The Inﬂuential Factors of Consumers’ Sustainable Consumption: A Case on Electric Vehicles in China

: As one of the internationally recognized solutions to environmental problems, electric vehicles feature zero direct emissions and can reduce dependence on petroleum. An increasing number of countries have attached importance to the electric vehicle and developed it, and it is predicted that it will become a main force in the transportation system. Hence, it is necessary to explore the factors that drive consumers to buy electric vehicles. This study analyzes the factors that inﬂuence the consumer’s intention to buy electric vehicles and tests the relationship between them, and intends to o ﬀ er information for the formulation of policies designed to popularize electric vehicles in order to reduce carbon emissions from transportation. As a result, consumer attitudes are the most important factor inﬂuencing the intention to purchase electric vehicles. The greatest e ﬀ ect is found in this line: Brand Trust → Perceived Beneﬁt → Attitude → Purchase Intention. This means that the brand can increase the consumer’s perceived beneﬁt of electric vehicles, make consumers more attracted to electric vehicles, and inﬂuence their ﬁnal purchase intention.


Background
Environmental deterioration is an inevitable problem caused by human activity [1], and global warming, the greenhouse effect, and acid rain are all nature's warnings against mankind. Moreover, the petroleum crisis and excessive carbon emissions are also regarded as the most urgent challenges confronting the modern world [2], and the survival of human beings has come to a critical juncture. Against such a backdrop, the member states of the United Nations put forth the 2030 Agenda for Sustainable Development (see Figure 1) in 2015. This agenda serves as a blueprint of peace and prosperity according to the present situation and the future development of mankind and the Earth [3]. It is specified in Goal 12 of the agenda: "Responsible consumption and production: ensure that the international community will develop towards green growth and a recycling economy.".
As one of the internationally recognized solutions to environmental problems, electric vehicles feature zero direct emissions and can reduce the dependence on petroleum [4,5]. An increasing number of countries have attached importance to the electric vehicle and have developed it [6], and it is predicted that it will become a main force in the transportation system [7]. Many countries have set corresponding goals and formulated relevant policies [8], including China [9]. Driven by the policies, China has evolved into the largest electric vehicle market in the world [6]. In order to be in tune with the times, vehicle manufacturers across the world have begun to develop electric vehicles [10]. It is As one of the internationally recognized solutions to environmental problems, electric vehicles feature zero direct emissions and can reduce the dependence on petroleum [4,5]. An increasing number of countries have attached importance to the electric vehicle and have developed it [6], and it is predicted that it will become a main force in the transportation system [7]. Many countries have set corresponding goals and formulated relevant policies [8], including China [9]. Driven by the policies, China has evolved into the largest electric vehicle market in the world [6]. In order to be in tune with the times, vehicle manufacturers across the world have begun to develop electric vehicles [10]. It is estimated that there will be over 145 million electric vehicles on the planet by 2035 [11]. According to a test in New York, electric vehicles play a positive role in reducing air pollution in urban areas [12]; they can improve the air quality by reducing 20% of carbon emissions [13] and decrease noise.
Aside from the endeavors of governments and automobile manufacturers across the world, consumers are also a key factor that have contributed to the popularization of electric vehicles. The more consumers who use electric vehicles, the less petroleum consumed and the less CO2 emitted [14]. At the critical juncture of the transformation from the production of traditional cars to the development and production of electric vehicles in the automobile industry, consumers who buy electric vehicles will have an enormous impact on the spreading of electric vehicles and the development of the whole industry; hence, it is necessary to explore the factors that drive consumers to buy electric vehicles.
This study analyzes the factors that influence consumer intentions to buy electric vehicles and tests the relationship between them, and intends to offer information for the formulation of policies designed to popularize electric vehicles in order to reduce carbon emissions from transportation.

