1. Introduction
Rapid global economic growth over recent decades thanks to increasing consumer consumption has made human life more convenient and comfortable. However, rising levels of consumption have led to environmental deterioration through the overuse of natural resources [
1], while the environment faces further degradation because of global warming, the depletion of the stratospheric ozone layer, water, air, noise, and light pollution, and the damage caused by acid rain and desertification [
2,
3]. Because of these negative impacts, the issue of environmental protection has become highly relevant. Indeed, environmental awareness has grown, from the 1960s ecology movement that focused on pollution and energy conservation, to the more recent efforts to reduce our carbon footprint [
4].
These environmental problems have challenged the ways people live, and have resulted in increased environmental consciousness, with consumers now choosing to integrate environmental considerations into their lifestyle choices [
5]. Indeed, some groups of consumers are willing to pay a premium for environmentally friendly products that satisfy their needs (e.g., [
6,
7,
8,
9]). Because of this growing consumer demand for eco-friendly products, socially responsible companies in all industrial sectors are constantly developing products and practices that minimize harmful environmental effects [
10]. This is certainly the case for the hospitality and tourism sectors, where businesses often rely on the health of the environment for their survival. According to the Stys [
11], the restaurants consume a vast volume of disposable products, water and energy in the USA. Referring to the research of Horovitz [
12], restaurants are the worlds’ largest energy users. The restaurants use almost five times more energy per square floor than any type of commercial building. Hu
et al. [
13] claimed that hospitality businesses could negatively influence the sustainability of the local environments in which they operate through their overconsumption of natural resources. As a way of assuming their environmental responsibility, the restaurants are expected to be subjected to green trends. As an important issue, we cannot ignore the consumers’ growing understanding of the effect of food consumption on health [
10], together with increasing environmental awareness throughout society, which has resulted in a growing trend of green restaurants.
Lorenzini [
14] defined a green restaurant as a restaurant with “new or renovated structures designed, constructed, operated, and demolished in an environmentally-friendly and energy-efficient manner”. Increased competition in the marketplace coupled with changing guest demand for green as well the need to ensure guest satisfaction are some of the driving factors for restaurants to go green. Despite this important trend, however, there has been still no green restaurant certified in Taiwan. Although there does not exist “green certified restaurants”, actually, there are many restaurants engaging in green practices, such as recycling, developing their energy and water efficiency, providing sustainable and organic food, and preventing pollution. This study specifically discusses the concept of green restaurants whereby restaurants engage in green practices, green building, and provide organic food.
According to marketing concepts, the consumers are the core element for promoting green restaurants. Increasing understanding of consumers’ attitudes and their responses to green restaurants is the aim of this research. A number of researches have focused on green consumers’ purchases of sustainable food, such as green food, organic food, eco-friendly food [
6,
15]. There are some researchers focused on the relationship between consumer attitude and behavior in response to green products [
2,
5].
In the literature, Value-Attitude-Behavior approach has proved to be a useful framework for investigating a wider range of consumer behavior [
16]. The VAB model is a popular theoretical model for explaining consumer behavior [
17]. It has been widely used as a framework to investigate the role of personal values and value orientation in consumers’ behavior toward the choice of leisure activities [
18], buying organic foods [
19], mall shopping [
20], and consumer behavior intention towards functional foods [
16].
Numerous studies have focused on consumers’ purchases of organic and eco-friendly food (e.g., [
6,
15]). Indeed, environmental concern is often used to measure the importance of the environment and its protection and has been cited as an indicator of the “greening” of consumption [
21]. Moreover, several studies have found a positive relationship between consumers’ environmental concerns and their subsequent environmentally-friendly behavior (e.g., [
13,
22,
23]). However, although some studies have focused on ecological initiatives within the hospitality industry [
24,
25,
26,
27,
28], few have specifically examined environmental issues in the restaurant industry [
13], especially the investigation of consumers’ perspectives of green restaurants. However, to our knowledge, the relationships between personal values and intention towards green restaurant remain untouched in the literature. This study helps to fill the current gap and proposes an integrated model that combines the value–attitude–behavior model with environment concern to understand the underlying values, attitudes, and intentions of Taiwanese consumers’ patron intention toward green restaurants. This study thus offers implications for academic and industry practitioners and contributes new knowledge to the theoretical understanding of consumers’ perceptions of green practices in the hospitality industry.
