctrine in cognitive psychology that affect is post-cognitive. He provided experimental evidence that discriminations (i.e. like-dislike) can be made in the complete absence of recognition memory.  Ulrich also cited evidence in support of affect being precognitive (Ulrich, 1986). Ulrich et al, (1991) proposed that:

immediate, unconsciously triggered and initiated emotional responses – not ‘controlled’ cognitive responses – play a central role in the initial level of responding to nature, and have major influences on attention, subsequent conscious processing, physiological responding and behavior.

He also suggested that an:

evolutionary perspective implies that adaptive response to unthreatening natural settings should include quick-onset positive affects and sustained intake and perceptual sensitivity.

These views are antithetical to the information processing approach which holds that, although preferences are generated extremely rapidly, they are nevertheless the result of cognitive processing. Lazarus (1982) suggested that Zajonc made the mistake of equating cognition with reality. Lazarus argued that this process occurs outside of conscious awareness and is virtually automatic. He:

regards emotion as a result of an anticipated, experienced, or imagined outcome of an adaptionally relevant transaction between organism and environment (and therefore) cognitive processes are always crucial in the elicitation of an emotion.

Lazarus considers that this approach in no way threatens the basic premises of the evolutionary-adaptional perspective

Twenty years later, Zajonc (2000) revisited the subject and stated:

behavioral, neuroanatomical, and neurophysiological evidence has been found—evidence that is clear and robust— that substantiates many of the suppositions that derive from the original conjecture that ‘preferences need no inferences’.

Herzog has examined this issue in several studies. Herzog (1984, 1985) included scenes which respondents viewed for 20 milliseconds (i.e. 1/50 sec) or 200 milliseconds (i.e. 1/5 second) and compared the responses with 15 seconds. Although there were differences in their ratings, the differences were slight (Figure 7). Nor is the difference in one direction – some are lower and some are higher. The findings are probably insufficient to prove Zanonc correct but it is difficult to comprehend complex cognitive processes being undertaken in as short a space as 1/50 second.

Fig 15 Herzog
Herzog, 1984, 1985
Figure 7 Effect of Viewing Times on Preferences

Wade (1982) examined whether landscape preferences were affected by respondents being given as much time as they desired to view scenes. He found no relationship between preferences and viewing time.

In a study which Korpela et al, (2002) believed provided support for the rapid and automatic affective evaluations of environmental scenes, they timed the interval between a stimulus of urban or nature pictures and a presentation of a vocal expression of joy, anger or emotional neutrality. The reaction times were all less than 0.7 second.

To ensure that participants rely on their affective faculties, the viewing time should be short as longer times encourage analysis of the scene and thereby use of the cognitive faculty. In the past when slides were shown, fixed viewing times were used, generally between 5 and 10 seconds. Now that Internet-based surveys are used, the participant can spend as little or long as they wish although the instructions are to rate the scenes quickly based on their first impressions and not to analyze or think too much about them. This is intended to replicate the way in which they view scenery, rapidly coming to like it or not. Table 10 showed the average time taken was between 5 and 10 seconds which is the time the participants have chosen to spend on each scene.

INFLUENCE OF THE OBSERVER ON PREFERENCES

There are two components to surveys of landscape quality – the landscape and the observer. How does the observer’s background influence their landscape preferences? According to the common saying, familiarity breeds contempt. Does this apply to landscapes? Should experts be used to rate landscapes? What about the influence of an observer’s culture on their preferences? These are the issues that this section examines:

Use of students in surveys Click here
Demographic data collected by surveys
Click here
Influence of respondent characteristics Click here
Influence of cultural background of participants Click here
Do expert participants give different ratings than lay participants? Click here
Influence of familiarity on ratings Click here
Children’s perceptions of landscapes Click here
Influence of personality Click here
Influence of environmental attitudes Click here
Perceptions by different groups
Click here
Reliability of ratings over time Click here
Influence of the observer on preferences – Summary Click here

Use of students in surveys

Because many of the surveys were academic studies in universities, generally tertiary students were used. Of 314 participants in a range of surveys, 143 were university students or staff, 46.6% of the total (Lothian, 2000). Only 37% were from the general community, park visitors or residents. In a study of river environments in Slovenia and Croatia, Stober et al, (2012) found no significance difference between students and experts in their assessment of the river landscapes. From a meta-analysis of 107 studies, Stamps (1999) found a high correlation of 0.83 between students and other respondents and 0.86 where the students served as representatives. However, Tveit (2009) found distinct differences in the rating of visual scale between students and the public.

With concern about “nature deficit disorder” influencing children and students (Louv, 2005), their lack of familiarity with natural landscapes and even fear towards the outdoors could influence their preferences (Aaron & Witt, 2011).

Demographic data collected by surveys

Of 227 surveys examined in a survey, 63% sought no demographic data on their participants, a surprisingly high proportion (Lothian, 2000). Age, gender, education, employment and socio-economic status were the most frequently sought. The author’s surveys generally ask age, gender, education and whether born in Australia as these are attributes for which national statistical data is available. The author’s survey of the Lake District (2013) included birthplace, postcode (to identify which part of the UK they resided), their familiarity with the region, and whether they lived in or near the Lake District.

Influence of respondent characteristics

Table 7 summarizes 17 studies that examined the influence of respondent characteristics – age, gender, education, socio-economic factors – on their preferences. Most of these surveys used the information to check that their sample was representative of the population – they were not generally used to assess the results of the survey.

Table 7 Influence of demographic characteristics (chronological order)

Table 15

In their study of the influence of socio-economic influences on landscape preferences, Tips & Savasdisara (1986d) found high correlations among the groups for age, gender and socio-economic status (see Table 8) and also high correlations among most religions (Table 16), with none of the differences being statistically significant.

Table 8 Spearman rank correlations by religion (Thailand)

Table 16
Tips & Savasdisara, 1986d

Of the demographic factors, only age, and to a lesser extent, gender, exhibited an influence on preferences. Eleven studies showed no difference for age, and five studies showed a slight or weak influence. Banerjee (1977) showed significant influence of age – respondents aged over 25 were more critical of artificial changes to the coastal landscape and more appreciative of natural elements.

In their lifespan analysis of landscape assessment, Zube et al, (1983) found that the preferences of young children (6 – 11 years) differed from adults (Figure 8), being very strongly correlated with the presence of water compared with other age groups. Preferences for other age groups rose with age to middle age and then declined in older age.

Fig 16 Landscape dimensions
Zube, Pitt & Evans, 1983       
Figure 8 Correlations of age groups with scenic ratings of landscape dimensions
Fig 17 Biomes
Lyons, 1983         Figure 9 Preferences of biomes by age group

Lyons (1983) asked children and adults to rate five vegetation biomes on the basis of a place or live or visit (Figure 9). The scale ranged from 1 (extremely undesirable) to 6 (extremely desirable). For most biomes, the ratings of the youngest children (grade 3) were higher than in subsequent ages, the exception being for coniferous forest the preference for which rose in middle age. Preferences for the biomes changed over the life span, being lowest in the elderly. Males and females showed similar preferences.

Stamps (1999) analyzed 107 studies involving 19,000 respondents and over 3,000 environments. He found a correlation of 0.82 between different demographic groups. The correlation for different ethnic groups was 0.87, political affiliation 0.86, gender 0.84, and age (≤12 vs. age > 12) 0.61.

Gender played no influence in twelve studies but did have an influence in four studies:

  • A preference of males for natural landscapes and by females for humanized landscapes (Bernaldez & Parra, 1979)
  • Females were more sensitive to lack of cover in savannah (Woodcock, 1982)
  • Males more likely to view ground, topography & ephemeral objects (Hull & Stewart, 1995)
  • Females were more positive regarding nature than males (Strumse, 1996)

The apparent contradiction between the first and last studies regarding females and natural landscapes may be explained by the first study indicating that females preferred humanized landscapes over natural landscapes if given the choice, but this does not mean that the females disliked the natural landscape.