Electric Vehicle
The electric vehicle is a hot research topic at present, and an increasing number of studies on the electric vehicle market have been conducted at home and abroad. According to Ewing and Sarigöllü, the price, performance, usage cost, and time cost are the key factors that influence the purchase of electric vehicles [15]. However, consumers lack an adequate understanding of the overall impacts of electric vehicles on the environment and the cost to possess electric vehicles [16,17]. This is one of the reasons why electric vehicles have not been widely used. Another factor that affects popularization is the battery life of electric vehicles [18]; therefore, some consumers would choose a plugged hybrid automobile instead [19]. If the problem of battery life is resolved, consumers will have a stronger intention to buy electric vehicles [20]. Liao et al. classified and summarized the influential factors of consumer preferences, such as socioeconomic variables, psychological factors, mobility condition, and social influence [21].
There are many factors that influence the purchase of electric vehicles; in addition to the aforementioned factors, brand effect is another reason. Moreover, the correlation between these factors and the degree to which they influence the consumer's purchase intention are also the focuses of this study. Aside from the endeavors of governments and automobile manufacturers across the world, consumers are also a key factor that have contributed to the popularization of electric vehicles. The more consumers who use electric vehicles, the less petroleum consumed and the less CO 2 emitted [14]. At the critical juncture of the transformation from the production of traditional cars to the development and production of electric vehicles in the automobile industry, consumers who buy electric vehicles will have an enormous impact on the spreading of electric vehicles and the development of the whole industry; hence, it is necessary to explore the factors that drive consumers to buy electric vehicles.
This study analyzes the factors that influence consumer intentions to buy electric vehicles and tests the relationship between them, and intends to offer information for the formulation of policies designed to popularize electric vehicles in order to reduce carbon emissions from transportation.

Electric Vehicle
The electric vehicle is a hot research topic at present, and an increasing number of studies on the electric vehicle market have been conducted at home and abroad. According to Ewing and Sarigöllü, the price, performance, usage cost, and time cost are the key factors that influence the purchase of electric vehicles [15]. However, consumers lack an adequate understanding of the overall impacts of electric vehicles on the environment and the cost to possess electric vehicles [16,17]. This is one of the reasons why electric vehicles have not been widely used. Another factor that affects popularization is the battery life of electric vehicles [18]; therefore, some consumers would choose a plugged hybrid automobile instead [19]. If the problem of battery life is resolved, consumers will have a stronger intention to buy electric vehicles [20]. Liao et al. classified and summarized the influential factors of consumer preferences, such as socioeconomic variables, psychological factors, mobility condition, and social influence [21].
There are many factors that influence the purchase of electric vehicles; in addition to the aforementioned factors, brand effect is another reason. Moreover, the correlation between these factors and the degree to which they influence the consumer's purchase intention are also the focuses of this study.

Purchase Intention and Attitude
Consumer purchase intentions and attitudes have been discussed in many models, including Theory of Reasoned Action (TRA) [22], Theory of Planned Behavior (TPB) [23,24], and Technology Acceptance Model (TAM) [25]. In these models, attitude is interpreted as a personal inner experience that influences the consumer's purchase intention, and purchase intention is the tendency of consumer Sustainability 2020, 12, 3496 3 of 16 action [23]. In this study, consumer attitudes toward electric vehicles are believed to have an effect on their purchase intention. Therefore, the following hypothesis is proposed:

Hypotheses 1 (H1):
Attitude has a significantly positive correlation with the consumer's intention to purchase electric vehicles.

