Wildlife Conservation through Economically Responsible Ecotourist: The Mediator Roles of Attitude between Anticipated Emotion and Intention to Stay in Local Homestays

If responsible ecotourists stay in a local homestay, this will benefit local people economically and lead to improved wildlife conservation. This study aims to examine the mediator roles of attitudes between anticipated emotion and intention. It was conducted in Penang National Park, Malaysia, and a stratified sampling method was used for collecting the data. In all, 320 sets of questionnaires were analysed using the SPSS Amos 24.0 Statistical Software Package to test the Structural Equation Modelling. The findings show that economically responsible ecotourist attitudes to staying in local homestays for wildlife conservation partially mediate the relationship between anticipated emotion and intention to stay in a local homestay for wildlife conservation. This study suggests that players in the ecotourism industry should incorporate emotional elements in their marketing strategies to promote local homestays to responsible ecotourists, which would benefit local economies.


Introduction
According to a 2020 World Wildlife Fund (WWF) report [1], wildlife populations have declined by 68% since 1970 due to their over-consumption by poor local people living in or near national parks. According to Duffy, St John, Büscher and Brockington [2], poverty is the main reason for illegal wildlife hunting by locals, who then sell the hunted wildlife at high prices as a source of income. CITES, the Convention of International Trade in Endangered Species of Wild Fauna and Flora, reports that international wildlife trade is estimated to be worth billions of dollars per year, affecting hundreds of millions of animal and plant specimens [3,4]. It is important to alleviate poverty among local people to save wildlife. Increased incomes will reduce the local dependence on wildlife. Ecotourism is defined by the International Ecotourism Society (TIES) as "responsible travel to natural settings that conserves the environment and enhances the well-being of local people" [5]. TIES is an example of a non-profit organisation devoted to aiding businesses by implementing ecotourism practices and fostering long-term community development. Poverty is the main cause of illegal wildlife hunting among local communities, causing widespread wildlife extinction, so ecotourist groups have introduced local ecotourism plans [6,7] to develop the local economy to eliminate the local over-dependence on wildlife. Through local ecotourism, communities have been given opportunities to run homestay businesses to obtain economic benefits from the resulting profits [8]. Many homestays have been established by locals in or near national parks to provide accommodation facilities and experiences to ecotourists [9]. However, it was found that tourists prefer [8] to stay in hotels or resorts located in cities [10,11] rather than choosing local homestays as their holiday accommodation [12][13][14][15]. As a result, local homestay businesses are unprofitable and this sector is stagnant [16,17]. Staying in a local homestay for wildlife conservation is deemed as responsible behaviour in this study. Ecotourists are those who travel in a natural area and spend a predetermined number of days developing the local economy [18,19]. They are responsible people who care about maintaining and protecting the natural environment [20,21]. Ecotourists display responsible economic behaviour as they are willing to do everything that others expect them to do, even when confronted with challenging situations [22]. They will stay in a local homestay for the purposes of wildlife conservation. However, to what extent do they intend to stay in local homestays for wildlife conservation during their visits to a national park? To answer this question, it is vital to understand ecotourists' responsible economic behaviour in relation to wildlife conservation so an effective marketing strategy for staying in local homestays can be implemented. As tourists and wildlife have a strong emotional connection, the present research study focuses on the relationship between anticipated emotion, attitude and intention.