Preferences were unaffected by education, socio-economic status, religion, occupation, childhood residence or current residence.

To reinforce this, Table 9 and Figures 10a and 10b summarizes the means for a range of respondent characteristics in six studies that I have undertaken.

Table 9 Respondent ratings from Scenic Solutions landscape surveys

Table 17
Lothian, 2005 a & b, 2006, 2009, 2013, 2015
Fig 18a
Figure 10a Respondent characteristics from Scenic Solution’s landscape surveys
Fig 18b
Lothian, 2005 a & b, 2006, 2009, 2013, 2015
Figure 10b Respondent characteristics from Scenic Solution’s landscape surveys –  Scale exaggerated

Overall, providing children are not used as subjects, the basic respondent characteristics of age, gender, education, employment and socio-economic status appear to have a nil or negligible influence on preferences. This again supports the evolutionary view that landscape preferences are innate and therefore, fairly uniform across all humans.

Influence of cultural background of participants

Cultural differences are often regarded as ensuring that there will be wide disagreement across races, nations and cultures in respect of landscape aesthetics. In fact, the opposite is true; the similarities are far greater than the differences. In the 19 studies summarized on Table 10, all but a couple indicate close similarities in the preferences between different cultures. Three of the studies indicate differences; in several studies, these differences were with sub groups of the study.

Table 10 Studies of cultural influence (Chronological order)

Table 18

Similarities were found between:

  • Balinese and Western tourists
  • Koreans and Western tourists
  • Asian and Western tourists
  • Italians and Australians
  • Sherpas and Australians
  • Americans, Virgin Islanders and Yugoslavians
  • Americans, Dutch, Swedes and Danes
  • Americans and Australians
  • Americans and Chinese
  • Americans and Scottish
  • Canary Islanders and tourists
  • British and Australians

Differences were found between:

  • Chinese and American: differences for junior students but similar for older age groups;
  • Americans, Irish and Senegalese: differences for coast, waters, woods & forests, and pastoral landscapes;
  • Dutch and Islamic immigrants: differences regarding wilderness images and non -urban landscapes, especially marshes & dunes.

Hull & Revell (1989) examined two distinctly different cultures, Balinese and Western tourists, yet they concluded that despite the “enormous differences which exist between the Balinese and western culture,” the results suggested “that there was perhaps more similarity than difference between the two groups in their scenic evaluations” of the Balinese landscape. 

Based on the study by Purcell et al, (1994), Figure 11 compares the responses by Italian and Australian students to photographs of landscapes from both countries. Preferences for natural vistas were generally higher among the Italian participants than among the Australian participants, but the differences were only slight.

Fig 19
Purcell et al, 1994
Figure 11 Comparison of Italian and Australian Landscape Preferences

Stamps (1999) carried out a meta-analysis of 107 studies involving over 19,000 respondents and over 3000 scenes. He found an overall correlation of 0.85 for cross-cultural studies involving respondents from different cultural backgrounds. The correlation for ethnic affiliation was even higher, 0.87.

In The Netherlands, Buijs et al, (2009) investigated the concepts of nature held by native Dutch and contrasted these with Islamic immigrants. While the Dutch related strongly to the wilderness image of nature (i.e. ecocentric values and the independence of nature), the immigrants related to the functional image (i.e. anthropocentric values and intensive management). The immigrants had lower preferences for non-urban landscapes, especially wild and unmanaged landscapes such as marshes and dunes. Age, gender and education had little influence.

Beza (2010) asked Australian tourists and Sherpas on the Mt Everest trek in Nepal to rate photographs of the landscapes and of rubbish and garbage along the route. While there was a close correlation in their ratings of the 68 scenes (r2 = 0.91), the Australians tended to rate the high-quality landscapes about 0.9 higher and the poorer-quality landscapes about 0.6 lower (Figure 12). Overall, 80% of the ratings by Australians and Sherpas were within one unit of each other’s ratings.

Fig 20
Beza, 2010. Note: Graph shows the ratings for the same photograph by the two groups
Figure 12 Ratings of Nepal landscapes by Australian tourists and Sherpas

Based on the premise that “culture and landscape interact in a feedback loop in which culture structures landscapes and landscapes inculcate culture,” Nassauer (1995) defined four principles regarding culture and landscape:

  • Human landscape perception, cognition, and values directly affect the landscape and are affected by the landscape.
  • Cultural conventions powerfully influence landscape pattern in both inhabited and apparently natural landscapes.
  • Cultural concepts of nature are different from scientific concepts of ecological function.
  • The appearance of landscapes communicates cultural values.

Overall, these studies indicate that the influence of culture is not as great as might be expected. Acculturation with Western values may be a partial explanation, but is not adequate. It lends support to the evolutionary view that landscape aesthetics are innate and therefore, fairly uniform across all peoples. The differences that occur reflect strong local cultural influences.

Do expert participants give different ratings than lay participants?

A key issue is whether participants should have some form of expertise in landscape (e.g. botanist, geologist, geographer, planner) or whether they should be lay people drawn from the community who are not selected for their expertise. In an early seminal study, Fines (1968) initially used respondents with no design training, but then rejected their ratings in preference to a smaller group with considerable training and experience. His justification of this was twofold: firstly, “such people (i.e. those with training) are most likely to seek and to obtain the greatest enjoyment from landscape” and secondly, the majority may someday aspire to similar values – a justification which appears quaint and elitist by contemporary standards.

Similarly, Carlson (1977) argued that the general public’s judgements of landscape should not provide the basis for the rating of landscapes, and the judgements of more environmentally sensitive people should be used. This would prevent the landscape’s aesthetic qualities being rendered to the lowest common denominator.

The assumption underlying the approach of both Fines and Carlson was that the landscape ratings of the majority would differ from that of the trained minority. Does the evidence support this assumption?

Table 11 lists 20 studies that have examined the differences in preferences between “experts” and lay people. There were ten studies which yielded similar preferences between expert and lay, and ten that yielded different preferences (several studies included sub-studies). That there are equal numbers of studies that find similarities and differences between expert and lay observers suggests that care should be taken in assuming that experts will provide similar ratings of landscape as the community.

Table 11 Studies of preferences by experts and lay (chronological order)

Table 19

Paradoxically, the one professional group whose preferences appear to differ from that of the community is landscape architects. More surveys found that their preferences differed (Anderson & Schroeder, 1983; Brown, 1985; Buhyoff et al, 1978; Miller, 1984, Hunziker et al, 2008, Tveit, 2009) than studies that found similarities (Craik, 1972, Schomaker, 1978 and Yu, 1995). From his meta-analysis of 107 studies, Stamps (1999) found a correlation of only 0.60 between expert and lay. While the preferences of natural resource managers and planners generally corresponded reasonably well with those of the community, the views of landscape architects were often at significant variance with the community. Yeiser & Shilling (1978) found with students that just the connotation of the terms used affected those with no professional knowledge of the terminology.

Influence of familiarity on ratings

The common adage that familiarity breeds contempt does not seem to apply to landscapes. A number of studies, including the author’s, have found that the more familiar a person is with a landscape, the higher their preferences for it.

To study the effect of familiarity on landscape preferences, Wellman & Buhyoff (1980) asked subjects from Virginia and Utah to evaluate mountainous scenes from the Rocky Mountains and Appalachians. They found no regional familiarity effect which may be because photographs of both regions are common and people from across the US are familiar with them.

Adevi & Grahn (2012) in Sweden examined the relationship between landscape preferences and childhood landscapes, where they grew up. This showed the influence of familiarity. They found that people feel more at home in the type of landscape they grew up in and often settled in a similar landscape: Of those born in the:

  • coast, 73% settled in a coastal area;
  • forested landscapes, 63% settled in forested areas;
  • hills and lakes areas, 54% settled in hills and lakes areas;
  • agricultural areas, 52% settled in agricultural areas.