Perceived Benefit
The fundamental purpose of a trade is to achieve value [26]. For consumers, the perceived value of a product or service is one of major factors that influence the consumer's intention to purchase [27]. Perceived benefit is the perceived possibility of the positive result of a purchase [28]. As a cognitive emotion, it has positive impacts on the consumer's intention and behavior [29]. The consumer's perceived benefits of electric vehicles can be divided into financial and non-financial benefits. Regarding financial benefit, the subsidy for the purchase of electric vehicles is high [30,31]. Meanwhile, the zero petroleum consumption of electric vehicles and the good after-sale services of manufacturers are two of the reasons why consumers choose electric vehicles. In terms of non-financial benefits, the zero petroleum consumption of electric vehicles indicates that electric vehicles are environmentally friendly [32]. Moreover, electric vehicles feature zero noise, high technology, and steady acceleration [33]. Perceived benefit is one of the main factors that influences the consumer's purchase of electric vehicles [34]; hence, this paper proposes the following hypotheses: Hypotheses 2 (H2). Perceived benefit has a significantly positive correlation with the consumer's intention to purchase electric vehicles.
Hypotheses 3 (H3). Perceived benefit has a significantly positive correlation with the consumer's attitude towards electric vehicles.

Perceived Risk
Perceived risk was originally a research topic in the realm of psychology, and referred to the consumer's predicted negative effects regarding the purchase of a specific product [35]. It is usually in a negative correlation with perceived benefit [36]. As electric vehicles have not been widely used, many consumers are still biased against them [37] in terms of safety [38], reliability [39], and battery life [40]. These are the factors that affect the consumer's selection of electric vehicles. The less consumers know about the electric vehicle, the more biased they will be against it and the more negative effects there will be. In addition, consumers would be influenced not only by the perceived benefit, but also the perceived risk in their intention and behavior, and they would balance benefit against risk before making the final decision and seeking the best solution [41]. Therefore, this paper proposes the following hypotheses: Hypotheses 4 (H4). Perceived risk has a remarkably negative correlation with the consumer's attitude towards electric vehicles.

Hypotheses 5 (H5).
Perceived risk has a remarkably negative correlation with the consumer's intention to purchase electric vehicles.
Hypotheses 6 (H6). Perceived risk has a remarkably negative correlation with the consumer's perceived benefit of electric vehicles.

New Product Knowledge
New product knowledge has a great influence on the consumer's purchase intention [42], and the more that consumers know about a new product, the more they intend to purchase it [43]. According to Wang and Hazen, consumers with more knowledge of green products and value would be more efficient in using the products [44]. Hence, it is important for consumers to have knowledge of electric vehicles [45]. If consumers know more about electric vehicles, they will be more likely to purchase the product [46]. In addition, new product knowledge is related to perceived benefit [47] and perceived risk [48]. More knowledge can further offset perceived risk [49] and motivate consumers to believe that electric vehicles would create more benefits for themselves and society [46]. Therefore, this paper proposes the following hypotheses: Hypotheses 7 (H7). New product knowledge has a significantly positive correlation with the consumer's perceived benefit of electric vehicles.

Hypotheses 8 (H8).
New product knowledge has a significantly negative correlation with the consumer's perceived risk of electric vehicles.

Brand Trust
Trust is one of the factors that must be considered in the explanation of the consumer's behavioral intention, as it plays an essential role if there is uncertainty and risk [50]. Brand trust refers to the relationship between the consumer's perceived quality of a product or service and the brand and reputation of manufacturers [51]. If consumers have a higher perceived quality of a desired brand product, they will show more trust in the brand and less perceived uncertainty and risk. Most existing studies on brand trust are about food [50,52], business [53,54], and network media [55][56][57]. As the electric vehicle is an emerging industry, most electric vehicle manufacturers and models are not popular, with the exception of Tesla. However, this study believes that brand trust would have an effect on consumers and eliminate their perceived uncertainty and risk [50]; thus, trust would influence the consumer's benefit. Hence, this paper proposes the following hypotheses: Hypotheses 9 (H9). Brand trust has a noticeably positive correlation with the consumer's perceived benefit of electric vehicles.

Hypotheses 10 (H10).
Brand trust has a noticeably positive correlation with the consumer's perceived risk of electric vehicles.

Proposed Theoretical Model
According to the aforementioned, this study proposes the following model (see Figure 2), which is comprised of six dimensions-"Purchase Intention," "Attitude," "Perceived Benefit," "Perceived Risk," "New Product Knowledge," and "Brand Trust"-and ten relevant research hypotheses.