Literature Review and Research Hypotheses
Ecotourism is an economic tool for wildlife conservation [23,24]. Previous studies involving wildlife conservation through ecotourism highlighted conservation learning that focused on captive wildlife, such as in zoos or aquariums [25][26][27][28], and conservation interpretation [29][30][31][32][33][34]. According to Myers [35], tourists are agents in wildlife conservation. They are responsible people who care about maintaining and protecting wildlife [20,21]. Responsible behaviour is the act of doing what should be done in any given situation, even if it is difficult, unpleasant or unclear [36]. There are three main types of responsible behaviour, namely environmentally responsible behaviour, socially responsible behaviour and economically responsible behaviour [37]. However, scholars of tourist responsible behavioural research tend to study environmentally responsible behaviour [38][39][40] and socially responsible behaviour [41,42]. For example, Xu, Kim, Liang and Ryu [43] conducted a case study of Nansha Wetland Park in China to examine the links between tourist involvement, experience and environmentally responsible behaviour. He, Hu, Swanson, Su and Chen [44] conducted a study of tourists' environmentally responsible tourism behaviour and perceptions of destinations and quality. Su and Swanson [41] analysed the effect of destination social responsibility on tourist behaviour. Luo, Tang, Jiang and Su [42] investigated socially responsible tourists' awareness of environmentally responsible behaviour. In tourism economics research, studies have proven that tourists are willing to pay for wildlife conservation [45][46][47][48][49][50][51][52]; however, no studies have been undertaken on tourists' economically responsible behaviour. In the present research, ecotourists staying in local homestays for wildlife conservation is referred to as responsible economic behaviour.
The Theory of Planned Behaviour (TPB) was developed by Ajzen in 1991 for studying the human decision-making process [53] and it is widely used in conservation behavioural research [54][55][56][57][58][59]. It consists of three important components, namely attitude, subjective norm and perceived behavioural control, which are used as important determinants in understanding human behaviours [53]. In psychological terms, attitude refers to a person's mental and emotional state [60]. It can enable a better understanding of how humans perceive the world and how they behave [61]. It involves an overall evaluation of attitude objects, e.g., favour or disfavour, or like or dislike [62,63]. In wildlife value orientation [64], a beneficial interaction between humans and wildlife exists when humans have a positive attitude towards wildlife conservation [65,66]. Previous studies indicate that local communities [67][68][69][70], stakeholders [71][72][73][74][75] and teachers [76] have positive attitudes towards wildlife conservation. However, ecotourist attitudes towards wildlife conservation remain unclear. According to Newhouse [77], ecotourist attitudes have major implications for wildlife conservation. Based on Ajzen [53] in the TPB, attitudes are antecedents of intention. Intention is the individual willingness to undertake a particular behaviour, so it has a direct relationship with behaviour [53]. The TPB explains that the more an attitude relates to a behaviour, the greater the intention to perform the behaviour [53]. In tourism research, several studies have confirmed that ecotourist attitudes have a direct effect on intention, as shown by research conducted by Clark Mulgrew, Kannis-Dymand, Schaffer and Hoberg [78]; Phu, Hai, Yen and Son [79]; Meng and Choi [80]; and Gstaettner, Rodger and Lee [81]. Based on the arguments above, the present study proposed the first research hypothesis as follows: Hypothesis 1 (H1). Attitude to staying in a local homestay for wildlife conservation has a significant direct effect on intention to stay at a local homestay for wildlife conservation.
The importance of emotions in the decision-making process is often overlooked by scholars because they believe rational thinking is more meaningful. They regard emotions as irrational phenomena that can lead to incorrect thinking. However, current ideas emphasise the significance of emotions in decision making. Loewenstein and Lerner [82] distinguish between two sorts of emotions encountered during decision-making: anticipated emotions and immediate emotions. Emotions that are anticipated (or predicted) are not experienced immediately, but are anticipated as a result of the rewards or losses arising from a decision. The two effects that exist in anticipated emotion are self-consistency (e.g., pride and guilt) and basic hedonic (e.g., pleasure and frustration). Self-consistency shows a long-term emotional effect while basic hedonic is more short-term [83]. Bagozzi, Belanche, Casaló and Flavián [84] mentioned that anticipated emotion plays an important role in purchasing decision making. Emotion is widely used by major companies and organisations in marketing strategies to influence buyers' emotions by placing emotional taglines in their marketing adverts. Some examples are the 'Choose Happiness' tagline, used in Coca Cola ads; the 'Stop Climate Changes Before It Changes You' tagline by the World Wildlife Fund, which adds the element of fear; the '30+ Years' Experience Successfully Representing Accident Victims' tagline by John Rapillo, a law office that uses the element of trust; the 'Help Me Read This' tagline by Children Of The World, which features a sad element; or 'You're Not You When You're Hungry' for the Snickers bar (Mars), which uses an element of humour by displaying an awesome picture of Godzilla, except when he is hungry. Surprisingly, adverts that include emotional elements are generally successful in influencing users/customers to buy products or services. In tourism research, studies also show that tourists have a strong emotional connection with wildlife through their wildlife view experience [85][86][87]. This research assesses how tourists' anticipated emotions in relation to wildlife affect their decisions to stay in local homestays for wildlife conservation. In the TPB, scholars tend to relate environmental research to norms [88][89][90][91][92]. However, in their study, Onwezen, Antonides and Bartels [93] suggest that exploring anticipated emotion in environmental research is critical. Moreover, a meta-analysis conducted by Rivis, Sheeran, and Armitage [94] mentioned that anticipated emotion can increase the variance in explaining attitude and intention in the TPB. In TPB research, previous studies by Kim, Njite and Hancer [95] and Londono, Davies and Elms [96] have shown that anticipated emotion has a significant direct effect on attitudes to behaviour and intention. Based on the arguments above, the present study proposed further research hypotheses as follows: Hypothesis 2 (H2). Anticipated emotion has a significant direct effect on attitude to staying in a local homestay for wildlife conservation.

