3.1. The Classroom Experiment
In order to answer the research questions of the current study, 85 spatial planning students at the undergraduate level (mainly of Austrian origin) were recruited for the classroom experiment [
21]. As an incentive to participate, students could earn points for the upcoming economics exam. As the sample of participants of the experiment is not representative, we do not expect reliable economic values in absolute terms (e.g., consumer surplus or willingness-to-pay in EUR). However, in order to test the hypotheses formulated below, we reasonably assume that the effect of the different modes of presentation (2D or 3D) alters economic values in the same direction, i.e., qualitatively similar to the whole population. The use of classroom experiments in general is a widely used and accepted approach in psychological, social and economic sciences—for instance, in behavioral economics (cf. [
22]).
For producing the stereoscopic images for our classroom experiment, we faced the challenges described above in
Section 2. According to the size of the spatial simulation laboratory, the position of participants and the size of the screen we identified a stereoscopic optimum with a distance of intersection for two images of 2.5 cm. We used the software “Free 3D Photo Maker” (Digital Wave Ltd., London, UK) in order to produce the (stereoscopic) 3D images on the basis of the 2D pictures of an Alpine pasture.
Students registered for the experiment during certain time slots and were then randomly assigned to a group (treatment); in order to guarantee equal conditions of the experiment, we limited each group size to 15 respondents, which is the maximum capacity of the spatial simulation laboratory. Prior to the experiment, students did not know whether they would be presented 2D or 3D images.
The spatial simulation laboratory at Vienna University of Technology is equipped with a 3D virtual reality (VR) environment including a 3D stereoscopic rear projector, tractors and pointers; the resolution of the images both in the 2D and the 3D group were equal (for further information on the laboratory, see [
23]).
Students of the 2D group were verbally introduced to the survey in a standard seminar room (see
Figure 1), and then handed the self-administered questionnaire; the 2D image and the two development scenarios (see
Figure 2) presented below were shown by means of an overhead projection. In total, 43 students were assigned to the 2D groups (and 42 to the 3D treatment, respectively).
In the spatial simulation lab, the verbal introduction and the whole experiment were completely equal to the 2D treatment, but students were in addition handed 3D glasses in order to watch the images of the original landscape and the two potential development scenarios.
Figure 1 presents impressions from both the classroom (seminar room) and the simulation laboratory. Students then filled in the self-administered questionnaire after watching the 2D or 3D pictures.
3.2. Development Scenarios of Alpine Pastures and Testable Hypotheses
As mentioned before, we presented respondents three pictures; the original one displayed an Alpine pasture which proved to mirror the usual image, e.g., in tourism brochures, of a typical Alpine pasture. Due to space restrictions, we cannot describe the procedure to test for similarities between pictures in detail. In a nutshell, we took several pictures of tourism brochures, analyzed them with respect to their contents (forests, open spaces, Alpine pastures, meadows, and blue sky) by breaking up the pictures in a 1 mm × 1 mm grid, and then took our own picture in high resolution in order to have a picture available that could be manipulated [
21].
Based on this picture, two scenarios were implemented by manipulating the original picture and adding (or removing) several components of landscape changes such as trees, hard-surfaced paths and roads, and tourism infrastructure—all jeopardizing the appearance of a typical Alpine pasture as a part of natural and cultural heritage. Scenario 1 mirrors a development that might reduce biodiversity of Alpine pastures, and change the landscape from the image of a landscape attractive for tourists to an area largely afforested owing to reducing or abandoning the traditional way of managing Alpine pastures (scenario “Afforestation”). Scenario 2 also assumes a drastic change in the management of the pasture and intensively develops tourism infrastructure in terms of an Alpine hut upgraded to a restaurant, with information panels, and other sports and leisure infrastructure (scenario “Tourism”). The landscape elements added to the original picture all represent probable and realistic elements of landscape changes according to the two scenarios.
After having manipulated the original 2D picture to display the two potential scenarios, we then produced the corresponding 3D pictures.
Figure 2 presents the original picture, and the two manipulated pictures displaying the scenarios 1 and 2.