New Product Knowledge
New product knowledge has a great influence on the consumer's purchase intention [42], and the more that consumers know about a new product, the more they intend to purchase it [43]. According to Wang and Hazen, consumers with more knowledge of green products and value would be more efficient in using the products [44]. Hence, it is important for consumers to have knowledge of electric vehicles [45]. If consumers know more about electric vehicles, they will be more likely to purchase the product [46]. In addition, new product knowledge is related to perceived benefit [47] and perceived risk [48]. More knowledge can further offset perceived risk [49] and motivate consumers to believe that electric vehicles would create more benefits for themselves and society [46]. Therefore, this paper proposes the following hypotheses:  Hypotheses 7 (H7). New product knowledge has a significantly positive correlation with the consumer's perceived benefit of electric vehicles.  Hypotheses 8 (H8). New product knowledge has a significantly negative correlation with the consumer's perceived risk of electric vehicles.

Brand Trust
Trust is one of the factors that must be considered in the explanation of the consumer's behavioral intention, as it plays an essential role if there is uncertainty and risk [50]. Brand trust refers to the relationship between the consumer's perceived quality of a product or service and the brand and reputation of manufacturers [51]. If consumers have a higher perceived quality of a desired brand product, they will show more trust in the brand and less perceived uncertainty and risk. Most existing studies on brand trust are about food [50,52], business [53,54], and network media [55][56][57]. As the electric vehicle is an emerging industry, most electric vehicle manufacturers and models are not popular, with the exception of Tesla. However, this study believes that brand trust would have an effect on consumers and eliminate their perceived uncertainty and risk [50]; thus, trust would influence the consumer's benefit. Hence, this paper proposes the following hypotheses:  Hypotheses 9 (H9). Brand trust has a noticeably positive correlation with the consumer's perceived benefit of electric vehicles.  Hypotheses 10 (H10). Brand trust has a noticeably positive correlation with the consumer's perceived risk of electric vehicles.

Proposed Theoretical Model
According to the aforementioned, this study proposes the following model (see Figure 2), which is comprised of six dimensions-"Purchase Intention," "Attitude," "Perceived Benefit," "Perceived Risk," "New Product Knowledge," and "Brand Trust"-and ten relevant research hypotheses.

Definition and Measure of Variables
This study designed the items of the questionnaire according to the research theme and relevant literature. The definitions of variable operability and reference scales are shown in Table 1.

Research Variable Operability Definition Reference Scale
Purchase intention It refers to the possibility that consumers will purchase electric vehicles.
Wang et al. [46]; Han et al. [58] Attitude It indicates the consumer's actual attitude towards, and the evaluation of, electric vehicles.
Wang et al. [46] Perceived benefit It is the consumer's perceived possibility of the positive results of purchase.
Kim et al. [59]; Kim et al. [60] Perceived risk It represents the consumer's predicted risk of purchasing electric vehicles.
Wang et al. [46] New product knowledge It implies the degree to which consumers know about electric vehicles.
Wang et al. [46]; Han et al. [58] Brand trust It signifies the degree to which consumers trust the chosen brand in the purchase of electric vehicles.

Analysis of Pre-Test Questionnaires
A 7-point Likert scale was adopted for the pre-test questionnaire of this study. The pre-test questionnaire was conducted from 15 January to 3 February 2020, during which, 60 questionnaire copies were distributed and 49 questionnaires were retrieved. For more accurate research results, the reliability and items of the pre-test questionnaire were analyzed to remove irregular items and enhance the reliability and discrimination of items.
As shown in Table 2, with the exception of "Brand Trust," the Cronbach's α values of all dimensions were higher than 0.6, which indicated that all the dimensions were highly reliable. As the Cronbach's α of "Brand Trust" was higher than 0.6, after Item BT2 was removed, the item was deleted. Meanwhile, the Cronbach's α of "New Product Knowledge" would have risen without Item NPK5; thus, this item was also removed. The official questionnaire copies were distributed after the removal of the above two items.