They identified eight “perceived dimensions of experiences in (these) landscapes,” the factors being: species richness, serene, prospect, social, culture, nature, refuge, and space. For example, those who grew up on the coast have a preference for prospect – vast expanses and views without many people, many species and comprising basically the sea, beach, cliffs and rocks. The authors stated that such qualities are tied to innate reflexes related to certain basic needs that are associated with people’s feeling of safety and security. They also found that the larger the town a person grows up in, and the more densely built and lacking in green areas this town is, the lower the person’s identification with the childhood landscape. Familiarity obviously has a major influence on one’s choice of a place to live.

Fig 21
Hammitt, 1979. Arbitrary scale.
Figure 13 Relationship between familiarity and preference

Hammitt (1979) showed photographs of wetlands to visitors before and after their visit and found that generally familiarity increased ratings (Figure 13). However, he also noted that the opposite can occur, a highly familiar scene may rate low preferences – “familiarity per se,” Hammitt said, is an insufficient basis for appreciation. One can be very familiar with non-preferred aspects of an environment.

In several of the author’s studies, participants were asked to indicate their level of familiarity with the region: not familiar, familiar, very familiar and (in some studies) extremely familiar. Figure 14 summarizes the influence of familiarity on the ratings of the landscape. In the coast, Flinders Ranges and Mt Lofty Ranges studies, familiarity had a clearly positive effect on ratings. Being very familiar with the coast increased mean ratings by 4.5% and for the Flinders Ranges by a large 12.5%. For the Mt Lofty Ranges, the effect was smaller, 2.5%. Conversely, however, in the study of the Lake District in England, being familiar with the area actually was found to reduce ratings by nearly 2% while being extremely familiar with it increased them by 2.5%. Overall, the average for the four studies was: somewhat familiar +2.6%, very familiar 4.3%, extremely familiar 2.6%. These figures suggest there may be a threshold of familiarity at the higher level.

Fig 22
Lothian, 2005, 2009, 2013, 2015
Figure 14 Influence of familiarity on mean ratings

Children’s perceptions of landscape

Water is a special part of the play world of the child observed Kates & Katz (1977) in a delightful study of children’s play in a day center and their understanding of the hydrological cycle. They describe the many games involving water – making volcanoes with water spurting up through sandcastle and using water to put out an imaginary fire. Asked where the water from the tap came through elicited many wonderful descriptions involving clouds, pipes, rivers, trucks bringing water from the sea, and rivers under every house that bring the water. Asked where the soapy water goes from the bath, one said it goes back into the river to clean it!

As mentioned earlier, Zube et al, (1983) found that the preferences of young children (6 – 11 years) differed from adults (Figure 17) being very strongly correlated with the presence of water compared with other age groups. Differences in the landscape preferences of children and adults can indicate the influence of acculturation (socialization) on these preferences and the extent to which preferences are inherent or are learnt. This reinforces the finding by Balling & Falk (1982) and Lyons (1983) that the preferences for savanna by children aged 8 – 11 years differed significantly from older children and adults (see Figure 18).

Bernaldez et al, (1987) examined the landscape preferences of children on the Canary Islands. Two age groups were used, 11 and 16 years old. Pairs of photographs were used, and the children asked to indicate their preference. Younger children differed from the older children: they disliked darker scenes with less detail, and they disliked harshness in scenes. The authors linked this with the common fear of darkness among children. The shift in the 11 and 16 years olds in this regard indicates the older children are less influenced by this fear and are more inclined to find it stimulating.

In a study in California of teenager’s valuing of outdoor areas, Owens (1988) found that 70% valued such areas where they could be with nature, 66% where they could get away from other people or be with their friends (30%). Valued outdoor places were places where they could go and view and not be seen. Natural parks and undeveloped agricultural land were their most popular areas, their beauty being their defining characteristic. The teenagers benefited by their proximity to Mt Diablo, a nearby State Park.

Yamashita (2002) gave cameras to adults and 5th & 6th grade (11 years) children to photo-graph a river environment in Japan. He found that while the adults’ photographs emphasized the flow rate and surface conditions of the water, the children’s photographs were more of water quality and flow rate (Figure 15). The children’s photographs also had more water surface than those of the adults.

Fig 23
Yamashita, 2002
Figure 15 Features photographed by adults and children

Tunstall et al (2004) gave cameras to 150 children age 9 – 11 in London and took them to two rivers to take photographs and to record what they thought about them. Over 500 photos were taken. The children focused more on the river environment rather than the rivers themselves which accounted for only 16% of the photos and comments. This may be because the rivers were regarded somewhat negatively, being littered and polluted, with 52% making critical comments and only 29% were positive. Some showed an aesthetic appreciation of the broader river landscape and wide open spaces in choosing to represent ‘views’ of the rivers or flood-plain meadows. They particularly appreciated the trees and plants near the river, partly because they offered opportunities for play. Parts of the river were seen to be dangerous because of steep slippery banks, and the children wanted cleaner, safer, more accessible rivers and varied, adventurous and manipulable play opportunities. Figure 16 summarizes their evaluation of the rivers.

Fig 24
Tunstall et al, 2004
Figure 16 Evaluation of two London rivers by children

In a study of agricultural landscapes on the Veneto Plain in Italy, Tempesta (2010) found that children rated the scenes higher than students or adults: mean children 7.83, students 6.17, adults 6.88.

In a South African study, Adams & Savahl (2013) found that children perceived the natural environment through the lens of safety as natural areas in their community are characterized by crime, violence, and pollution.

An assessment of the influence of nature on children’s self-discipline in high-rise apartments in Chicago (Taylor et al, 2002) found that the more natural the view was from the girl’s home, her self-discipline improved by about 20%; however, for boys no relationship was detected.

In a meta-analysis of studies involving children, Stamps (1999) found an overall correlation of only 0.61 between children (≤12 years) and people over 12 years.

The few studies that have included children indicate that their landscape preferences differ significantly from adults. In their high preference for water and savannah there are suggestions of an evolutionary influence. Children were particularly perceptive about water cleanliness. Generally, they are positive about nature and natural landscapes.

Influence of personality

Spanish researchers have examined the influence of personality on preferences. The research design involved use of paired photographs of scenes together with a personality test to identify personality types. Factor analysis was used to identify the differences. Maciá (1979) separated the results for male and female. For men, he found:

  • men with mature personalities who dealt with reality prefer humanized landscapes;
  • men who score high in emotional control prefer pleasant landscapes;
  • extroverted men prefer landscapes with diffuse forms and rounded trees.

For women, Maciá found:

  • women with a sensitive, insecure personality prefer natural, unaltered landscapes;
  • women with astute, worldly personalities prefer dry, cold landscapes;
  • extroverted women prefer landscapes with diffuse forms and rounded trees.

Maciá concluded that personality structure conditions landscape choice, and gender can influence preference, either directly or be influenced by personality factors. It is worth noting a reversal of a previous finding by Bernaldez & Parra (1979) that males preferred natural landscapes while females preferred humanized landscapes.

Abello & Bernaldez (1986) found that individuals having low emotional stability prefer landscapes exhibiting “recurrent patterns” and “structural rhythms or patterns.” They favor environmental regularity and avoid vegetation spontaneity and vigor. They reject hostile, cold, wintry scenes with defoliated vegetation, although the same scenes are more legible and generally appreciated.

These results provide tantalizing indications of the influence of personality upon landscape preferences.

Influence of environmental attitudes

In southern Norway, Kaltenborn & Bjerke (2002) examined the influence of environmental value orientations – anthropocentric, ecocentric and apathetic – on landscape preferences. Their survey found close positive correlations between the ecocentric orientation and preference for wildlands with water, and for cultural landscapes. The anthropocentric orientation linked with a preference for farm landscapes. In contrast, environmental apathy was negatively linked with wildlands and cultural landscapes.

Perceptions by different groups

Do different groups in the community rate the landscape differently? Several studies have examined suggested they do.

