1. Introduction
With respect to a historical viewpoint of the New Ecological Paradigm (NEP) conceptualization and applicability, only a few research works were devoted to the adjustment of the NEP scaling to strictly environmental issues between the years 1980–1990. Nevertheless, by the end of 1980s, handling of the out-of-border environmental issues of aquifer pollution, land desertification, protection of endangered ecosystems and atmospheric degradation necessitates the adaptation of NEP scaling towards recognizing and confronting the reality of global environmental change [
1].
In response to this environmental transition from the local to global level of manipulation, a measurable tool of 3 revisions of the original NEP Scale was developed by Dunlap [
1], with the latest revision rapidly replacing the 1978 version in most studies. Under the original NEP Scale, the questions rotate around three aspects of environmental beliefs (these are the sub-scales which frame the psychological tendency to evaluate the favoring or disfavoring of the natural environment): “(1) a belief in humans’ ability to upset the balance of nature, (2) the existence of limits to growth, and, (3) humans’ right to rule over the rest of nature” [
2]. Consequently, the uni-dimensionality of NEP scaling towards environmental attitudes reflects an ecologically oriented point of view (a high NEP score reflects positiveness for the preservation of natural sources) or an anthropogenic (a low NEP score reflects positiveness for the exploitation of natural sources) approach, accordingly.
Dunlap [
1] reviewed the current (at the time of publication) application of the several formats of the NEP Scale and examined the criticism of them. The original NEP Scale was created during this study, as the authors concluded that 12 items offer good reliability (alpha = 0.81) and are sufficiently correlated to be treated as a single rating scale. Although Dunlap [
1] included questions to capture three different aspects (or what the author later termed “facets of the NEP Scale”), he argued that since they constituted a coherent worldview, all the items should hang together.
Under the context of NEP (new ecological paradigm scale), the authors specifically designed a revised NEP scaling where the original scale was elaborated by two new aspects of ecological orientation, having the following revised performance, compared to the former one:
The introduction of items dealing with the likelihood of eco-crises was made, in alignment with the growing awareness of global ecological problems.
The inclusion of three items per each one of the formulated subscales was implemented. Therefore, the development of 8 pro-NEP and 7 anti-NEP items consists of the arrangement of the 15-item NEP scaling; thus, no subscale is measured with items located in only one (positively or negatively expressed) direction.
The author removed outdated expressions like «mankind». Furthermore, the author subsumed the NEP into the relevant social-psychological theory, by indicating that the NEP questions capture the primitive inner-beliefs about the relationship between humans and their environment.
Under the revised NEP Scale, Dunlap [
1] modified the original NEP scaling towards psychometrically robust and contemporarily structured terminology. Indeed, the original NEP scaling contained an uneven ratio of pro-trait to con-trait items and gender-oriented items (“mankind” was interchangeably mentioned as “humans”). Hawcroft and Milfont [
2] methodologically followed, by using specific versions of the NEP scaling. Specifically, the authors revealed the comparability between the 12- and 15- item formulations. Nevertheless, the validity of the scoring and the comparability were prominent while selecting fewer items out of the whole NEP Scale.
Considering the previous theoretical background, this study aims to measure and analyze citizens’ perceptions based on the NEP Scale in a Greek area in order to evaluate their environmental awareness and locate the most important facets of the NEP Scale. The analysis includes hypothesis tests between the sample demographics and the NEP Scale’s mean score. Furthermore, based on the above results, the relationship between respondents’ NEP Scale score and their willingness to pay (WTP) for an expansion of renewables into the current energy mix will be examined. Relevant research concerning the applicability and evaluation of the NEP Scale is presented in the review section.
3. Materials and Methods
The research area is the prefecture of Evia, which belongs to the region of Central Greece. This research expands on similar research, for the area of Evia [
6]. The region of Central Greece occupies an area of 15,449 km
2, representing 11.8% of the country. Evia is one of the largest islands of Greece, with an area of 4167 km
2. The ground is mainly mountainous and semi-mountainous, while the plains occupy only 20% of the total area. Its eastern part is islanded and includes the island of Skyros and other islets. The sea borders of Evia are defined by the northern part of the island, which is covered by the Pagasetic Gulf, on the eastern side of the island, which extends into the Aegean Sea and the western side of the island, which is washed by the Evia Gulf (
Figure 2). The island of Evia has a length of 180 km, and its width ranges from 8 to 50 km, while its coastline is about 680 km long. The capital of the prefecture is Chalkida. The territory of the island consists of 44% of pastures, 30% of forest land and 26% of arable land. Regarding the population of the Prefecture of Evia, according to the last two censuses carried out by the Statistical Service, in 2001 it amounted to 207.305 inhabitants while in 2011 it increased to 210,815 inhabitants [
14].