Sample and Data Collection
The official questionnaire of this study was carried out on the Internet to collect data, and the subjects were from China. There were 24 items as the estimation parameters of the questionnaire, and 496 samples were collected. According to the study by Jackson [61], the ratio of estimation parameters to samples should be 1:20; thus, the collection of the questionnaire copies was stopped. After invalid samples were removed, the number of the remaining samples was 417, and was still higher than the minimum quantity of samples (1:10) [61]; hence, the remaining samples were used for data analysis in the later stage. The valid copies accounted for 84.1%. The data about the samples of the valid copies were statistically analyzed to obtain the information about the gender and age of the samples. The distribution of the demographic variables is shown in Table 3.

Convergent Validity
This study used AMOS v22.0 software for structural equation model analysis. Because a large number of studies have used AMOS for analysis, AMOS is proven to be a reliable structural equation modeling software. According to the research of Anderson and Gerbing, data analysis can be divided into two stages [62]. The first stage was the Measurement Model, where the Maximum Likelihood Estimation method was adopted, and the estimation parameters included factor loading, reliability, convergent validity, and discriminant validity. According to the studies by Hair et al. [63], Nunnally and Bernstein [64], and Fornell and Larcker [65]-and those by Chin [66] and Hooper et al. [67] to probe into standardized factor loading-to explore convergent validity, the standardized factor loadings of this study ranged from 0.441 to 0.917, as shown in Table 4, which were within the acceptable scope. This meant that most of the items were reliable. The composite reliabilities of the dimensions were between 0.672 and 0.917, and most were above 0.7, which met the criterion suggested by scholars and showed that most of the dimensions were internally consistent. The average variance extractions ranged from 0.407 to 0.736, and most were higher than 0.5 [63], which indicated that most of the dimensions had a high level of convergent validity.

Discriminant Validity
The results of Fornell and Larcker [65] were used to test the discriminant validity of this study. If the Average Variance Extracted (AVE) square root of each dimension was higher than the correlation coefficient between dimensions, it would mean that the model had discriminant validity.
As shown in Table 5, the AVE square root of each dimension in the diagonal line was higher than the correlation coefficient beyond the diagonal line; hence, each dimension of this study had a high level of discriminant validity.

Structural Model Analysis
The nine goodness-of-fit indices, as obtained in the study by Jackson et al. [68], are the most widely used in SSCI journals, and were adopted to report the research results of this study. Kline [69] and Schumacker et al. [70] suggested that the goodness of fit of the model should be evaluated with diverse goodness-of-fit indices, rather than with the p value alone. In theory, a lower "χ 2 " is better; however, as "χ 2 " is sensitive to the quantity of samples, "χ 2 /df" was utilized to facilitate the evaluation, and its ideal value should be lower than 3. In addition, Hu and Bentler [71] argued that each index should be separately evaluated, and that more rigorous model fit indices should be adopted to control the error of the dominant "I," such as the "Standardized RMR < 0.08" and "CFI > 0.90" or "RMSEA < 0.08". Finally, the Satorra-Bentler scaled chi-square test [72,73] was used to modify the chi-square different statistics and the model fit. The structural model fit of this study was as follows (see Table 6):