Zube (1974) asked a group of 30 environmental designers (landscape architects, planners, architects) to participate in a field-based study and a group of 30 resource managers (foresters, wildlife managers, hydrologists, etc.) to participate in an office-based study. The study involved using a semantic differential scale of 25 items (e.g. simple – complex, hard – soft, unity – variety) to describe the landscape, write a free description of the landscape, and rank scenes of aerial photographs of landscapes. He found a reasonable level of agreement between the two groups in their evaluation and description of scenic resources. Correlations between the groups in their semantic descriptions of landscapes averaged 0.80 – 0.88. Descriptions of land form, landscape materials or features, and land use consistently dominated in their free descriptions of landscapes.

Dearden (1984) asked planners, Sierra Club members and the community to evaluate scenes of peri-urban, rural and wilderness. While finding no difference on the basis of age, gender, income, education or occupation, he found that wilderness scenes were evaluated very differently by Sierra Club members compared with the other groups. He also found that those living in low density housing felt more positively about the rural and wilderness scenes than respondents from high-density housing.

Brush (1979) asked private owners of commercial forest land in Massachusetts about the scenic attractiveness of woodlands and compared their ratings with students with and without forestry training. He found that the landowners preferred large, enclosed spaces and spaces created by thinning stands of trees than dense overstocked stands. A study was made of the ability of physical, artistic, and psychological descriptor dimensions to predict aesthetic preferences for rural river, forest, and agricultural landscape scenes. Some descriptors were effective in predicting preference across a range of landscape types, while others were effective within a particular landscape type.

DSC_0798C
Swiss alpine landscape. Grindlewald & Wetterhorn

In central alpine Switzerland, Hunziker (1995) investigated the perception by tourists and residents of reafforestation of abandoned farmlands.  He found that most people prefer “partially reafforested landscapes with a high diversity. Partial ingrowth of forest into an agricultural landscape is even assessed as an improvement of its visual quality. However, if the resulting forest patches become too big and homogeneous, a negative feedback can be expected.”

Holland 049PSC
Dutch landscape

In Holland, Van den Berg et al, (1998) asked three groups, farmers, residents (non-farmers) and visiting cyclists, to rate agrarian landscapes and computer simulations of nature development plans. Interestingly, they found that the beauty ratings of residents and visitors were positively related to typical characteristics of nature development plans (wetness, roughness and non-cultivatedness), while famers’ beauty ratings were negatively related to these characteristics.

Gómez-Limón et al (1999) found in a study in central Spain where change of use has resulted in widespread revegetation of the landscape, that the livestock farmers preferred open landscapes in contrast with recreationists and environmental managers who preferred denser vegetation.

Brush et al, (2000) examined whether different groups view the landscape differently. They had dairy farmers, professional foresters and logging contractors, groups that earn their living from the land, together with lake association members and tourists, view videos of driving along Wisconsin highways and rate the enjoyability of driving through each landscape type. They found significant differences between the groups, but only their reported knowledge regarding land management had a significant impact on their preferences.

Williams & Carey (2002) asked urban and rural groups in Australia to rate various vegetation associations. They found close similarity in their ratings, the difference averaging just 3.4%.

Dramstad et al, (2006) asked locals and non-local students to evaluate scenes of the Norwegian landscape, including preferences, and found nearly identical mean preference scores: locals 3.2, students 3.25.

In Flanders, Rogge et al, (2007) compared the perception of rural landscapes among farmers, landscape experts and the general public. They found the groups looked at landscapes in different ways, attaching importance to different landscape features and funding different functions appropriate for the considered landscape. In regard to landscape preferences, the following factors were important to each group:

  • Farmers: openness and maintenance of the landscape;
  • Experts: vegetation and openness of the landscape;
  • General public: appearance of vegetation, the openness and maintenance of the landscapes.

In Japan, Natori & Chenoweth (2008) asked farmers and naturalists to rate rice paddies and woodland landscapes. For the rice paddy landscapes, perceptions of stewardship and openness were much more important to the farmers, while to the naturalist, naturalness and biodiversity were more important. No difference was found in the two group’s perception of the woodlands. 

From these few studies, it is evident that people with a stake in the land, particularly employment that is derived from the land, see landscape beauty in terms of productivity rather than aesthetics. A farmer told the author that he rated the landscapes of farmland according to how high the crops were! It is not that these people are blind to the aesthetics but rather that their priority and focus is on making a living from the land. The surveys involved land in which they derived a living; had landscapes elsewhere been used, they would probably have rated them as aesthetic objects.

Reliability of ratings over time

The reliability of observer responses has been assessed by examining the extent to which they change over time.

Coughlin & Goldstein (1970) examined the consistency of ratings one month after the initial rating. While they found a reasonably good correlation of 0.73 between the two ratings, this seems rather low for a gap of only one month. 

Hull & Buhyoff (1984) reassessed preferences after the elapse of more than twelve months. Individual observer reliability averaged nearly 80% while group consensus values were very reliable (r = 0.956). The authors recommended that group data be used in preference to individual responses.

The reliability of preferences was tested by Cook & Cable (1995) in rating of shelterbelts in the Great Plains of Kansas where some respondents repeated the rating, time interval unstated. They found the ratings of both sessions highly correlated (0.90).

In Denmark, Jensen (1999) reviewed surveys in 1976 – 78 of the public who had recreated in the forests and compared them with surveys in 1993 – 95. He found that the forest preferences were generally stable over this period, although some changes were noted.

In a study in Massachusetts, Palmer (2004) found that because of land use change from forestry to residential, the overall landscape quality had decreased. Although the landscape had changed, he found that the landscape scenic norms were relatively unchanged, with a preference for natural appearing landscapes with a mosaic of open and forested land. The model of using spatial landscape metrics retained its predictive efficacy after 20 years.

These few studies suggest a reasonable constancy of landscape preferences over time.

Summary – Observer influences on landscape preferences

This section examined the influence of the observer upon their landscape preferences. It found that although many studies used tertiary students as their subjects, there is a high correlation (0.83) between their preferences and that of other respondents. Surprisingly, nearly two-thirds of studies collected no demographic data on their subjects.  Of the demographic factors, only age and to a lesser extent, gender, had an influence on preferences. Landscape preferences of young children differed from adults, they have particularly high preferences for water and savannah landscapes. Providing children are not used as subjects, the characteristics of age, gender, education, employment and socio-economic status had nil or negligible influence on preferences.

It is often believed that cultural differences will result in wide differences in landscape preferences, but in fact, the similarities are far greater than the differences. Experts in landscape are sometimes used as subjects instead of lay people with no expertise but studies indicate that there can be differences between expert and lay (correlation only 0.60), and that care should be taken in assuming that experts will provide similar ratings as the community. In particular, while the preferences of natural resource managers and planners were reasonably close to those of the community, the ratings of landscape architects often varied significantly from those of the community.

While it is commonly believed that familiarity breeds contempt, in respect of landscapes, a common finding is that the more familiar a person is with a landscape, the higher their preferences for it. A few studies have suggested that the subject’s personality influences their landscape preferences, but the evidence is thin. Surveys involving different groups in the community have found that those with a stake in the land, particularly employment that is derived from the land, see landscape beauty in terms of productivity rather than aesthetics. Finally, longitudinal studies have found remarkable constancy in people’s landscape preferences in repeated viewings and also over time, even as long as 20 years.