Hypothesis 3 (H3)
. Anticipated emotion has a significant direct effect on intention to stay in a local homestay for wildlife conservation.
Mediation analyses are often used in social psychology [97] to investigate the underlying mechanism or process by which one variable influences another via a mediator variable in order to better understand a known relationship. When no evident direct relationship exists between an independent variable and a dependent variable, mediation analysis might enable a better understanding of the relationship [98]. In tourism behavioural research, attitude has been identified as functioning in a mediator role [99]. Studies by Rahman, Rana, Hoque and Rahman [100] showed that tourists' attitude has the effect of mediating between service and satisfaction. Mehmood, Liang and Gu [101] determined that attitudes mediate the relationship between user-generated content and travel intentions. Studies conducted by Gilchrist, Masser, Horsley and Ditto [102]; and Taylor, Ishida and Donovan [103], indicate that attitude mediates the relationship between anticipated emotion and intention. The proposed research framework, illustrating the relationship between attitude, anticipated emotion and intention, is shown in Figure 1. Based on the arguments above, the present study proposed a further research hypothesis as follows: Hypothesis 4 (H4). Attitude to staying at a local homestay for wildlife conservation mediates the relationship between anticipated emotion and intention to stay at a local homestay for wildlife conservation.
Mediation analyses are often used in social psychology [97] to investigate the underlying mechanism or process by which one variable influences another via a mediator variable in order to better understand a known relationship. When no evident direct relationship exists between an independent variable and a dependent variable, mediation analysis might enable a better understanding of the relationship [98]. In tourism behavioural research, attitude has been identified as functioning in a mediator role [99]. Studies by Rahman, Rana, Hoque and Rahman [100] showed that tourists' attitude has the effect of mediating between service and satisfaction. Mehmood, Liang and Gu [101] determined that attitudes mediate the relationship between user-generated content and travel intentions. Studies conducted by Gilchrist, Masser, Horsley and Ditto [102]; and Taylor, Ishida and Donovan [103], indicate that attitude mediates the relationship between anticipated emotion and intention. The proposed research framework, illustrating the relationship between attitude, anticipated emotion and intention, is shown in Figure 1. Based on the arguments above, the present study proposed a further research hypothesis as follows: Hypothesis 4 (H4). Attitude to staying at a local homestay for wildlife conservation mediates the relationship between anticipated emotion and intention to stay at a local homestay for wildlife conservation.