Path Analysis
As shown in Table 7  The power of "Perceived Risk (PR)," "Brand Trust (BT)," and "New Production Knowledge (NPK)" to explain "Perceived Benefit (PB)" was 40.4%; the power of "Brand Trust (BT)" and "New Production Knowledge (NPK)" to explain "Perceived Risk (PR)" was 3.3%; the power of "Perceived Benefit (PB)" and "Perceived Risk (PR)" to explain "Attitude (ATT)" was 44.6%; the power of "Perceived Benefit (PB)," "Perceived Risk (PR)," and "Attitude (ATT)" to explain "Purchase Intention (PI)" was 62.3%. It is obvious that the research results support the model and research questions of this study. Table 7 shows the normalization coefficient of the SEM model in this study. The higher coefficient implies that the independent variable plays a more important role in the dependent variable. With the exception of H5, H6, and H8, the remaining hypotheses of this model are valid. Figure 3 shows the influence between variables in the structural model.  Table 7 shows the normalization coefficient of the SEM model in this study. The higher coefficient implies that the independent variable plays a more important role in the dependent variable. With the exception of H5, H6, and H8, the remaining hypotheses of this model are valid. Figure 3 shows the influence between variables in the structural model.

Results and Discussion
This study utilized the structural equation model to determine the factors that influence the consumer's intention to purchase electric vehicles, draw conclusions, and give some suggestions, with the intention of offering information for the formulation of policies designed to popularize electric vehicles in order to reduce the carbon emissions of transportation. The results of the empirical analysis have revealed some important findings, which are discussed as follows.
H1 is valid, which means that attitude has a remarkably positive correlation with the consumer's intention to purchase electric vehicles. Moreover, the path coefficient is the highest, which shows that the consumers who have a more positive attitude towards the use and purchase of electric vehicles are more willing to buy the products [33,74]. The direct effect of attitude on intention is manifested in TRA [22], TPB [23,24], and TAM [26]. In addition, the consumer's awareness of environmental protection has gradually enhanced in recent years [75,76], which indicates that attitude is a supportive index for predicting the consumer's purchase intention.
H2 is valid, which implies that perceived benefit has a noticeable correlation with the consumer's intention to purchase electric vehicles. The fact that H3 is also valid means that perceived benefit has an obviously positive correlation with the consumer's attitude toward electric vehicles. Moreover, the path coefficient is relatively high, indicating that the consumer's perceived benefit would influence their attitude towards electric vehicles and their purchase intention. Instead of exerting direct influence on purchase intention, the consumer's perceived benefit affects purchase intention through attitude [77]. Consumers perceive that electric vehicles overtake traditional automobiles with a combustion engine for zero petroleum consumption, little pollution, and smooth movement [78], and that electric vehicles enjoy supporting policies, such as "better access to get a license plate" and "a higher purchase subsidy" [79]. Therefore, they have developed a positive attitude towards electric vehicles. Finally, other factors, such as environmental protection and petroleum price, would also influence the consumer's purchase intention. Thus, perceived benefit is a supportive index for predicting the consumer's purchase intention [80].
H4 is valid, which implies that perceived risk has a significantly negative correlation with the consumer's attitude towards electric vehicles. Neither H5 nor H6 are valid, which indicates that there is no noticeable correlation between perceived risk and the consumer's intention to purchase electric vehicles or their perceived benefit of electric vehicles. This also means that consumers will develop a more negative attitude towards electric vehicles if their perceived risk of electric vehicles is higher. However, the fact that there was no remarkable correlation between perceived risk and perceived benefit is inconsistent with the findings of previous studies [81]. The possible reason for this is that the consumer's perceived risk of electric vehicles is focused on the weaknesses of existing electric vehicles, such as low safety, a short battery life, and the long time required for charging [82], and is not directly connected with the consumer's perceived benefit of the strengths of the products, including zero petroleum consumption, little pollution, and smooth movement [78]. According to the results of this study, perceived risk does not have an immediate effect on the consumer's purchase intention; instead, it influences purchase intention through attitude. This indicates that these weaknesses (low safety, a short battery life, and a long time for charging) will be constantly reduced with the development of the electric vehicle industry and technological advancements; however, the electric vehicles still fail to meet the expectations of consumers. Worse still, the concern about the weaknesses will result in a more negative attitude from consumers, and then affect their purchase intention. Moreover, the dimension correlation also shows that the effect of perceived risk on attitude is far greater than that of perceived benefit on attitude, which means that consumers believe that the strengths of electric vehicles, including zero petroleum consumption (low cost) [83], little pollution (environmentally friendly), and smooth movement (user experience) [30] can offset the risk caused by the weaknesses. The reasons for this are as follows: firstly, most consumers drive electric vehicles in urban areas, which reduces the cost of petroleum consumption caused by traffic jams; secondly, there are many charging points in urban areas and these charging points are near to each other, which reduces the concern caused by the weaknesses.
