Greg Michalec & Jenny Hershberger
NSII Final - May 6, 1998
This study attempts to discover whether or not the preferences we have for certain elements in a natural landscape have sociobiological roots. A current theory in the field of sociobiology suggests that those elements that we find to be aesthetically pleasing in nature actually are the same elements that were preferable to our prehistoric ancestors for survival. Thus, following natural selection theory, those who developed an ability to identify more suitable environments for habitation would be more likely to pass on these genes, hence our current aesthetic preference for these elements.
We tested this theory by digitally manipulating an image several times in order to isolate specific variables that would be either beneficial or detrimental to early human life. We then created a survey out of these images, asking people to rate each image based in its aesthetic value.
Our results show that features such as presence of water and healthy vegetation do increase a landscape's aesthetic value. However, many of the other elements returned inconclusive results. These results may be linked to flaws in our experimentation method.
Human beings have strong preferences, and reactions to the aesthetic values of landscapes. As Stephen Kellert points out, it doesn't matter where people are, "Central Avenue or Central Park...in prison or on 5th Avenue, people react with combinations of wonder and admiration at the sight of a glorious sunset, a stretch of pristine beach, a colorful bird, or a wide variety of other creatures and landscapes." (Kellert, 1997) Regardless of our culture, era, or geography, we as a people have found beauty in nature, and we have found there are certain specific elements of the natural world we find more pleasing than others.
In his book, Kinship to Mastery Stephen R. Kellert proposed a theory for this belief that certain elements in nature would be visually attractive. "...people's aesthetic preference for certain landscapes and species has also been related to the increased likelihood of encountering safety, sustenance, and security." (Kellert, 1997) This theory brings with it the question of whether or not humans are in some way 'genetically seeded' or 'programmed' with specific preferences to natural landscapes.
One argument indicates that there are certain cultural influences regarding natural preferences and aesthetic values. Bruce Hull and Grant Revell in their "Case Study in Bali"" bring into play several points that could go against a 'genetic seed' answer for how we determine aesthetic beauty in landscapes. They are as follows:
1. Landscape preference is a learned value. The usage's of 'scenic beauty' may be a culture specific term, not used the same way across the cultures
2. Preference is dependent on your expectations and information - your experiences. This includes religious influences, and purpose. For example, a tourist would view a landscape differently than someone who was raising a family there would.
In their study, Hull and Revell noticed that there were similar responses, but that these responses were based on different cultural reasoning. They even pointed out that the Balinese people, who live in Indonesia, had adopted western culture through the existence of tourism in their city, and that this played a role in the similarity between responses. They concluded that culture does play a role in our preferences of landscape aesthetics. However, Hull and Revell also made note of the fact that natural scenes with open lush and green landscapes were chosen. This is evidence that points toward a genetic preference. Hul and Revell go on to say that most studies show that "...there is a universally agreed upon scenic standard, and that there are inherited mechanisms which are responsible for these similarities. These mechanisms are posited as having adaptive, evolutionary significance." (Hull and Revell, 1989) This evolutionary significance is a speculation that Roger Ulrich supports.
Ulrich states, "...if biophilia is represented in the gene pool it is because a predisposition in early humans for biophilic responses to certain natural elements and settings contributed to fitness or chances for survival." (Ulrich, 1993) In order to determine which specific elements of natural landscapes humans find preferable for beauty, we must look to the elements of nature that our prehistoric ancestors found preferable for survival.
Humans that were able to locate and select suitable habitats were more likely to survive and pass their genes onto the next generation than those who couldn't find suitable homes. In this way, selection of landscapes play major role in natural selection. "The needs of our ancestors were the same as ours: to find adequate food and water and to protect themselves from the physical environment, predators, and hostile conspecifics." (Heerwagen & Orians, 1993) Based on this, and the study of the evolutionary functionality of natural elements, the following factors have been found to potentially aid in human survival. According to the hypotheses of Biophilia and Sociobiology, landscapes that contain these elements should be the ones that are most aesthetically pleasing to humans as well.
Preferred Elements in Nature and Purpose
1. The presence of clean, fresh water. Beneficial to human life, signifies nourishment and would support the growth of nature.