Target Area
In terms of richness of biodiversity, Malaysia is one of the top 17 countries, with an estimated 12,500 plant species, 306 mammal species, 742 bird species and 547 reptile species [104]. As a protected area gazetted under Malaysia National Parks [105], Penang National Park (PNP) functions to protect and conserve wildlife. PNP is home to 417 species of flora and 143 species of fauna, including numerous rare and endangered species. It is located at 5°26′53″N 100°11′36″E in Peninsular Malaysia and has an area of 2,563 hectares (9.9 sq mi). In this study, PNP was selected as the target area to conduct the research, as it is a developed eco-tourism area that receives high numbers of domestic and foreign tourists. Among the tourism activities at PNP are visiting the interpretation centre, camping at Keranchut Beach and Teluk Kampi; wildlife observation, e.g., squirrels, beavers, moles, ants and pythons; and various bird species such as eagles; walking at the Canopy Bridge;

Target Area
In terms of richness of biodiversity, Malaysia is one of the top 17 countries, with an estimated 12,500 plant species, 306 mammal species, 742 bird species and 547 reptile species [104]. As a protected area gazetted under Malaysia National Parks [105], Penang National Park (PNP) functions to protect and conserve wildlife. PNP is home to 417 species of flora and 143 species of fauna, including numerous rare and endangered species. It is located at 5 • 26 53 N 100 • 11 36 E in Peninsular Malaysia and has an area of 2563 hectares (9.9 sq mi). In this study, PNP was selected as the target area to conduct the research, as it is a developed eco-tourism area that receives high numbers of domestic and foreign tourists. Among the tourism activities at PNP are visiting the interpretation centre, camping at Keranchut Beach and Teluk Kampi; wildlife observation, e.g., squirrels, beavers, moles, ants and pythons; and various bird species such as eagles; walking at the Canopy Bridge; doing water activities; picnicking; fishing; jungle trekking; mountain climbing; and visiting the lighthouse, the Turtle Conservation Centre and Lake Meromiktik. As with national parks in other countries, PNP also faces its own challenges in reviving the local community economy for wildlife conservation [106,107]. Figure 2 depicts a map of Penang National Park, the study area for this research.
doing water activities; picnicking; fishing; jungle trekking; mountain climbing; and visiting the lighthouse, the Turtle Conservation Centre and Lake Meromiktik. As with national parks in other countries, PNP also faces its own challenges in reviving the local community economy for wildlife conservation [106,107]. Figure 2 depicts a map of Penang National Park, the study area for this research.

Method
In this study, the non-experimental quantitative research method was used [108].
Although this method appears highly flexible, the results obtained through the quantitative research procedure are extremely reliable because data is collected, processed and presented in numbers, which are not deceptive. Moreover, in non-experimental research, a predictor variable or subjects cannot be manipulated, proving that this method is highly systematic. Non-experimental quantitative research is often used because it allows researchers to gain a better understanding of social science through questionnaires, polls and survey data analysis. Where the data is easily communicated using statistics and figures, this method is performed to study the relationships between variables in an existing phenomenon [109].

Sampling
Tourists formed the sample in this study, and they were selected using the stratified sampling method [110]. Initially, tourists who came to Penang National Park were asked whether they had checked in to local accommodation or were staying overnight during their visit to Penang National Park. Only tourists who had checked in to local accommodation or were staying overnight were selected for the sample in this study. Secondly, only a tourist who agreed to be a respondent received a set of questionnaires from the research team. The data collection was conducted carefully face to face with tourists in Penang National Park and they were given a token of appreciation after the questionnaire had been completed. A summary of the sample background is shown in Table 1.

Method
In this study, the non-experimental quantitative research method was used [108].
Although this method appears highly flexible, the results obtained through the quantitative research procedure are extremely reliable because data is collected, processed and presented in numbers, which are not deceptive. Moreover, in non-experimental research, a predictor variable or subjects cannot be manipulated, proving that this method is highly systematic. Non-experimental quantitative research is often used because it allows researchers to gain a better understanding of social science through questionnaires, polls and survey data analysis. Where the data is easily communicated using statistics and figures, this method is performed to study the relationships between variables in an existing phenomenon [109].