The per capita GDP of Evia for 2013 was 13,315 euros, well below the corresponding size of the prefecture of Viotia (21,159 Euros), but very close to the average price of the Region of Central Greece (14,858 Euros). The decline in the per capita GDP of the Prefecture of Evia in the year 2013 compared to 2011 is of the order of 10%, while for the whole Greece and the prefecture of Attica this decrease was almost 12%.
Our research was selected to take place in this specific area due to its high wind and solar potential. Southern Evia has already attracted investment interest in wind energy investment. According to Baltas and Dervos [
16], three wind priority areas are distinguished in Greece. Evia belongs to the 2nd area, while the total installed capacity of wind farms in South Evia is about 212 MW. Consequently, due to the numerous wind farms in the area, the issue of social acceptance of the local community for further expansion of investment is of great importance.
Concerning sample size, the estimation was done by using the equation of simple random sampling with substitution [
13]. For our calculations, the confidence level is set at 95%, meaning that a 5% significance level is set. According to Eng [
17], when the examined variable measures proportions, the equation for sample size in per cent sampling is expressed in Equation (1):
where
n is the total sample size,
p is the pre-study estimate of the proportion to be measured,
e is the accepted error (in our case 5%) and,
Z1−a/2 is the standard normal deviate which takes a standard value based on the significance level set [
17]. Thus, for the calculation of the proper sample size from the proportional variables, calculated from the preliminary survey, we use respondents’ willingness to pay (WTP) for renewable energy (61% replied “Yes”), for 95% confidence level and an accepted error term of 5%, presented in Equation (2):
The data for this research were collected through questionnaires which were completed between September 2017 and October 2017. The database was analyzed with SPSS v.20.
Sample’s demographics are provided in
Table 1.
The analytical composition of the research was undertaken in compliance with the following steps shown in
Figure 3:
4. Results
Initially, a reliability test was run in order to measure the NEP Scale’s internal consistency in the current research. A Cronbach’s a value equal to 0.714 was obtained, which shows an acceptable level of reliability.
Based on methods used by researchers deploying similar data [
18], principal component analysis (PCA) was carried out in order to verify the existence of NEP Scale’s facets and the relationship between them [
18]. According to this method, each component interprets a proportion of the variance that has not been interpreted by the other created components [
13]. The results of Κaizer-Meyer-Olkin’s Measure of Sampling Adequacy (KMO = 0.785) and Bartlett’s test of sphericity (p-value = 0.000) indicated that the data were suitable for PCA. The PCA was carried out using varimax rotation, a method which minimizes the number of large weight variables and makes them more interpretable.
Table 2 confirms the existence of the NEP Scale’s five components in the context of the current research, based on the number of components with eigenvalues greater than 1.
The following
Table 3 shows how the NEP Scale’s Items are grouped into the five extracted components, based on their loadings.
Examining the NEP Scale’s variables scores enabled the measurement and evaluation of the environmental concern of the respondents. NEP Scale variables are measured using a 7-point Likert scale. The scale items are coded from 1 = “strongly disagree” to 7 = “strongly agree”, while 4 is “neutral”. The following
Table 4 sums up the agreement and disagreement statements’ cumulative percent, as well as the undecided statement’s percent. This modification improves result interpretation [
18]. Negative phrases are reverse-coded. Mean scores for the five facets of the NEP Scale are presented in
Figure 4.
Based on the following
Figure 4, we conclude that the dominant social paradigm does not characterize the respondents, meaning that they do not believe that humans are superior to other all other species, the Earth provides unlimited resources, and that progress is an inherent part of human history [
19]; thus, they are preferably characterized by a pro-NEP attitude and seem environmentally sensitive. Furthermore, standard deviations are relatively low (anti-anthropocentrism has the lowest one), which means that the respondents’ views are somewhat similar.