The rest of the paper is organized as follows. First, a literature review and hypotheses are provided in
Section 2. Research methodology is discussed in
Section 3, and results and discussion are provided in
Section 4 and
Section 5, followed by the conclusion in
Section 6.
4. Results
4.1. Exploratory Factor Analysis
In order to reduce the number of variable dimensions and to identify the underlying factors for 17 personal values items, a series of exploratory factor analyses with principle component and varimax rotation were conducted. Items with low factor loadings (less than 0.50), high cross loading with other items (greater than 0.40), or low communalities (less than 0.30) were all deleted [
73]. Only factors exhibiting an eigenvalue greater than 1 and factor loading of 0.50 or greater were retained. From this procedure, three items of personal values were deleted. Two factors were extracted for the personal values and labeled as factor PV1 = Benevolence and factorPV2 = Universalism (
Table 1). The loadings of the items for each consequent factor were all above 0.50. The three factors explained approximately 64.092% of the total variance. KMO measuring of sample accuracy of 0.933 has met the recommended index of 0.60 and the Bartlett Test of Sphericity was 3175.052 (
p = 0.000) [
74]. The two factors have Cronbach’s α greater than 0.80, indicating an appropriate level of internal consistency. Both factors are the sub-dimensions of personal values as aforementioned with the proposed model.
Similarly, three factors with an eigenvalue greater than one explained 71.473% of the variance of environmental concern scale using the principal components factor analysis. One item with loading factors less than 0.5 was removed from the scale. The varimax-rotated factor pattern implies that the first factor relates to “Human over nature” (4 items, α = 0.883). The second factor relates to “Balance of nature” (4 items, α = 0.808).The third factor concerns “Limits to growth” (3 items, α = 0.810).The arithmetic means of the three multi-item factors were used to build the construct environmental concern for subsequent analysis. The result of the factor analysis for environmental concern was shown in
Table 2.
In the same way, two factors with an eigenvalue greater than one explained 74.894% of the variance of general attitudes scale using the principal components factor analysis. The varimax-rotated factor pattern implies that the first factor relates to “Health attitude” (4 items, α = 0.845). The second factor relates to “Environment attitude” (2 items, α = 0.792).The arithmetic means of the two multi-item factors were used to build the construct general attitude for subsequent analysis. The result of the factor analysis for general attitude was shown in
Table 3.
Table 1.
Factor analysis of personal value.
Table 1.
Factor analysis of personal value.
Factor/Item | Factor Loading | Variance Explained (%) | Cumulative Variance Explained (%) | Cronbach α |
---|
PV1:Benevolence | 32.083 | 32.083 | 0.899 |
Mature Love | 0.820 |
Responsible | 0.817 |
True Friendship | 0.804 |
Loyal | 0.695 |
Honest | 0.623 |
A world at peace | 0.616 |
Wisdom | 0.600 |
PV2:Universalism | 32.009 | 64.092 | 0.901 |
A world of beauty | 0.813 |
Broad minded | 0.796 |
Meaning in life | 0.741 |
Inner harmony | 0.707 |
Equality | 0.700 |
Helpful | 0.639 |
Unity with nature | 0.574 |
Table 2.
Factor analysis of environmental concern.
Table 2.
Factor analysis of environmental concern.
Factor/Item | Factor Loading | Variance Explained (%) | Cumulative Variance Explained (%) | Cronbach α |
---|
EC1: Human over nature | 27.195 | 27.195 | 0.883 |
Humankind was created to rule over the rest of nature. | 0.881 |
Humans need not adapt to the natural environment because they can remake it to suit their needs. | 0.866 |
Humans have to the right to modify the natural environment to suit their need. | 0.865 |
Plants and animals exist primarily to be used by humans. | 0.826 |
EC2: Balance of nature | 23.956 | 51.150 | 0.808 |
Humans must live in harmony with nature in order to survive | 0.902 |
When humans interfere with nature, it often produces disastrous outcomes | 0.797 |
To maintain a healthy economy, we will have to develop a steady-state economy where industrial growth is controlled. | 0.663 |
The balance of nature is very delicate and easily upset. | 0.610 |
EC3:Limits to growth | 20.323 | 71.473 | 0.810 |
We are approaching the limit of the number of people the earth can support. | 0.843 |
Humankind is severely abusing the environment. | 0.772 |
There are limits to growth from which our industrialized society cannot expand. | 0.733 |
Table 3.