Sampling
Tourists formed the sample in this study, and they were selected using the stratified sampling method [110]. Initially, tourists who came to Penang National Park were asked whether they had checked in to local accommodation or were staying overnight during their visit to Penang National Park. Only tourists who had checked in to local accommodation or were staying overnight were selected for the sample in this study. Secondly, only a tourist who agreed to be a respondent received a set of questionnaires from the research team. The data collection was conducted carefully face to face with tourists in Penang National Park and they were given a token of appreciation after the questionnaire had been completed. A summary of the sample background is shown in Table 1.  Table 1 shows that tourists aged 18 to 23 years old formed the largest group of respondents, 39.7% of the total, followed by respondents aged from 24 to 29 (22.5%) and 30 to 35 (18.8%). However, the smallest group was tourists aged above 41, which was only 4.4% of the total. Table 1 also shows that females comprised the largest group of respondents in terms of gender, with a total of 58.1%, while the rest were male. Respondents from the Malaysian group were found to be the largest in terms of nationality, with 82.2% of the total; the rest were from the non-Malaysian group. The table shows that tourists with employed status were the largest group of respondents in this field, at 71.6%, while the smallest group comprised pensioners, with 1.6% of the total. Tourists visiting for leisure comprised the largest group of respondents in terms of the purpose of visiting, with 74.1% of the total; followed by the 'other purposes' group, with 16.9%; and business purposes, with 9.1% of the total.

Measurement
Since this study aimed to predict tourist intention to stay in a local homestay for wildlife conservation, the survey method was deemed to be the most suitable. The most appropriate instrument used to survey tourist intentions is a questionnaire form [111]. The questionnaire consisted of two parts. The first part related to the tourist background, and respondents were required to answer questions regarding their age, gender, nationality, working status and their purpose of visiting PNP. The second part of the questionnaire contained three measurements, namely anticipated emotion (four items), attitude (four items) and intention (three items). All the items were measured using a five-point Likerttype scale, from 5-strongly agree to 1-strongly disagree. All items were adapted from the literature research. The items used to measure anticipated emotion were based on Bagozzi et al. [84], such as 'If I visit Penang National Park next time, it would be a real pleasure if I stay at a local homestay that contributes to wildlife conservation'. The items used to measure attitude and intention were based on Ajzen [112], such as 'I would like very much to stay at a local homestay if this contributes to wildlife conservation in Penang National Park' and 'I intend to stay at a local homestay when visiting Penang National Park this year for wildlife conservation'. Appendix A shows the items used in the questionnaire. Two steps were involved in conducting this research. The first step was the pilot study. At this stage, 80 tourists had been selected as respondents. The internal consistency shows that the questionnaire had very high reliability, with Cronbach's Alpha values between 0.89 and 0.94. The questionnaire had also gone through the validation process before the pilot study was conducted. The questionnaire had been checked by the supervisors to determine whether the items would achieve the appropriate measurement of the constructs. The second step in this research process was conducting the actual research. The data collection process took place over four weeks from 19 December 2020 to 10 January 2021and was conducted during the weekend. Since 20 sets were not answered completely by the respondents, only 320 sets of questionnaires were used for data analysis out of the 340 set questionnaires that had been returned to the research team.

Measurement Modelling
In this study, the data was analysed using the SPSS Amos 24.0 Statistical Software Package to test the Structural Equation Modelling (SEM) [113]. SEM is widely used in social science research [114] because it provides a flexible framework for constructing and analysing complicated interactions across different variables, allowing researchers to verify the validity of a theory using empirical models [115]. Its capacity to manage measurement error, one of the most significant constraints of most studies, is perhaps its greatest advantage [116]. SEM is a combination of factor analysis and multiple regression analysis, and it is used to analyse the structural relationship between measured variables and latent constructs [117]. Conducting SEM in data analysis involves two steps [117]. The first step is conducting a measurement of the model and the second step is conducting a structural equation of the model. In a measurement model, a confirmatory factor analysis (CFA) process needs to be conducted initially. CFA is a validation process and it is used to determine the validity and reliability of a latent construct [117]. There are three types of validity: construct validity, convergent validity and discriminant validity. The former is achieved when the Fitness Index Model has reached its standard. Four indicators [113] are suitable for testing the Fitness Index Model; the Root Means Square of Error Approximation (RMSEA) (good if RMSEA < 0.08); the Comparative Fit Index (CFI) (good if CFI > 0.9); the Tucker-Lewis Index (TLI) (good if TLI > 0.9) and Chi Square/degree of freedom (Chisq/df) (good if Chisq/df < 3.0). between 0.89 and 0.94. The questionnaire had also gone through the validation process before the pilot study was conducted. The questionnaire had been checked by the supervisors to determine whether the items would achieve the appropriate measurement of the constructs. The second step in this research process was conducting the actual research. The data collection process took place over four weeks from 19 December 2020 to 10 January 2021and was conducted during the weekend. Since 20 sets were not answered completely by the respondents, only 320 sets of questionnaires were used for data analysis out of the 340 set questionnaires that had been returned to the research team.