H7 is valid, which means that there is a markedly positive correlation between new product knowledge and the consumer's perceived benefit of electric vehicles. H8 is invalid, which implies that there is no noticeable correlation between new product knowledge and the consumer's perceived risk of electric vehicles, and that the consumers who have more new product knowledge of electric vehicles would perceive more benefits. The new product knowledge of electric vehicles, as defined in this study, includes various strengths, such as comfort, a high accelerated speed, low pollution, and little noise [83], as well as other features, such as time of charging and application scope. The greater the new product knowledge of consumers, the more they know about the features (strengths and weaknesses) of electric vehicles. Meanwhile, consumers with more new product knowledge would be clearer about the strengths of electric vehicles compared with traditional automobiles with a combustion engine, and, thus, have more perceived benefit, which will influence their attitude towards electric vehicles and purchase intention. There is no marked correlation between new product knowledge and the consumer's perceived risk of electric vehicles, which indicates that the consumer's knowledge of electric vehicles may not reduce their concern. Regarding the previous paragraph, the perceived risk of electric vehicles comes from the weaknesses of electric vehicles (low safety, a short battery life, and a long time for charging). As electric vehicles are still in the stages of development at present, it is impossible to reduce the perceived risks of consumers with more new product knowledge before a better solution to the weaknesses of electric vehicles is found in the overall industry [84].
H9 is valid, which means that brand trust has a significantly positive correlation with the consumer's perceived benefit of electric vehicles, and the path coefficient is relatively high. H10 is valid, which implies that there is a noticeably negative correlation between brand trust and the consumer's perceived risk of electric vehicles; it also shows that the consumer's brand trust in electric vehicles influences their perceived benefits and risks, and this greater brand trust leads to more perceived benefits and less perceived risks. With the gradual development of the electric vehicle industry, some new brands have been formed, such as Tesla and NIO. Famous brands would reduce the consumer's perceived risk and increase their perceived benefit; consumers tend to trust the quality and service of well-known brands [85] and feel less concern. Moreover, brand trust has a greater effect on perceived benefit than new product knowledge, which indicates that consumers tend to trust the reliability created by brands rather than evaluating the benefit of electric vehicles with their own knowledge. This also implies that consumers tend to believe that the products of their favorite brands will give them a better experience than those equipped with the same functions and performance.
This study established 10 hypotheses overall, of which, seven are supported (H1-H4, H7, H9, and H10), which means the research model is acceptable in explaining the factors that influence the consumer's purchase decisions on electric vehicles. From this, it is known that consumers will consider a variety of factors when choosing whether to buy an electric vehicle, and the most influential factor is their attitude towards electric vehicles. In addition, their attitudes are affected by other factors, including perceived benefit, perceived risk, new product knowledge, and brand trust. These factors have different degrees of influence on consumer attitudes and electric vehicle purchase decisions. Perceived benefit and perceived risk, as factors that directly impact attitudes, greatly influence the consumer's final decisions. The consumer's perception on the strengths of the products, including zero petroleum consumption, little pollution, and smooth movement, in perceived benefit, and safety considerations, endurance ability considerations, or long charging time in perceived risk all reflect their cost considerations for electric vehicles [86]. These cost considerations include the cost of value, the cost of use, the cost of time, and the cost of risk of electric vehicles. In this study, among the perceived factors, new product knowledge and brand trust was selected for discussion to mine the influencing factors that consumers can perceive more intuitively from various cost considerations. From the research results, consumers have a certain degree of perception of new product knowledge and brand trust, which act on perceived benefit and perceived risk, which, in turn, affect attitude and purchase intention. This means that it is indispensable to enhance the product power of electric vehicles and carry out corresponding brand marketing and promotion to increase the consumer's brand trust in order to influence their attitudes and purchase decisions.