Figure 4 depicts the respondents’ mean scores according to the five NEP subscales (components), as they are obtained from the PCA and presented in the literature review section.
In the aftermath of the analysis, we test the possible differences between environmental awareness as captured by NEP Scale mean scores and respondents’ demographics. According to comparable studies [
18,
20], respondents’ demographics are statistically related to the mean NEP score index. Furthermore, in a recent study, income and area of residence were found to be related to respondents’ positive view towards renewable energy sources [
21]. Using a one-sample Kolmogorv-Smirnov test [
22], it was concluded that the NEP mean score is normally distributed (p-value = 0.196); thus, one-way Analysis of Variance (ANOVA) will be applied to test the significance of the difference between the means of demographics and NEP Scale mean score.
The one-way ANOVA results presented in
Figure 5 reveal that there is no statistically significant difference between the means of the NEP score and respondents’ gender (p-value = 0.264), education (p-value = 0.938), occupation (p-value = 0.171) and income (p-value = 0.163); on the contrary, the test revealed a statistically significant difference between the NEP Scale score and respondents’ place of residence (p-value = 0.000). By looking at
Figure 5, we see that the mean score on the NEP Scale is higher in rural than in the semi-urban and urban areas of residence. This is an important finding, since people in rural and agricultural areas who live close to the natural environment seem to appreciate it more as they are in direct contact with it. As the distance between the place of residence and nature increases, it seems that the inner ecological beliefs become weaker. Post Hoc tests show a statistically significant difference between rural and urban areas of residence (p-value = 0.000) and rural and semi-urban areas of residence (p-value = 0.030), while there is no statistically significant difference between the urban and semi-urban areas of residence (p-value = 0.263).
In order to further examine the NEP Scale score and the impact of the five facets, a multiple regression model is constructed. The NEP Scale score is set as the dependent variable, while NEP Scale components (
Table 3) as derived from the PCA are the independent ones. All the variables are normally distributed (p-value > 0.05). Based on the following
Table 4, we present Equation (3), which describes the relationship between NEP Scale scores and its subscales:
According to the adjusted R-squared index, Equation (3) explains 96.3% of NEP Scale score variance, but there is no autocorrelation in the residuals as proved by Durbin-Watson’s test value which is equal to 2.01. According to the coefficient values in
Table 5, the subscale “reality to limits of growth (COMP_1)” is the most important subscale in explaining NEP Scale score conformation, while, the least important one is “anti-exceptionalism (COMP_4)”.
At this point, we annotate that the highly adjusted R squared index can be explained since all the PCA’s components are added in the regression model. The purpose is to have the highest adjusted R squared with the least number of variables. In fact, in a regression model where only COMP_1 and COMP_2 are used, an adjusted R squared index equal to 0.727 is obtained, showing that these two are the variables with the most significant contribution. However, we used all the PCA’s components as our intention to see the effect of all the extracted components. Thus, in order to further test the significance of the models, the Akaike Information Criterion (AIC) will be used [
23]. Comparing the AIC’s value for Equation (3) model which is equal to 15.11 and this of the model including only COMP_1 and COMP_2 where the AIC’s value is equal to 16.68, we conclude that the model of Equation (3) has the most parsimonious fit, outperforming the other.
According to several studies, the NEP Scale score can affect willingness to pay (WTP) for renewable energy expansion [
24,
25]. A first indication is obtained by using the independent samples t-test. More specifically, the binary variable “WTP”, which represents respondents’ willingness to pay for a further expansion of green energy into the energy mix, will be the grouping variable for an independent samples t-test concerning the New Environmental Paradigm Scale (NEP) mean scores, as presented in
Table 6.
According to Levene’s test for equality of variances result, p-value = 0.001 is used for the t-test for equality of means. Based on this, a statistically significant difference for the NEP score mean is recorded; the NEP score mean is higher in the group of respondents who declare a positive willingness to pay for renewable energy. To further examine the above-recorded tendency, a binary logistic regression model was constructed. In this model, willingness to pay is the dependent variable, while NEP score and respondents’ demographics are the independent ones. The model’s initial category concerning respondents’ willingness to pay is the negative one,
Table 6.