Measurement Modelling
In this study, the data was analysed using the SPSS Amos 24.0 Statistical Software Package to test the Structural Equation Modelling (SEM) [113]. SEM is widely used in social science research [114] because it provides a flexible framework for constructing and analysing complicated interactions across different variables, allowing researchers to verify the validity of a theory using empirical models [115]. Its capacity to manage measurement error, one of the most significant constraints of most studies, is perhaps its greatest advantage [116]. SEM is a combination of factor analysis and multiple regression analysis, and it is used to analyse the structural relationship between measured variables and latent constructs [117]. Conducting SEM in data analysis involves two steps [117]. The first step is conducting a measurement of the model and the second step is conducting a structural equation of the model. In a measurement model, a confirmatory factor analysis (CFA) process needs to be conducted initially. CFA is a validation process and it is used to determine the validity and reliability of a latent construct [117]. There are three types of validity: construct validity, convergent validity and discriminant validity. The former is achieved when the Fitness Index Model has reached its standard. Four indicators [113] are suitable for testing the Fitness Index Model; the Root Means Square of Error Approximation (RMSEA) (good if RMSEA < 0.08); the Comparative Fit Index (CFI) (good if CFI > 0.9); the Tucker-Lewis Index (TLI) (good if TLI > 0.9) and Chi Square/degree of freedom (Chisq/df) (good if Chisq/df < 3.0).   Convergent validity of a model is achieved if the Average Variance Extracted (AVE) value of the latent construct is above 0.5, the Composite Reliability (CR) value of the latent construct is above 0.6 and the factor loading of the item is above 0.6 [113]. Table 2 shows that the model had a high internal consistency and reliability. The results indicate that the  When a model is free from redundant items, it means the discriminant validity is achieved. The discriminant validity is assessed by comparing the square root of AVE and the correlation coefficient of the latent constructs. It is achieved if the diagonal value with bold (the square root of AVE) is larger than the value in its row and column (correlation coefficient) [113]. Table 3 shows that the discriminant validity of all the latent constructs was achieved, since the results indicate that all the square root AVE values were larger than the correlation coefficient values. In the study, the data distribution analysis was also conducted to evaluate whether the data was normally distributed. Table 4 presents the results of the analysis. The results indicate that the data was normally distributed. The values of the skewness item are lower than 1.0, indicating the data was normally distributed.

Structural Equation Modelling
After the measurement modelling had been conducted, the final step implemented the structural equation modelling. In this step, the Fitness Index Model also needed to be run Sustainability 2021, 13, 9273 9 of 16 structurally [117]. Figure 4 shows that the Fitness Index Model for SEM was achieved, since the results were as follows: RMSEA = 0.047, CFI = 0.993, TLI = 0.991 and ChiSq/df = 1.692.