Conclusions
The greatest contribution of this study is that it has established a theoretical model regarding the factors that influence the consumer's intention to purchase electric vehicles through various dimensions, such as new product knowledge, brand trust, perceived risk, and perceived benefit. In addition, the relevant effect analysis of this study shows that all of the above dimensions exert direct or indirect effects on the consumer's intention to purchase electric vehicles. This study has aimed to explore the meaning of the consumer's perceived benefit and perceived risk with electric vehicles and selectively discuss the more intuitive possibilities, thus, setting up the foundation for subsequent in-depth research. Meanwhile, the conclusions of this study can be taken as reference information for governments, consumers, and those working in the field of electric vehicles to promote purchase and reduce the carbon emission of transportation [87].
According to the analysis results, consumer attitudes are the most important factor in influencing their intention to purchase electric vehicles, and the factors that influence attitudes include perceived benefit (direct and positive), perceived risk (direct and negative), new product knowledge (indirect and positive), and brand trust (indirect and positive). The greatest effect is found in this line: Brand Trust→Perceived Benefit→Attitude→Purchase Intention. This means that the brand can increase the consumer's perceived benefit of electric vehicles, make consumers more attracted to electric vehicles, and influence their final purchase intention. Meanwhile, the consumer's trust in the brand can also reduce their perceived potential risk of purchasing electric vehicles and contribute to their more positive attitude towards the products. The concern about the use of electric vehicles and their higher requirements are the main obstacles that affect the consumer's attitude towards electric vehicles and their purchase intention [84]. In addition, the consumer's knowledge of electric vehicles would create an indirect positive effect on their attitude and purchase intention; the more they know about electric vehicles, the clearer their understanding is of whether they need the product.
For most potential consumers of electric vehicles, the weaknesses of electric vehicles, low popularization, and inadequate demand are the reasons why they have not purchased electric vehicles. Therefore, this study offers the following suggestions: 1.
On government policies: (1) at least maintain the existing policies on purchase subsidy for electric vehicles in the near future and arouse the consumer's initiative to purchase electric vehicles; (2) popularize the knowledge of electric vehicles through market-oriented publicity and incentives; (3) encourage the manufacturers and enterprises of electric vehicles to generate better products, including batteries and engines; (4) cooperate with relevant enterprises to establish more charging points for electric vehicles.

2.
On electric vehicle manufacturers: (1) make greater effort to develop electric vehicles and increase functions according to government policies; (2) improve the purchase experience and after-sale services, such as encouraging consumers to take a trial drive and adopting a new marketing model that features the combination of online reservation and offline purchase; (3) organize more driver social activities on a more regular basis, such as holding electric vehicle track days and forming an electric vehicle culture in order to further the development of the electric vehicle industry and market.

Future Research Directions
The limitations of this study may indicate some future research directions.

1.
This study probed into the factors influencing the consumer's intention to purchase electric vehicles from the perspective of consumers; however, the effects of continuously improving government policies for consumers were not considered. Hence, future researchers can focus on this issue.

2.
This study did not analyze the samples according to their social or financial conditions; therefore, emphasis can be placed on the differences in attitudes towards electric vehicles and the purchase intentions among consumers from different regions, with different earnings, and of different ages in future studies. 3.
The correlation between some dimensions in the model of this study is not marked, which is probably because some latent variables or sub-dimensions were not explored. For that reason, future researchers can add new dimensions, including sub-dimensions and mediating variables, to improve the model by strengthening its explanatory power.

4.
This research has focused on the constructed model, without any in-depth or specific discussion on cost, price difference, performance difference, etc. In the future, a deeper discussion can be carried out on the basis of this research model.