REFERENCES

Aaron, R.F. & P.A. Witt, 2011. Urban students’ definitions and perceptions of nature. Children, Youth & Envs., 21:2, 145 – 167.
Abello, R.P. & F.G. Bernaldez, 1986. Landscape preference and personality. Landscape & Urban Plg., 13, 19 – 28.
Acar, C. & C. Sakici, 2008. Assessing landscape perception of urban rocky habitats. Building & Env., 43. 1153- 1170.
Adams, S. & S. Savahl, 2013. Children’s perceptions of the natural environment: A South African perspective. Children’s Geographies, 11.
Adevi, A.A. & P. Grahn, 2012. Preferences for landscapes; a matter of cultural determinants or innate reflexes that point to our evolutionary background? Landscape Research, 37:1, 27 – 49.
Anderson, E., 1978. Visual Resource Assessment: Local Perceptions of Familiar Natural Environments. PhD (Natural Resources) dissertation, University of Michigan.
Anderson, L.M., 1981. Land use designations affect perception of scenic beauty in forest landscapes. Forest Science, 27:2, 343 – 400.
Anderson, L.M. & H.W. Schroeder, 1983. Application of wildland scenic assessment methods to the urban landscape. Landscape Planning, 10, 219 – 237.
Appleton, J., 1975. Landscape evaluation: the theoretical vacuum. Trans. Inst. Brit. Geog., 66, 120 – 123.
Balling J.D. & J.H. Falk, 1982, Development of visual preference for natural environments. Env. & Behav., 14:1, 5 – 28.
Banerjee, T., 1977. Who values what? Audience reactions to coastal scenery. Landscape Architecture, 65:3, 240 – 243.
Barroso, F.L, T. Pinto-Correia, I. L. Ramos, D. Surová & H. Menezes, 2012. Dealing with landscape fuzziness in user preference studies: Photo-based questionnaires in the Mediterranean context. Landscape & Urban Plg., 104:3-4, 329 – 342.
Beckett, P. H. T., 1974. The interaction between knowledge and aesthetic appreciation. Landscape Research, 1:8, 5 – 9.
Bergen, S.D., C.A. Ulbricht, J.L. Fridley & M.A. Ganter, 1995. The validity of computer-generated graphic images of forest landscape. Jnl. Env. Psych., 15:2, 135 – 146.
Bernáldez, F.G. & F. Parra, 1979. Dimensions of landscape preferences from pairwise comparisons. In: Elsner & Smardon (Eds), 256 – 262.
Bernáldez, F.G., D. Gallardo & R.P. Abello, 1987. Children’s landscape preferences: from rejection to attraction. Jnl. Env Psych, 7, 169 – 176.
Bernáldez, F. G., J. P. Ruiz, J. Benayas & R. P. Abelló, 1988. Real landscapes versus photographed landscapes: preference dimensions. Landscape Research, 13:1, 10 – 11.
Beza, B., 2010. The aesthetic value of a mountain landscape: A study of the Mt. Everest Trek. Landscape & Urban Plg, 97, 306–317.
Bishop, I.D. & C. Karadaglis, 1996. Combining GIS based environmental modeling and visualization. www.ncgia.ucsb.edu
Bishop, I.D. & C. Karadaglis, 1997. Linking modeling and visualization for natural resources management. Env & Plg B., 24:3, 345 – 358.
Bishop, I.D., 2000. Environmental information and technology: is it all too much? Landscape & Urban Plg., 47:3-4, 111 – 114.
Bishop, I.D., J-A.R. Wherrett & D.R. Miller, 2000. Using image depth variables as predictors of visual quality. Env. & Plg. B., 27:6, 865 – 875.
Bishop, I.D. & C. Stock, 2002. Immersive, interactive exploration of changing landscapes. Int. Modeling, 30 – 35.
Bishop, I.D. & B. Rohrmann, 2003. Subjective responses to simulated and real environments: a comparison.  Landscape & Urban Plg., 65, 261 – 277.
Bishop, I.D. & C. Stock, 2006. Linking GIS with real-time visualization for exploration of landscape changes in rural community workshops. Virtual Reality, 9:4, 260 – 270.
Bishop, I.D., 2011. Landscape planning is not a game: Should it be? Landscape & Urban Plg., 4:4, 390 – 392.
Bishop, I.D., 2015. Location based information to support understanding of landscape futures. Landscape & Urban Plg., 142, 120 – 131.
Bourassa, S.C., 1991. The Aesthetics of Landscape. Belhaven Press, London.
Briggs, D. J. & J. France, 1980. Assessing landscape attractiveness: A South Yorkshire study. Landscape Research, 6:2, 2 – 5.
Brown, G. & L. Brabyn, 2012. An analysis of the relationships between multiple values and physical landscapes at a regional scale using public participation GIS and landscape character classification.  Landscape & Urban Plg., 107, 317 – 331.
Brown, S., 1985. Landscape evolution – anathema or essential. Landscape Design, 157, 9 – 11.
Brown, T.C. & T.C. Daniel, 1987. Context effects in perceived environmental quality assessment: scene selection and landscape quality ratings. Jnl. Env. Psych., 7, 233 – 250.
Brown, T.C., T.C. Daniel, M.T. Richards & D.A. King, 1988. Recreation participation and the validity of photo-based preference judgements. Jnl. Leisure Research, 20:4, 40 – 60.
Brown, T.C. & T.C. Daniel, 1990. Scaling of Ratings: Concepts and Methods. Rocky Mountain Forest and Range Experiment Station, Fort Collins, Colorado. Research Paper Rm-293.
Brush, R.O. & E.L. Shafer, 1975. Application of a landscape-preference model to land management. In: Zube, E.H. et al, 1975.
Brush, R.O. 1979. The attractiveness of woodlands: perceptions of forest landowners in Massachusetts. Forest Science, 25:3, 495 – 506.
Brush, R., R.E. Chenoweth & T. Barman, 2000. Group differences in the enjoyability of driving through rural landscapes. Landscape & Urban Plg., 47:1 – 2, 39 – 45.
Buhyoff, G.J. & W.A. Leuschner, 1978. Estimating psychological disutility from damaged forest stands. Forest Science, 24:1, 424 – 432.
Buhyoff, G.J., J.D Wellman, H. Harvey & R.A. Fraser, 1978. Landscape architects’ interpretations of landscape preferences. Jnl. Env. Mgt., 6, 255 -262.
Buhyoff, G.J., J.D Wellman, N.E. Koch, L. J. Gauthier & S. Hultman, 1983. Landscape preference metrics: an international comparison. Jnl. Env. Mgt., 16, 181 – 190.
Buijs, A.E., B.H.M. Elands & F. Langers, 2009. No wilderness for immigrants: cultural differences in images of nature and landscape preferences. Landscape & Urban Plg., 91:3, 113 – 123.
Calvin, J.S., J.A. Dearinger & M.E. Curtin, 1972. An attempt at assessing preferences for natural landscapes. Env. & Behav., 4:4, 447 – 470.
Carlson, A.A., 1977. On the possibility of quantifying scenic beauty. Landscape Plg., 4: 131–172.
Chapman, V., 1974. Reactions to developments in the countryside in terms of landscape change. Landscape Research News, 1:7, 4 – 7.
Chenoweth, R., 1984. Visitor employed photography: A potential tool for landscape architecture. Landscape Journal, 3:2, 136-143.
Cherem, G.J. & B.L. Driver, 1983. Visitor employed photography: a technique to measure common perceptions of natural environments. Jnl.  Leisure Research, 15:1, 65 – 83.
Cook, P.S. & T. T. Cable, 1995. The scenic beauty of shelterbelts on the Great Plains. Landscape & Urban Plg., 32:1, 63-69.
Coughlin, R.E. & K.A. Goldstein, 1970. The Extent of Agreement Among Observers on Environmental Attractiveness. RSRI Discussion Paper Series No 37.
Craik, K.H., 1972. Appraising the objectivity of landscape dimensions. In: J.V. Krutilla (Ed), Natural Environments, Studies in Theoretical and Applied Analysis. Resources for the Future, John Hopkins University Press.