Structural Equation Modelling
After the measurement modelling had been conducted, the final step implemented the structural equation modelling. In this step, the Fitness Index Model also needed to be run structurally [117]. Figure 4 shows that the Fitness Index Model for SEM was achieved, since the results were as follows: RMSEA = 0.047, CFI = 0.993, TLI = 0.991 and ChiSq/df = 1.692. The direct effect results shown in Table 5 indicate that the path coefficient of attitude to staying at a local homestay for wildlife conservation in terms of intention to stay at a local homestay for wildlife conservation is 0.47, so it is statistically significant. Thus, hypothesis H1 is supported by the results. The path coefficient of anticipated emotion to attitude to staying at a local homestay for wildlife conservation is 0.85, which shows that it is statistically significant. Thus, hypothesis H2 is also supported by the results. Finally, the path coefficient of anticipated emotion on intention to stay at a local homestay for wildlife conservation is 0.4 and the p-value result indicates that it is statistically significant. Thus, hypothesis H3 is also supported by the results. Mediation analysis is prominent in psychological theory and research [118]. This study was conducted in part to quantify the extent to which attitude to staying at a local homestay for wildlife conservation was a mediating variable. A mediating variable is one that illustrates a relationship between dependent and independent variables [118]. There are two ways to analyse the mediating effect, which are to determine the direct and the The direct effect results shown in Table 5 indicate that the path coefficient of attitude to staying at a local homestay for wildlife conservation in terms of intention to stay at a local homestay for wildlife conservation is 0.47, so it is statistically significant. Thus, Hypothesis H1 is supported by the results. The path coefficient of anticipated emotion to attitude to staying at a local homestay for wildlife conservation is 0.85, which shows that it is statistically significant. Thus, Hypothesis H2 is also supported by the results. Finally, the path coefficient of anticipated emotion on intention to stay at a local homestay for wildlife conservation is 0.4 and the p-value result indicates that it is statistically significant. Thus, Hypothesis H3 is also supported by the results. Mediation analysis is prominent in psychological theory and research [118]. This study was conducted in part to quantify the extent to which attitude to staying at a local homestay for wildlife conservation was a mediating variable. A mediating variable is one that illustrates a relationship between dependent and independent variables [118]. There are two ways to analyse the mediating effect, which are to determine the direct and the indirect effects. Based on Baron and Kenny, several steps are needed to determine the existence of a mediator variable using the direct effect. First, there should be a correlation between an independent variable and a mediator variable; second, there should be a correlation between the mediator variable and the dependent variable; and third, if the relationship between the independent variable and the dependent variable is significant, this means the existence of a partial mediator. The results reveal a partial mediating effect of attitude to staying at a local homestay for wildlife conservation on the relationship between anticipated emotion and intention to stay at a local homestay for wildlife conservation, which is supported by the findings for Hypothesis H4. The indirect effect tests also show that there is a mediating effect of attitude to staying at a local homestay for wildlife conservation between the relationship of anticipated emotion and intention to stay at a local homestay for wildlife conservation. These results are presented in Figure 5. Based on Baron and Kenny, if the indirect effect of axb < c, and there is a significant effect at both sides, then a partial mediator exists in the relationship. In the results of this study, axb = c and both sides were significant. These results reveal that there is a partial mediating effect of attitude to staying at a local homestay for wildlife conservation on the relationship between anticipated emotion and intention to stay at a local homestay for wildlife conservation; this supported the findings for Hypothesis H4. relation between the mediator variable and the dependent variable; and third, if the relationship between the independent variable and the dependent variable is significant, this means the existence of a partial mediator. The results reveal a partial mediating effect of attitude to staying at a local homestay for wildlife conservation on the relationship between anticipated emotion and intention to stay at a local homestay for wildlife conservation, which is supported by the findings for hypothesis H4. The indirect effect tests also show that there is a mediating effect of attitude to staying at a local homestay for wildlife conservation between the relationship of anticipated emotion and intention to stay at a local homestay for wildlife conservation. These results are presented in Figure 5. Based on Baron and Kenny, if the indirect effect of axb<c, and there is a significant effect at both sides, then a partial mediator exists in the relationship. In the results of this study, axb=c and both sides were significant. These results reveal that there is a partial mediating effect of attitude to staying at a local homestay for wildlife conservation on the relationship between anticipated emotion and intention to stay at a local homestay for wildlife conservation; this supported the findings for hypothesis H4. Recently, scholars have tended to confirm the result of a mediator using bootstrapping analysis. Bootstrapping is a non-parametric test based on resampling, with many instances of replacement, e.g., 1000. Based on the bootstrapping results in Table 6, the significant results for both the direct and indirect effects in this study have confirmed that attitude to staying at a local homestay for wildlife conservation partially mediates the relationship between anticipated emotion and intention to stay at a local homestay for wildlife conservation. Partial mediation since direct is also significant Recently, scholars have tended to confirm the result of a mediator using bootstrapping analysis. Bootstrapping is a non-parametric test based on resampling, with many instances of replacement, e.g., 1000. Based on the bootstrapping results in Table 6, the significant results for both the direct and indirect effects in this study have confirmed that attitude to staying at a local homestay for wildlife conservation partially mediates the relationship between anticipated emotion and intention to stay at a local homestay for wildlife conservation. Partial mediation since direct is also significant