Factor analysis of attitude.
Table 3.
Factor analysis of attitude.
Factor/Item | Factor Loading | Variance Explained (%) | Cumulative Variance Explained (%) | Cronbach α |
---|
AT1:Health attitude | 42.230 | 42.230 | 0.845 |
Green restaurant foods are more tasty | 0.881 |
Green restaurant foods have superior quality | 0.803 |
Green restaurant foods are more attractive | 0.731 |
Green restaurant foods are healthier | 0.703 |
AT2: Environment attitude | 32.664 | 74.894 | 0.792 |
I practice environmental conservations tasks | 0.876 |
Unless we do something, environmental damage will be irreversible | 0.871 |
Reliability for each of the factors was obtained using the calculation of a Cronbach α coefficient. The Cronbach α coefficients ranged from 0.792–0.901 (see
Table 1,
Table 2 and
Table 3). All factors were above the cut-off criterion of 0.7 recommended by Nunnally [
75].
4.2. Confirmatory Factor Analysis
Confirmatory factor analysis (CFA) was then conducted using AMOS16 [
76] to test the convergent validity of the constructs used in subsequent analysis. The fit indices suggested by Joreskog and Sorbom [
77] and Hair, Black, Anderson, and Tatham [
78] were used to assess the model adequacy. Convergent validity of CFA results should be supported by item reliability, construct reliability and average variance extracted [
78]. As shown in
Table 4,
t-values for all the standardized factor loadings of the items were found to be significant (
p < 0.01). In addition, construct reliability estimates ranging from 0.70–0.87 exceeded the critical value of 0.7 recommended by Hair
et al. [
78], indicating it was satisfactory. The average variances extracted for all the constructs fell between 0.55 and 0.77, and were greater than the value of 0.5 suggested by Hair
et al. [
78]. Composite scores for each construct were obtained from the mean scores across items representing that construct.
Table 4.
Reliability and validity of each variable.
Table 4.
Reliability and validity of each variable.
Construct | Items | Standardized Factor Loading | t Value | CR | AVE | Cronbach’s α |
---|
Personal Values | Benevolence | 0.893 | --- | 0.87 | 0.77 | 0.937 |
Universalism | 0.856 | 13.933 *** |
Environmental Concern | Balance of Nature | 0.950 | --- | 0.82 | 0.70 | 0.871 |
Limits to Growth | 0.707 | 9.896 *** |
General Attitudes | Healthy | 0.843 | --- | 0.70 | 0.55 | 0.848 |
Environment | 0.616 | 7.682 *** |
Behavioral intention | Willing to patronize | 0.698 | --- | 0.87 | 0.62 | 0.859 |
Considerable chance of patronizing | 0.933 | 13.165 *** |
Predominantly patronize | 0.812 | 12.006 *** |
Recommend others to patronize | 0.692 | 10.334 *** |
The proposed conceptual model was tested by using the fourth constructs: namely personal values, environmental concern, general attitudes and behavioral intentions. Factors of “benevolence” and “universalism” were served as the measurement variables of personal values. Factors of “balance of nature” and “limits to growth” were served as the measurement variables of environmental concern. In addition, factors of “health attitude” and “environment attitude” were used as the measurement variables of general attitudes.
Fornell and Larcker [
79] indicated that discriminant validity exists when the proportion of variance extracted in each construct exceeds the square of the coefficient that represents its correlation with other constructs. As shown in
Table 5, all the AVE values were greater than the squares of the correlations between constructs; hence, discriminant validity was satisfactory for all constructs.
Table 5.
Discriminant validity for the measurement model.
Table 5.