Craik, K.H., 1975. Individual variations in landscape description. In: Zube, E.H., R.O. Brush & J.G. Fabos (Eds).
Cronbach, L.J., 1970. Essentials of Psychological Testing. Harper International Edition, New York.
Daniel, T.C., Wheeler, L., R.S. Boster & P.R. Best, 1973. Quantitative evaluation of landscapes: an application of signal detection analysis to forest management alternatives. Man – Env. Systems, 3:5, 330 – 344.
Daniel, T.C. & R.S. Boster, 1976. Measuring Landscape Esthetics: The Scenic Beauty Estimation Method. USDA Forest Service Research Paper RM-167.
Daniel, T.C., L.M. Anderson, H.W. Schroeder & L. Wheeler III, 1978. Mapping the scenic beauty of forest landscapes. Leisure Sciences, 1:1, 35 – 52.
Daniel, T.C. & M.M. Meitner, 2001. Representational validity of landscape visualizations: the effects of graphical realism on perceived scenic beauty of forest vistas. Jnl. Env. Psych.., 21:1, 61 – 72.
De Vries, S., R-K. J. Lankhorst & A.E. Buijs, 2007. Mapping the attractiveness of the Dutch countryside: a GIS-based landscape appreciation model. Forest Snow & Landscape Research, 81:1-2, 43 – 58.
Dearden, P., 1980. Landscape assessment: the last decade. Canadian Geog., 24:3, 316 – 325.
Dearden, P. 1984. Factors influencing landscape preferences: an empirical investigation. Landscape Plg., 11, 293 – 306.
Dearden, P. & B. Sadler (Eds), 1989. Landscape Evaluation: Approaches and Applications. University of Victoria, BC.
Díaz, P., D. Ruiz-Labourdette & A.R. Darias, 2010. Landscape perception of local population: the relationship between ecological characteristics, local society and visitor preferences. In: C. A. Brebbia, F. D. Pineda (Eds), Sustainable Tourism IV. 4th International Conference on Sustainable Tourism, Ashurst, New Forest, England, 5-7 July 2010. WIT Press.
Downes, M. & E. Lange, 2015. What you see is not always what you get: A qualitative, comparative analysis of ex-ante visualizations with ex-post photography of landscape and architectural projects.  Landscape & Urban Plg, 142, 136 – 146.
Dramstad, W.E., M.S. Tveit, W.J. Fjellstad & G.L.A. Fry, 2006. Relationships between visual landscape preferences and map-based indicators of landscape structure. Landscape & Urban Plg., 78:4, 465-474.
Dunkel, A., 2015. Visualizing the perceived environment using crowdsourced photo geodata. Landscape & Urban Plg., 142, 173 – 186.
Dunn, M.C., 1976. Landscape with photographs: testing the preference approach to landscape evaluation. Jnl. Env. Mgt., 4, 526.
Dupont, L, M. Antrop & V. van Eetvelde, 2015. Does landscape related expertise influence the visual perception of landscape photographs? Landscape & Urban Plg, 141, 68 – 77.
Elsner, G.H. & R.C. Smardon (Eds), 1979. Proceedings of Our National Landscape Conference on Applied Techniques for Analysis and Management of the Visual Resource, 23 – 25 April, 1979, Incline Village, Nevada.
Fines, K.D., 1968. Landscape evaluation: a research project in East Sussex. Regional Studies, 2, 41 – 55.
Germino, M.J., W.A. Reiners, B.J. Blasko, D. McLeod & C.T. Bastian, 2001. Estimating visual properties of Rocky Mountain landscape using GIS. Landscape & Urban Plg, 53:1-4, 71-83.
Ghadirian, P. & I.D. Bishop, 2008. Integration of augmented reality and GIS: A new approach to realistic landscape visualization.  Landscape & Urban Plg, 86:3 – 4, 226 – 232.
Gómez-Limón, J. & J.V. de Lucio Fernánde, 1999. Changes in use and landscape preferences on the agricultural-livestock landscapes of the central Iberian Peninsula (Madrid, Spain). Landscape & Urban Plg, 44:4, 165-175.
Green, D.M. & J.A. Swets, 1966. Signal Detection Theory and Psychophysics. John Wiley, New York.
Guildford, J.P. 1954. Psychometric Methods. McGraw-Hill, New York.
Hammitt, W.E., 1979, Measuring familiarity for natural environments through visual images. In: Elsner & Smardon (Eds), 1979.
Hart, W.J. & W.W. Graham, 1967. How to rate & rank landscape. Landscape Arch., 57:2, 121 – 122.
Hehl-Lange, S., 2001, Structural elements of the visual landscape and their ecological functions. Landscape & Urban Plg, 5, 105 – 113.
Hepburn, R.W., 1963, Aesthetic appreciation of nature. British Jnl.  Aesthetics, 35/10/20143, 195 – 209.
Herzog, T.R., 1984. A cognitive analysis of preference for field-and-forest environments. Landscape Research, 9, 10 – 16.
Herzog, T.R., 1985. A cognitive analysis of preference for waterscapes. Jnl. Env. Psych., 5, 225 – 241.
Herzog, T.R., E.J. Herbert, R. Kaplan & C. Crooks, 2000. Cultural & development comparison of landscape perceptions & preferences. Env. & Behav., 32:3, 323 – 346.
Hodgson, R.W. & R.L. Thayer, 1980. Implied human influence reduces landscape beauty. Landscape Plg, 7, 171 – 179.
Hull, R.B., G. J. Buhyoff & T.C. Daniel, 1984. Measurement of scenic beauty: The Law of Comparative Judgement and scenic beauty estimation procedures. Forest Science, 30:4, 1084 – 1096.
Hull, R.B. & G. J. Buhyoff, 1984. Individual and group reliability of landscape assessments. Landscape Plg., 11, 67 – 71
Hull, R.B., 1986, Sensitivity of scenic beauty assessments. Landscape & Urban Plg., 13, 319 – 321.
Hull, R.B., 1987. Interpreting scenic beauty estimates. Landscape Jnl., 4 – 27.
Hull, R.B. & M.M. McCarthy, 1988, Change in the landscape.  Landscape & Urban Plg, 15, 265 -278.
Hull, R.B. & G.R. Revell, 1989, Cross-cultural comparison of landscape scenic beauty evaluations: a case study in Bali. Jnl. Env. Psych.., 9, 177 – 191.
Hull, R.B. & W.P. Stewart, 1992. Validity of photo-based scenic beauty judgements. Jnl. Env. Psych.., 12, 101 – 114.
Hull, R.B. & W.P. Stewart, 1995. The landscape encountered and experienced while hiking. Env. & Behav., 27:3, 404 – 426.
Hunziker, M., 1995. The spontaneous reafforestation in abandoned agricultural lands: perception and aesthetic assessment by locals and tourists. Landscape & Urban Plg., 39:1-3, 399 – 410.
Jackson, J. B. 1999. Landscape in Sight: Looking at America. Yale University Press.
Jensen, F. S., 1993. Landscape managers’ and politicians’ perception of the forest and landscape preferences of the population. Forest & Landscape Research, 1:1, 79-93.
Jones, G.R., 1978. Landscape Assessment … where logic and feelings meet. Landscape Arch, 113 – 116.
Kaltenborn, B.P. & T. Bjerke, 2002. Associations between landscape preferences and place attachment. A study in Røros, Southern Norway. Landscape Research, 27:4, 381-396.
Kane, P.S., 1976. Evaluation of Landscape Attractiveness – a Review of Problems and Methods, and a Technique Developed for the National Trust of South Australia. National Trust of South Australia.
Kane, P.S., 1981. Assessing landscape attractiveness: a comparative test of two methods. Applied Geography, 1, 77 –96.
Kaplan, R. & E.J. Herbert, 1987. Cultural and sub-cultural comparisons in preferences for natural settings. Landscape & Urban Plg, 14, 281 – 293.
Kates, R.W. & C. Katz, 1977. The hydrologic cycle and the wisdom of the child. Geog. Review, 67:1, 51 – 62.