Discussion and Conclusions
This research examined the mediator roles of attitude towards staying in a local homestay for wildlife conservation between the relationship of anticipated emotion and intention to understand ecotourist responsible economic behaviour. The study revealed that attitude towards staying in a local homestay for wildlife conservation has a significant direct effect on intention to stay in local homestays for wildlife conservation. This result is in line with studies conducted by Clark Mulgrew, Kannis-Dymand, Schaffer and Hoberg [78]; Phu, Hai, Yen and Son [79]; Meng and Choi [80]; and Gstaettner, Rodger and Lee [81]. The study also proves that anticipated emotion can influence attitude since the results indicate that anticipated emotion has a significant direct effect on attitude towards staying in a local homestay for wildlife conservation. These results support those of previous studies by Kim, Njite and Hancer [95] and Londono, Davies and Elms [96]. The study also shows that anticipated emotion can highly influence intention since the results indicate that anticipated emotion has a significant direct effect on intention to stay in a local homestay for wildlife conservation, which is in line with previous studies [95,96]. The study also confirmed that attitude is a mediator between anticipated emotion and intention, since the study revealed the mediating effect of attitude to staying in a local homestay for wildlife conservation between anticipated emotion and intention to stay in a local homestay for wildlife conservation. This result is also supported by the findings of previous studies by Gilchrist, Masser, Horsley and Ditto [102] and Taylor, Ishida, and Donovan [103]. The high variance in explaining attitude and intention shows that anticipated emotion plays an important role in understanding economically responsible ecotourists' attitudes and intentions to stay in local homestays for wildlife conservation. In practical terms, the current study has proven that a responsible ecotourist's attitudes and intentions are influenced by his/her anticipated emotions. Therefore, it is suggested that an emotional element is considered in advertising that promotes a local homestay. This would resemble the actions of major companies and organisations, e.g., Mars, Coca Cola, and the World Wildlife Fund, which have promoted their products and services in this way. For example, players in the ecotourism industry can use narratives and storytelling strategies to market a local homestay by creating an emotional connection between tourists and wildlife. Since colours have a scientifically proven association with human emotions, stakeholders could also practise this technique in their marketing strategies. Although the influences of colours vary according to gender, stakeholders could consider the power of colour itself to influence others. For example, black and purple are related to being strong, powerful and masterful; red is arousing; blue is soft. The colour Facebook uses is blue and that of Coca-Cola is red; this is no coincidence. Word-of-mouth techniques (through testimonials, customer reviews, logos, or real customer stories) are also important in instilling trust in a tourist. Trust will eventually produce a positive emotional reaction, such as 'If other people trust them, I should too'. Airbnb is well-known for using the simple yet strong word-of-mouth method. Since this study has determined that attitude mediates the relationship between anticipated emotion and intention, it is also suggested that players in the ecotourism industry conduct strategic planning on how to improve the attitudes of responsible ecotourists to wildlife conservation. The more positive the attitude to wildlife conservation is, the higher the chance that a responsible ecotourist will stay in a local homestay to help the local economy. An effective marketing strategy would induce responsible ecotourists to choose a local homestay rather than other accommodation.
To summarize, ecotourists are particularly responsible for wildlife conservation when they are willing to stay in local homestays and pay for their accommodations to aid the local economy through changes in their attitudes and emotions. This study has some limitations due to time constraints. Although this study found that attitude and anticipated emotion greatly influence behavioural intentions, the study did not consider the effect of behavioural intentions on the actual behaviour of responsible ecotourists. According to Ajzen [53], not all intentions have a significant effect on actual behaviour. It is recommended that future researchers conduct a study of economically responsible ecotourist behaviour using the Theory of Planned Behaviour.

Conflicts of Interest:
The authors declare no conflict of interest.