2. Lush, green leaves and bright foliage. These signify ample nourishment, as opposed to dry and brown leaves that are dying.
3. Large open areas. This openness provides better visibility of predators and prey, as well as allowing for increased mobility.
4. Scattered clumps of trees and bushes. This shrubbery provides shelter and an escape from most predators. (Ulrich, 1993; Heerwagen & Orians, 1993)
After conducting research and gaining preliminary information on our subject we conducted a survey to determine if landscapes containing the preferred elements of nature would be found to be more aesthetically pleasing than those which did not. The survey consisted of 9 photographs. Each photograph was shown on a page alone, and participants were asked to give it an aesthetic value ranging from one to ten.
In an attempt to eliminate variables, and keep the scenes as consistent as possible, seven of the nine photographs were digitally altered, and all of the photographs were from the same front, or heads on perspective. Our original photo was of an open green pasture containing blue skies and a 'fairy ring', this picture was then altered for color tint, to show lush greenness, and also brown and dying areas to see which landscape would be preferred. A few clumps of trees were added to make another photograph, and there was a photograph containing alot of dense trees close together, to test for open area preference and the relevancy of small clumps of trees. There were two photos with a lake in them, one had all of the 'good' elements of nature in it, and then the lake was added to the original photograph. There were two desert type scenes to illustrate dry, or exotic Savannah areas. One had small bushes and the original photo's background, where the second desert scene had a large cactus imposed in front of the previous desert scene.
In order to reach as many respondents as possible, and in an attempt to eliminate large amounts of homogenous cultural values, we placed the survey on the Internet, allowing for a larger sampling area of people. We also asked a few brief questions at the beginning of the survey for categorization purposes, such as: What is your age, gender, and do you live in a rural, urban, or suburban area? We also explained that the results would be used for a school project and stressed that we needed the respondents first and primary response to the aesthetic value of our photos, and also that taking the survey more than once would skew our data.
The survey data was collected into a text file, which we imported into Microsoft Excel. We then organized the data into groups, each group relating to one of the four specific variables (visibility, water, vegetation, and tree clumps), as well as one group for the combination photos.
In order to interpret our data, we first took the mean scores for each photo and charted them in order to get an overall idea of where each variable fell (Figure 1).
We then created five groups of data, each testing for one of the specific variables. In the visibility group, we tested the low visibility photo versus the default. For presence of water, we compared the water photo to the default to the dry. We grouped the dead vegetation, default, and lush vegetation photos to test the healthiness of the vegetation, and compared the image with tree clumps to the default for evaluating our hypothesis concerning trees. Finally, we arranged the images containing all of the "good" and "bad" features, along with the original photo, in order to test the overall comparison. Each of these groups were placed in a "Paired Comparisons" table in StatView in order to obtain P-Values(Figures 2-5).
We also ran the data through a series of steps in an attempt to eliminate variations in scoring - some people tend to score very high (6-10), some low (1-6), and some across the entire scale (1-10). We did not want these variations to affect our results. Therefore, we first obtained the mean score for each individual respondent. We then subtracted that value from each of their responses, thereby obtaining the difference from the individual's mean score given to each image. However, after all this, the results remained the same - only the scale was changed - and we don't believe that this method did anything to eliminate range variations (Figure 7).
We also tested the data to see if gender had any influence on our results, thereby suggesting a cultural influence on aesthetic preference rather than a biological root. However, gender differences, tested through ANOVA tables, proved to be insignificant in all categories except for the "tree clumps" variable. However, we do not believe that there is any real significance in this finding.
Our results are very mixed in relation to the hypotheses for which we were testing. In terms of overall ranking, most variables fell as we had expected them. However, the photo testing for lack of water - "dry" - scored surprisingly high, whereas the photo showing the small clusters of trees ended up with unexpectedly low results. The test for low visibility also scored slightly higher than we would expected it, surpassing the default photo.
As for the individual tests, again our results followed our hypotheses in some areas, but proved inconclusive in most. Our most successful test was that of healthiness of vegetation(Figure 4). Both the photos of dying vegetation and lush vegetation were significantly different varied from the default photograph, and returned an extremely low P-value (.0006) in relation to each other.