Kearney, A.R. & G.A. Bradley, 2011. The effects of viewer attributes for forest scenes: contributions of attitudes knowledge, demographic factors, and stakeholder group membership. Env. & Behav., 43:2, 147 – 181.
Kellomaki, S. & R. Savolainen, 1984. The scenic value of the forest landscape as assessed in the field and the laboratory. Landscape Plg., 11, 97 – 107.
Kellomäki, S. & T. Pukkala, 1989. Forest landscape: A method of amenity evaluation based on computer simulation. Landscape & Urban Plg., 18:2, 117 – 125.
Korpela, K.M, T. Klemettilä & J.K. Hietanen, 2002. Evidence for rapid affective evaluation of environmental scenes. Env. & Behav., 34:5, 634 – 650.
Kroh, DP. & R.H. Gimblett, 1992. Comparing live experience with pictures in articulating landscape preference. Landscape Research, 17:2, 58 – 69.
Lamb, R.J. & A.T. Purcell, 1990. Perception of naturalness in landscape and its relationship to vegetation structure. Landscape & Urban Plg., 19, 333 – 352.
Lange, E. & W.A. Schmid, 2000. Ecological planning with virtual landscapes: three examples from Switzerland. Landscape Research, 19:1-1, 156 – 165.
Lange, E., 2001. The limits of realism: perceptions of virtual landscapes. Landscape & Urban Plg., 54, 163-182.
Lange, E. & S. Hehl-Lange, 2010. Making visions visible for long-term landscape management. Futures, 2, 693 – 699.
Lange, E. & S. Hehl-Lange, 2011. Citizen participation in the conservation and use of rural landscapes in Britain: The Alport Valley case study. Landscape Ecol Eng., 7, 223 – 230.
Lange, E., 2011. 99 volumes later: We can visualize. Now what?  Landscape & Urban Plg., 100:4, 403 – 406.
Lazarus, R.S., 1982. Thoughts on the relations between emotion and cognition. American Psychologist, 37:9, 1019 – 1024.
Legge Smith, E., I.D. Bishop, K.J.H. Williams & R. M. Ford, 2012. Scenario Chooser: an interactive approach to eliciting public landscape preferences. Landscape & Urban Plg., 106:3, 230- 243.
Lim, E.M., T Honjo & K Umeki, 2006. The validity of VRML images as a stimulus for landscape assessment. Landscape & Urban Plg., 77:1-2, 80 – 93.
Lindzey, G., R.F. Thompson & B. Spring, 1988. Psychology. Worth Publishing.
Linton, D.L., 1968. The assessment of scenery as a natural resource. Scot. Geog. Mag., 84:3, 219 – 238.
Lovett, A., K. Appleton, B. Warren-Kretzschmar & C. Von Haaren, 2015. Using 3D visualization methods in landscape planning: An evaluation of options and practical issues. Landscape & Urban Plg.,142, 85 – 94.
Louv, R. 2005. Last Child in the Woods: Saving Our Children from Nature-Deficit Disorder. Algonquin Books.
Lowenthal D., 1962. Not every prospect pleases: What is our criterion for scenic beauty? Landscape, 12:2, 19 – 23.
Lyons, E., 1983. Demographic correlates of landscape preference. Env. & Behav., 15:4, 487 – 511.
Macaulay Land Use Research Institute, 2013. Review of Existing Methods of Landscape Assessment and Evaluation. (online).
Macia, A., 1979. Visual perception of landscape: sex and personality differences. In: Elsner, G.H. & R.C. Smardon (Eds), 1979.
Meitner, M.J., 2004 Scenic beauty of river views in the Grand Canyon: relating perceptual judgments to locations. Landscape & Urban Plg., 68:1, 3-13.
Miller, P.A., 1984. A comparative study of the BLM Scenic Quality Rating Procedure and landscape preference dimensions. Landscape Jnl., 3:2, 123 – 135.
Morgan, R. 1999. Some factors affecting coastal landscape aesthetic quality assessment. Landscape Research, 24:2, 167-184.
Mosley, M.P., 1989, Perceptions of New Zealand river scenery. New Zealand Geographer, 45:1, 2 – 13.
Natori, Y. & R. Chenoweth, 2008. Differences in rural landscape perceptions and preferences between farmers and naturalists. Jnl. Env. Psych., 28:3, 250 – 267.
Newell, P.B., 1997. A cross-cultural examination of favorite places. Env. & Behav., 28:4, 495 – 514.
Orland, B., K. Budthimedhee & J. Uusitalo, 2001, Considering virtual worlds as representations of landscape realities and as tools for landscape planning. Landscape & Urban Plg., 54:1, 139 – 148.
Osgood, C.E. & G.J. Suci, 1955. Factor analysis of meaning. Jnl. Experimental Psychology, 50:5, 325 – 338.
Owens, P.E., 1988. Natural landscapes, gathering places, and prospect refuges: characteristics of places valued by teens. Children’s Environments Quarterly, 5:2, 17 – 24.
Paar, P., 2006. Landscape visualizations: Applications and requirements of 3D visualization software for environmental planning. Comp. Env. & Urban Systems, 30:6, 815–839.
Palmer, J.F., 2004. Using spatial metrics to predict scene perception in a changing landscape, Dennis, Mass.  Landscape & Urban Plg., 69:2/3, 201 – 218.
Penning-Rowsell, E.C., 1982. A public preference evaluation of landscape quality. Regional Studies, 16:2, 97 – 112.
Pitt, D.G. & E. H. Zube, 1979. The Q-sort method: Use in landscape assessment research and landscape planning. In: Elsner & Smardon (Eds), 1979.
Pomeroy, J.W., J.E. FitzGibbon & M.B. Green, 1989. The use of personal construct theory in evaluating perceptions of landscape aesthetics. In: Dearden, P. & B. Sadler (Eds), 1989.
Porteous, J.D., 1982. Approaches to environmental aesthetics. Jnl. Env. Psych., 2, 53 – 66.
Pukkala, T. & S. Kellomäki, 1988. Simulation as a tool in designing forest landscape.  Landscape & Urban Plg., 16:3, 253 – 260.
Purcell, A.T., R.J. Lamb, E.M. Peron & S. Falchero, 1994. Preference or preferences for landscape? Jnl. Env. Psych., 14, 195 – 209.
Regan, C.L. & S.A. Horn, 2005. To nature or not to nature: associations between environmental preferences, mood states and demographic factors. Jnl. Env. Psych., 25:1, 57 – 66.
Robinson, D.G., I.C. Laurie, J.F. Wager & A.L. Traill, 1976. Landscape Evaluation, Report of the Landscape Evaluation Project to the Countryside Commission for England and Wales. University of Manchester.
Rogge, E., F. Nevens & H. Gulinck, Perception of rural landscapes in Flanders: Looking beyond aesthetics. Landscape & Urban Plg., 82, 159 – 174.
Roth, M. 2006. Validating the use of Internet survey techniques in visual landscape assessment – an empirical study from Germany. Landscape & Urban Plg., 78:3, 179-192.
Sargent, F., 1966.  Ideas and attitudes – A scenery classification system. Jnl.  Soil & Water Cons, 21:1, 26 – 27.
Schomaker, J.H., 1978. Measurement of preferences for proposed landscape modifications. Landscape Research, 3, 5- 9.
Schroeder, H.W., 1984. Environmental perception rating scales – a case for simple methods of analysis. Env. & Behav., 16:5, 573 – 598.
Scolozzi, R., U. Schirpke, C. Detassis, S. Abdullah & A. Gretter, 2015. Mapping alpine landscape values and related threats as perceived by tourists. Landscape Research, 40:4, 451-465.
Sevenant, M. & M. Antrop, 2011. Landscape representation validity: A comparison between on-site observations and photographs with different angles of view. Landscape Research, 36:3, 363-385.
Shafer, E.L., J.F. Hamilton & E.A. Schmidt, 1969. Natural landscape preference: a predictive model. Jnl.  Leisure Research, 1:1, 1 – 19.
Shafer, E.L. & M. Tooby, 1973. Landscape preferences: an international replication. Jnl.  Leisure Research, 5, 60 – 65.