Discriminant validity for the measurement model.
Construct | PV | EC | AT | BI |
---|
PV | 0.77 | | | |
EC | 0.312 ** | 0.70 | | |
AT | 0.141 ** | 0.128 ** | 0.55 | |
BI | 0.129 ** | 0.102 ** | 0.302 ** | 0.62 |
After testing the reliability and validity of the measurement model, we next determined the goodness of fit of the structural model in order to test H1–H6. According to Gefen
et al. [
80], between 100 and 150 responses are necessary to carry out structural equation modeling. Thus, the fact that we received 254 responses in the present study implies that the sample size was sufficiently large.
Table 6 shows that six goodness-of-fit indices yielded values above the recommended threshold levels. Consequently, the goodness of fit between the proposed model and the observed data in the present study was deemed acceptable [
80].
Table 6.
Recommended and actual values of fit indices.
Table 6.
Recommended and actual values of fit indices.
Fit Index | χ2/df | GFI | AGFI | CFI | NFI | NNFI | RMSEA | RMR |
---|
Recommended value | <3 | >0.90 | >0.80 | >0.90 | >0.90 | >0.90 | <0.08 | <0.05 |
Actual value | 3.331 | 0.931 | 0.869 | 0.947 | 0.927 | 0.948 | 0.096 | 0.049 |
4.3. Testing the Hypothesized Relationships
The path coefficients estimated using structural equation methods and the results of hypothesis testing are presented in
Table 7 and
Figure 2. H1 proposed that personal values (PV) positively influence general attitudes (AT). The path coefficient from PV to AT (beta = 0.278,
p < 0.05) was statistically significant at the 5% level, indicating the positive effect of PV on AT. Thus, H1 was supported.
H2, H3 and H6 proposed that personal values (PV), general attitudes (AT), and environmental concern (EC) positively influence behavioral intention (BI) to patronize a green restaurant. First, the path coefficients from PV to BI (beta = 0.182, p < 0.05) and from AT to BI (beta = 0.624, p < 0.001) were statistically significant at the 0.05 level, indicating the positive effects of PV and AT on BI. However, the path coefficient from EC to BI (beta = −0.324, p > 0.05) was not statistically significant at the 5% level. Thus, only H2 and H3 were supported, but H6 was not. According to the size of the presented beta values, the magnitude from AT to BI (beta = 0.624, p < 0.001) was larger than that from PV to BI (beta = 0.182, p < 0.05).
H4 proposed that personal values (PV) positively influence environmental concern (EC). The path coefficient from PV to EC (beta = 0.680, p < 0.001) was statistically significant at the 1% level, indicating the positive effect of PV on EC. Thus, H4 was supported. H5 proposed that environmental concern (EC) positively influences general attitudes (AT). The path coefficient from EC to AT (beta = 0.223, p < 0.05) was statistically significant at the 5% level, indicating the positive effect of EC on AT. Thus, H5 was supported.
Table 7.
Results from hypothesis testing.
Table 7.
Results from hypothesis testing.
Hypothesis | Path | Path Coefficient | t Value | Results |
---|
H1 | PV→AT | 0.278 | 2.552 * | Supported |
H2 | AT→BI | 0.624 | 5.784 *** | Supported |
H3 | PV→BI | 0.182 | 2.006 * | Supported |
H4 | PV→EC | 0.680 | 10.838 *** | Supported |
H5 | EC→AT | 0.223 | 2.008 * | Supported |
H6 | EC→BI | −0.034 | −0.324 | Not supported |
Figure 2.
The Hypothesized Model.
Figure 2.
The Hypothesized Model.
** p < 0.01; *** p < 0.001.
5. Findings
The present study investigated consumers’ values, attitudes, environmental concerns, and intentions to patronize green restaurants in Taiwan. It demonstrated that the proposed model fits the data well; thus did the findings enable us to draw the following four main conclusions.