While the image testing for water scored significantly higher than original photograph, its position above the photograph showing "dryness" was not statistically significant (Figure 3). Surprisingly, the "dry" image, which we had predicted people would tend to dislike, ranked higher than the default picture, although this difference was not significant.
Both the tests for visibility (Figure 2) and presence of small clusters of trees (Figure 5) also proved to be inconclusive. Both of these images, when compared to the original image, returned rather high P-values (.85 and .28, respectively). However, it is somewhat beneficial to our hypotheses that neither of these results are statistically significant, for, as previuosly stated, the photo with low visibility unexpectedly ranked higher than the original, while the image displaying tree clumps ended up lower than the unaltered image.
The group testing the overall comparison of the variables, comparing the two images that displayed all the values that we had declared to be negative and positive along with the original image, also returned varied results. (Figure 6) While there was a statistically significant difference between the two extremes, as well as between the "All Good" image and the unaltered photo, the default image and the "all bad" image did not return a significant variation.
The results that we obtained do support some elements of our proposed hypothesis. From our study, one may conclude that the presence of bodies of water and of lush, healthy vegetation in a landscape do result in greater aesthetic preference than do those that do not have such features. Our study also shows that dry, dying vegetation does serve as a successful deterrent as far as aesthetic landscape preference goes. One may also logically infer that the existence of the elements that we stated as being preferable in a landscape has a greater effect on aesthetic ratings than does the presence of those we stated as being less amiable - people like the "good" elements more than they dislike the "bad elements." This is evident in the fact that while the photo showing the presence of water, as well as the one with all of the favorable elements, were significantly distant from the original photo, while their counterparts did not achieve statistical significance in relation to the default. However, as far as the questions of visibility and the presence of trees as shelter, our field study did not provide conclusive results.
Our experiment had a great deal of flaws in it. Probably the greatest error in our data is that we did not reach a very heterogeneous population. The vast majority of our respondents were 18-22 year olds who grew up in the suburbs, and most of them were Western majors right here at good ol' Miami U. Thus, it is very hard to say whether or not culture had a significant impact on our study, as we did not test a very culturally diverse group. It would have also behooved us to have reached more participants in general - the more data you have, the more conclusive your study will be. Another significant was the images which we used. Although we attempted to isolate the variables for which we were testing as much as possible, the images were in no way perfect representations of what we had in mind. Had we been able to access a greater source of landscape photography, and had better mastery over image-manipulating software, we might have obtained results closer to those which we expected. The fact that the survey was on the internet might have affected our results as well. It is rather hard to imagine a digital image as an environmental image, especially when you staring at a tiny image on a computer screen with a mouse in your hand. It would have been nice to be able to handle out large, full-color glossy photos, or maybe create a shadowbox-type technique, but hey, we're poor college kids. One item that would have probably aided our study and been easy to implement is to provide the respondents with some sort of guide on the scale of one to ten, such as labeling one through three as "Not Very Aesthetic," four through six as "Mildly Aesthetic," etc.
Overall, this study does suggest that there may be some biological/evolutionary links concerning what we find aesthetic in a natural landscape, however a more refined version of our study is necessary before conclusive results might be obtained.
1. Heerwagen, Judith H. & Orians, Gordon H., "Humans, Habitats, and Aesthetics," in Kellert, Stephen R. & Wilson, E. O., eds., The Biophilia Hypothesis. Washington, D.C.: Island Press, 1993
2. Hull IV, Bruce R. & Revell, Grant R. B., "Cross-Cultural Comparison of Landscape Scenic Beauty Evaluations: A Case Study in Bali" in Journal of Environmental Psychology 9(1989):177-191
3. Kellert, Stephen R., Kinship to Mastery: Biophilia in Human Evolution & Development. Washington D.C.: Island Press, 1997
4. Ulrich, Roger S., "Biophilia, Biophobia, and Natural Landscapes," in Kellert, Stephen R. & Wilson, E. O., eds., The Biophilia Hypothesis. Washington, D.C.: Island Press, 1993
For Further Info on this Topic, Check out this WWW Site: http://dameat.novasys.com .
Return to the Topic Menu
IMPORTANT: For each Research Response, make sure the title of the response is different than previous titles shown above!