Shuttleworth, S., 1980. The use of photographs as an environment presentation medium in landscape studies. Jnl. Env. Mgt., 11, 61 – 76.
Simpson, C. J., T.L. Rosenthal, T.C Daniel, G.M. White, 1976. Social-influence variations in evaluating managed and unmanaged forest areas. Jnl. Appl. Psych., 61:6, 759 – 763.
Smith, E.L., I.D. Bishop, Williams, K.J.H & R.M. Ford, 2009. Landscape Visualization Tools and Methods: Decision Making with Scenarios. 18th World IMACs/ MODSIM Congress.
Sonnenfeld, J., 1967. Environmental perception and adaptation level in the Arctic. In: Lowenthal, D., Environmental Perception and Behavior, Dept. of Geog. Research Paper 109, University of Chicago.
Sonnenfeld, J., 1969. Equivalence and distortion of the perceptual environment. Env. & Behav., 1:1, 83 – 99.
Stamps, A.E., 1990. Use of photographs to simulate environments: a meta-analysis. Percp. & Motor Skills, 71, 907 – 913.
Stamps, A.E., 1999. Demographic effects in environmental aesthetics: a meta-analysis. Jnl.  Plg Lit, 14:2, 155 – 175.
Stevens, S.S., 1946, On the theory of scales of measurement. Science, 103:2684, 677 – 680.
Stewart, T.R., P. Middleton, M. Downton & D. Ely, 1984. Judgements of photographs vs field observations in studies of perception and judgement of the visual environment. Jnl.  Env. Mgt., 4, 283 – 302.
Stober, D., A. Pogačnik & S. Lončar-Vicković, 2012. Comparison of students and experts’ attitudes on planning and managing the Drava and Mura River landscape (Croatia). Technical Gazette, 19:3, 573 – 581.
Strumse, E., 1996. Demographic differences in the visual preferences for agrarian landscapes in Western Norway. Jnl. Env. Psych., 16, 17 – 31.
Sullivan, W.C., 1994. Perceptions of the rural-urban fringe: citizen preferences for natural and developed settings. Landscape & Urban Plg., 29, 85 – 101.
Taylor, A.F., F.E. Kuo & W.C. Sullivan, 2002. Views of nature and self-discipline: evidence from inner city children. Jnl. Env. Psych., 22:1-2, 49 – 63.
Tempesta, T., 2010. The perception of agrarian historical landscapes: a study of the Veneto plain in Italy. Landscape & Urban Plg., 97, 258-272.
Thurstone, L.L., 1927. A law of comparative judgement. Psychology Review, 34, 273 – 286.
Tips, W.E.J. & T. Savasdisara, 1986a. Landscape preference evaluation and sociocultural background: a comparison among Asian countries. Jnl.  Env. Mgt., 22, 113 – 124.
Tips, W.E.J. & T. Savasdisara, 1986b, The influence of the environmental background of subjects on their landscape preference evaluation. Landscape & Urban Plg., 13, 125–133.
Tips, W.E.J. & T. Savasdisara, 1986d. The influence of the socio-economic background of subjects on their landscape preference evaluation. Landscape & Urban Plg., 13, 225 – 230.
Torgenson, W.S., 1958. Theory and Methods of Scaling, John Wiley & Sons, New York.
Trent, R.B., E. Neumann & A. Kvashny, 1987, Presentation mode and question format artifacts in visual assessment research.  Landscape & Urban Plg., 14, 225 – 235.
Tunstall, S., S. Tapsell & M. House, 2004. Children’s perceptions of river landscapes and play: what children’s photographs reveal. Landscape Research, 29:2, 181 – 204.
Tveit, M.S., 2009. Indicators of visual scale as predictors of landscape preference. Jnl.  Env Mgt., 90:9, 2882–2888.
Ulrich, R.S., 1981. Natural versus urban scenes, some psychophysiological effects. Env. & Behav., 13:5, 523 – 556.
Ulrich, R.S., 1986. Human responses to vegetation and landscapes. Landscape & Urban Plg., 13, 29 – 44.
Ulrich, R.S., R.F. Simons, B.D. Losito, E. Fiorito, M.A Miles & M. Zelson, 1991. Stress recovery during exposure to natural and urban environments. Jnl. Env. Psych., 11, 201 – 230.
Van den Berg, A.E., C.A.J. Vlek & J.F. Coeterier, 1998. Group differences in the aesthetic evaluation of nature development plans: a multilevel approach.  Jnl. Env. Psych., 18:2, 141 – 157.
Vodak, M.C., P.L. Roberts, J.D. Wellman & G.J. Buhyoff, 1985. Scenic impacts of eastern hardwood management. Forest Science, 31:2, 289 – 301.
Wade, G., 1982. The relationship between landscape preference and looking time: a methodological investigation. Jnl.  Leisure Research, 14:3, 217 – 222.
Weinstein, N.D., 1976. The statistical prediction of environmental preferences problems of validity and application. Env. & Behav., 8:4, 611 – 626.
Wellman, J.D. & G.J. Buhyoff, 1980. Effects of regional familiarity on landscape preferences. Jnl. Env. Mgt., 11, 105 – 110.
Whitmore, W., E. Cook & F. Steiner, 1995. Public involvement in visual assessment: The Verde River Corridor study. Landscape Jnl., 14:1, 27 – 45.
Whittow, J.B., 1976. Reviews: Landscape assessment. Landscape Research, 2:1, 13 – 14.
Williams, K. & J. Carey, 2002. Landscape preferences, ecological quality and biodiversity protection. Env. & Behav., 34:2, 257 – 274.
Woodcock, D.M., 1982. A Functionalist Approach to Environmental Preference. PhD dissertation, University of Michigan.
Yamashita, S., 2002. Perception & evaluation of water landscapes: Use of photograph-projective method to compare child and adult residents’ perceptions of a Japanese river environment. Landscape & Urban Plg., 62:1, 3 – 17.
Yang, B-E. & R. Kaplan, 1990. The perception of landscape style: a cross-cultural comparison. Landscape & Urban Plg, 19, 251 – 262.
Yang, B-E. & T.J. Brown, 1992. A cross-cultural comparison of preferences for landscape styles and landscape elements. Env. & Behav., 24:4, 471 – 507.
Yeiser, J.L. & C.L. Shilling, 1978. Student responses to selected terms in natural resource management. Jnl.  Forestry, 76:8, 497-498.
Yu, 1995. Cultural variations in landscape preference: comparisons among Chinese sub-groups and Western design experts. Landscape & Urban Plg., 32:2, 107 – 126.
Yu, S-M, S-S Han & C-H Chai, 2005. Modeling the value of view in real estate valuation: a 3-d GIS approach. Jnl.  Pacific Rim Real Estate Soc. 11:3, 1 – 22.
Zajonc, R.B., 1980. Feeling and thinking: preferences need no inferences. American Psychologist, 35:2, 151-175.
Zajonc, R.B., 2000. Feeling and thinking: Closing the debate over the independence of affect. In: Forgas, Joseph P. (Ed), Feeling and Thinking: The Role of Affect in Social Cognition. Studies in Emotion and Social Interaction. CUP.
Zube, E.H., 1973. Rating everyday rural landscapes of the north-eastern U.S. Landscape Arch, 63:4, 370 – 375.
Zube, E.H., 1974. Cross-disciplinary and intermode agreement of the description and evaluation of landscape resources. Env. & Behav., 6:1, 69 – 89.
Zube, E.H., Brush, R.O. & Fabos, J.G., 1975. Landscape assessment: values, perceptions, and resources. Stroudsburg: Dowden, Hutchinson & Ross.
Zube, E.H. & L.V. Mills, 1976, Cross-cultural explorations in landscape perception. In: Zube, E.H. (Ed), 1976.
Zube, E.H. (Ed), 1976. Studies in Landscape Perception. Institute for Man and Environment, Pub No R-76-1, Uni. of Mass.
Zube, E.H. & D.G. Pitt, 1981. Cross-cultural perceptions of scenic and heritage landscapes. Landscape Plg, 6, 69 – 87.
Zube, E.H., D.G. Pitt & G.W. Evans, 1983. A lifespan developmental study of landscape assessment. Jnl. Env. Psych., 3:2, 115 – 128.