In order to test for differences in the perception, and consequently, in the economic valuation of the two scenarios, we structured our questionnaire as follows. First, a block of questions ascertained preferences of respondents and the frequency of vacations in Alpine landscapes, including several dimensions to describe an Alpine pasture. Second, the status quo of the appearance of the landscape, and the two scenarios, were valued by respondents, including their hypothetical frequency of vacations if one of the two scenarios were to be become reality. Third, respondents could state their willingness-to-pay for a landscape conservation funds to manage and maintain Alpine pastures, and, alternatively, to avoid scenario 1 or 2. Fourth, the questionnaire included statements regarding the 2D and the 3D presentation, respectively. The 2D group was asked about their opinions about a potential 3D presentation, e.g., whether they would expect more information if they would have had a 3D presentation instead. The 3D group was asked whether they thought that 2D pictures were enough for perceiving landscape changes. Both groups were also asked whether they thought that 3D could provide more information, or an improved spatial perception than 2D pictures. Finally, some socio-economic questions (e.g., income) were asked.
Our main six hypotheses to be tested are (for a list of variables, see
Table 1):
H1: Travelfrequency2D,Status quo ≠ Travelfrequency3D,Status quo
H2: Travelfrequency2D,Scenario1 ≠ Travelfrequency3D,Scenario1
H3: Travelfrequency2D,Scenario2 ≠ Travelfrequency3D,Scenario2
H4: WTP2D,Scenario1 ≠ WTP3D,Scenario1
H5: WTP2D,Scenario2 ≠ WTP3D,Scenario2
H6: WTP∅(2D,3D),Scenario1 ≠ WTP∅(2D,3D),Scenario2
The first group of hypotheses tests whether travel frequency is different between the two groups (2D vs. 3D), depending on the scenario presented. The second group tests for differences of willingness-to-pay (WTP) with respect to the form of presentation, as well as the scenarios. As this paper answers the research questions and tests for the hypotheses in an exploratory approach, we do not have a priori assumptions about the direction of each test. For instance, if 3D presentation contain more information and offer a more comprehensive spatial perception of landscape changes, respondents may state a higher travel frequency and a higher WTP in the 3D group, as several studies showed that an increased level of information may also lead to a higher willingness-to-pay (e.g., [
24,
25,
26,
27]).
Besides analyzing descriptive evidence regarding potential differences according to the hypotheses outlined above, this paper specifically tests the hypotheses by means of two equations which are estimated econometrically:
Equation (1) outlines the empirical model in order to estimate a “demand function” for vacations in Alpine areas, specifically with the focus on Alpine pastures, on the basis of a travel cost model (cf., e.g., [
28]). The frequency of vacations in Alpine landscapes, denoted by the variable “Travelfrequency”, should thus be dependent on T
i (travel cost) consisting of both conservative estimates of transportation and travel time costs. As we do not aim to provide a representative study, nor measure recreation benefits in absolute terms, neither the choice of costs of travel time nor transportation costs influences the ordinal ranking with respect to scenarios or groups of respondents ([
29]).
Gi contains a vector of grouping variables denoting the scenarios 1 and 2, and differentiating the sample of respondents between the 2D versus 3D presentations.
Si denotes the income of respondent i which can be reasoned by the standard assumption that travel frequency as a normal consumer good, as well willingness-to-pay for landscape conservation, is positively correlated with individual (or household) income. Other socioeconomic characteristics that usually determine economic values such as age, and education, are not included since our (non-representative) sample of participants in the experiment is too homogenous for testing the influence of these variables in the econometric estimations presented below.
Pi forms a vector of variables comprising the strength of agreement to a number of statements, e.g., regarding the role of Alpine pastures for Alpine landscapes as part of cultural and natural heritage, and the experience of respondents with developments at Alpine pastures potentially considered unfavorable (e.g., afforestation, intensive tourism development). The questionnaire also included an elicitation of activities during the vacations in the Alps. As Alpine pastures are mainly used by hikers during the summer season, the statistical analysis also does not account for these rather homogenous activities as explanatory variables in the estimations.
With respect to equation (2) presented above, we basically assume comparable influences of the variables, and include all variables in our willingness-to-pay (WTP) function except for travel costs (see
Section 4.1 and
Section 4.2 below).