Based on the following
Table 6, Equation (4) is presented, describing the relationship between willingness to pay for green energy and the independent variables that are statistically significant:
The optimal solution is found after three iterations, where
R2 = 14.9%. The Hosmer and Lemeshow test [
26] indicates that the total model is statistically significant (p-value > 0.05). Based on the binary logistic regression equation, two of the independent variables are correlated with WTP. These variables are the mean NEP score and respondents’ income.
From the interpretation of the coefficients, we notice that by keeping income unchanged, an increase of one unit for the NEP score reduces the likelihood that a person will not accept additional payment for WTP expansion (WTP = no) by 1 − Exp(B) = 1 − 0.617 = 0.383 = 38.3%. Furthermore, by keeping the NEP score unchanged, the probability of a person staying in the null category of “no”, concerning the additional payment, is 1 − Exp(B) = 1 − 0.829 = 0.171 = 17.1%. This means that environmental sensitivity as expressed by the NEP Scale affects respondents’ WTP more than their income.
As it is already argued, a statistically significant relationship exists between the NEP score and income. We expand on this finding in
Table 7, by examining the impact of the additional amount of money respondents are willing to pay for the expansion of renewable energy. For this reason, the 5-point Likert scale capturing the additional payment intention (“Amount”) is set as the dependent variable in an ordinal regression model, while NEP score and demographics are the independent ones.
Based on the results of the above
Table 7, it can be stressed that both the NEP score and income are statistically significant variables. However, it is noted that income is partially statistically significant in the low category and high categories.
According to the estimates, it is noteworthy that people with lower income are less likely to spend an additional amount of money than people with annual income higher than 30000 Euros (Income = 7). Furthermore, the NEP Scale is positively correlated with respondents’ willingness to pay for further expansion of green energy into the energy mix energy. This model explains 32.5% of the dependent variable’s variance as indicated by the Nagelkerke Pseudo R-Square [
27].
To assess model validity, it is of the utmost importance to point out that the regression coefficients are the same for all the categories, as provided by the test of parallelism, provided in
Table 8; thus, the ordinal regression model is appropriately selected for this case. Furthermore, all the model’s predictors are necessary, as the final model fitting shows. Finally, by using Pearson’s and deviance goodness of fit tests, it is perceived that the model fits well with the data, according to the significance level p-values.
6. Conclusions
This study revealed that the respondents in the area of Evia-Greece are NEP inspired, meaning that they adopt an ecological worldview. The possibility of an eco-crisis has the highest score amongst the components of the NEP Scale while the understanding that growth must be controlled has the second highest score, reveling that respondents are ecologically sensitive. Particularly, concerning respondents’ beliefs about the depletion of natural resources on Earth (NEP Scale’s component 1) these are the utmost important components in shaping their total NEP score, as shown by the regression analysis.
The mean NEP score was found to be statistically different between respondents who live in urban and rural areas. Respondents in rural areas have a higher mean NEP score and thus increased ecological sensitivity compared to those with a permanent residence in urban areas. This finding suggests that an important variable that could be used to further study the degree of ecological sensitivity is the proximity and access frequency to elements of the natural environment.
The NEP score also was found to be correlated with respondents’ willingness to pay for an expansion of renewables into the Greek energy mix. The most remarkable component of this correlation was that the NEP score was more important in shaping WTP for renewable energy than respondents’ income. This is a significant outcome, as in relevant studies, WTP for renewable energy was found to be correlated with environmental sustainability which is one of the principal outstanding goals to be achieved by humanity [
13,
31]. Thus, promoting the principles of NEP among people will have a significant contribution in shaping the ecological consciousness to motivate them towards the notion of environmental sustainability.
The study confirmed the importance of promoting peoples’ ecological consciousness as in this way, environmental sustainability would be a feasibly accomplished goal. Based on this analysis, there are opening new avenues for future research. The first one concerns the ways of increasing Greek peoples’ NEP score by raising their environmental awareness through actions such as educational reformation and activities (at municipal and city levels) for bringing people closer to the natural environment. Furthermore, an issue of further analysis is how the NEP score is applied in different populations [
7] and how it is affected by variables such as the area of residence and income. Lastly, future research works should be oriented to examine the possible effects of NEP score in shaping modern economic development as there is already evidence for the correlation between renewable energy usage, environmental sustainability, and broader socio-economic development [
31,
32,
33,
34,
35].