First, the supported data showed that personal values significantly influence consumers’ attitudes, environmental concerns and their intention of visiting a green restaurant. Specifically, these demonstrate that the values of universalism and benevolence are important to consumers’ intentions to choose a green restaurant. As the significant effects of the personal values on attitude, environmental concern and behavior intention, which demonstrated that more personal value of self-transcendence, the more likely that consumers would have a more positive attitude and behavioral intention to consume the green restaurants. This confirms using the VAB model and is consistent with previous studies [
83,
84]. The findings of the personal value effects provide green restaurant practitioners some insights regarding consumers’ green restaurants patronizing behavioral. Based on the results, practitioners need to identify which group of consumers has higher self-transcendence values, such as charitable organizations. Attending their council meetings and advertising during their activities for the purpose of exploring the benefit of green restaurants may be advisable.
As shown by the support for H1, personal values significantly influence consumers’ attitudes by guiding their actions and by helping them to develop positive attitudes toward relevant objects and situations. Several authors have found similar results, including Rajani [
36], who identified that personal values (particularly universalism) significantly and positively affect environmental concern. Consumers that display values of universalism thus consider the consequences of such behavior on the environment [
44]. However, we cannot overlook the potential role of personal values in predicting the use of green restaurants.
Second, attitude toward intention to visit a green restaurant is the prominent variable in our model among those constructs tested. This finding is in line with the attitude toward environmental sustainability, which increases intention to dine at a green restaurant and implies that future research should take more account of the effect of favoring a healthy and sustainable attitude.
Third, the present research discusses both the direct and the indirect effects among the investigated variables, in order to explain their interrelationships. The results verified that consumers’ attitudes partially mediate the influence of personal values on intent to visit a green restaurant. In this regard, our findings imply that the values of universalism and benevolence are likely to help individuals positively evaluate the behavioral consequences of eating in a green restaurant. Thus, it is crucial for green restaurant managers to enhance individuals’ attitudes in order to maximize the influence of personal values on visiting intention.
Finally, we found that environmental concern does not have a significant relationship with intent to visit green restaurants nor does it mediate the effect of personal values on visit intention. These results show that environmental concern cannot predict intent to visit a green restaurant. In other words, although consumers may show concern for the environment, they may not follow through with this line of behavior and actually eat in a green restaurant.
6. Conclusions
There are four academic contributions in this study. First, we combine the concepts of personal value, attitude, environmental concern, and behavioral intention to propose a research framework for green restaurants. Second, we develop a framework to enhance understanding green restaurants’ patron behavior. We prove that personal value positively influences attitude and visiting intention for green restaurants. Third, we demonstrate that attitude not only has a positive effect on behavioral intention but also plays a mediating role in the relationship between personal value and behavioral intention. Fourth, this paper extends the research of the value–attitude–behavior model into the field of green hospitality.
In terms of future research avenues, first of all, we suggest adding perspectives on universalism, environment compassion, willingness to give back, and benevolence values to ethics courses starting in basic education such as elementary school, also emphasizing environment ethics in tertiary education programs, because cultivating such values when consumers are students would help to develop social responsibility. Consequently, when they have the opportunity to dine out with family or friends, they may increasingly favor a green restaurant over a traditional one. Secondly, we would also suggest that the relevant government agencies formulate a policy to encourage the patronage of green restaurants as a type of benevolent and health-conscious behavior. Thirdly, as with the prior study by Crompton [
85], it is increasingly evident that resistance to action in society to address environment crises lies in cultural values. Debates on the consequences of cultural values and mechanisms become vigorous regarding the evidence base for sustainability civil society campaigns. Echoing this view, we suggest future study on consumers’ behaviors toward green restaurants should use the VAB model integrating cultural values to enrich the model context. Finally, a revised NEP scale [
86,
87] should be considered for use in future research.
As with all research, this study has several limitations. First, we confined our sample to major cities in Taiwan, which may have similar levels of environmental awareness based on the similar profiles of respondents. Future research should be conducted with consumers with different degrees of environmental knowledge, especially pertaining to green restaurants. Future samples should also be diversified in terms of the cultural backgrounds of respondents, by sampling respondents from other countries, for example. Second, the dependent variable used in this study is patronage intention of a green hotel rather than actual behavior itself, even though behavior intention is an adequate proxy of actual behavior [
88]. Following up on surveys by examining actual patronage behaviors is suggested.