What Drives Rural Consumers to Change E-Commerce Attitude and Adopt E-Commerce through the Moderating Role of Corporate Social Responsibility in an Emerging Market? An Empirical Investigation in the Chinese Context

Abstract

The present study aims to introduce a comprehensive framework that may help to better understand how to make rural consumers hold a favorable attitude and enhance their willingness to adopt e-commerce. To empirically assess the hypotheses posited in this paper, we started by conducting a qualitative interview-based study of 104 rural consumers. This analysis elucidates new problems or challenges faced by rural consumers in impoverished areas after several years of experience in e-commerce. To further understand the service quality and cultural context effects, we conducted a quantitative study in 434 rural consumers in relatively underdeveloped areas of China. Using a partial least squares of structural equation modeling (SEM) approach through smart PLS, this study empirically tested the hypotheses posited in the paper. The SEM results demonstrate a positive relationship between logistics and training service quality, subjective norms, self-efficiency sense, and rural consumers’ attitudes toward e-commerce platforms, which in turn positively contributes to their willingness to engage in word-of-mouth e-commerce promotion. In addition, it is shown empirically that corporate social responsibility positively moderates the effects of logistics and training services, subjective norms, and attitudes toward the use of e-commerce platforms. The findings from these two studies contribute to a better understanding of, and have major implications for, successful e-commerce entrepreneurial practices in areas undergoing the process of transition to an important emerging e-commerce marketplace.

1. Introduction

Poverty is a universal economic phenomenon. Nowadays, approximately 85% of the global population spend less than USD 30.00, two-thirds spend less than USD 10.00, and 10% spend less than USD 1.90 for daily living expenses [1]. Remarkable achievements have been made in reducing global poverty over the past few decades, but the downward trend has slowed down since 2010. Poverty eradication is the biggest challenge globally, and addressing it is an inevitable requirement for adhering to sustainable development [2].

Governmental departments and non-governmental organizations (NGOs) of different countries have attempted to eliminate poverty in various forms using different methods. Many international organizations, such as the United Nations and the World Bank, or national governmental agencies like the Department for International Development (DFID) in the United Kingdom have suggested that the government guide firms to fulfill their social responsibilities in poverty alleviation [3]. Corporate social responsibility (CSR) fulfillment helps solve unemployment, protect the environment and resources, narrow social gaps, and eliminate social instabilities. Guiding firms to fulfill their social responsibilities correctly is significant to promoting economic society’s sustainable development. At present, CSR fulfillment has been well-practiced. However, compared with the rapid economic growth, CSR fulfillment is not widely adopted to address many problems.

An epidemic has caught the world off guard. As various countries fight COVID-19, online shopping has become increasingly popular and inclusive of young people in urban and rural areas. Online shopping allows people to buy their favorite products online without leaving home and provides opportunities for people to start businesses in rural areas. E-commerce has extended to poor rural areas and brought new energy to poverty alleviation and rural revitalization, facilitating urban consumption and helping poor areas eliminate poverty and become rich. IT firms are proven to have a bright future in e-commerce poverty alleviation. They share resources and technologies throughout the internet, contributing to poverty eradication [4,5], and maintain normal business operations, even during an epidemic [6].

E-commerce poverty alleviation is a project promoted by information and communication technology, requiring multi-party and multi-pronged efforts to solve the poverty problem [7]. In China, e-commerce poverty alleviation is in full swing. E-commerce has been used in many poor areas and hoped to change the current poor living conditions. However, the opposite has happened, because many young and middle-aged people in rural areas work in cities, leaving behind a group of old people, women, and children who have limited knowledge about, and low enthusiasm for, e-commerce. In 2018, the e-commerce alliance helped Tailai County in Heilongjiang Province of China promote Tailai rice. Compared with previous years, its annual sales increased significantly. Accordingly, the income of 3400 poor people increased. However, peasants could not keep up with the brand’s development, given the lack of actual productivity, and the benefits of poverty alleviation have not been evident [8].

Poverty alleviation has entered a crucial stage. In the post epidemic era, popularizing ecommerce to the bottom of society, including poverty-stricken areas that are easily forgotten, and alleviating e-commerce poverty are essential to sustainable development. The previous literature has shown that e-commerce poverty alleviation, dominated by the government, should focus on cultivating potential participants [9]. An ecosystem should be established in poverty-stricken areas based on e-commerce platforms, and the system should meet the demand of self-sufficiency to alleviate poverty [7,10]. The previous literature did not pay attention to the attitudes of rural consumers toward e-commerce in poverty-stricken areas. It did not cover solving the remaining problems in e-commerce poverty alleviation. Thus, the objective of the present study is to explore the influencing factors on rural consumers’ attitude toward the use of e-commerce in poverty-stricken areas. It also aims to determine and understand the mechanisms that make rural consumers in poverty-stricken areas change their prejudice and attitude toward e-commerce to promote e-commerce poverty alleviation and benefit poverty-stricken areas. In doing so, this study proposes that logistics and training service quality, subjective norms, and self-efficiency sense will have a positive impact on rural consumers’ attitude toward e-commerce platforms which in turn is argued to positively shape the willingness of these consumers to engage in word of mouth (WOM). Furthermore, the study proposes that rural consumers’ attitude toward e-commerce may not be influenced equally by the logistics and training service quality, subjective norms, and self-efficiency sense, because they face different external conditions. In particular, this paper argues that rural consumers’ perceptions regarding the external force in local markets, namely, the CSR engagement of e-commerce firms, helps motivate rural consumers who perceive a high quality of logistics and training service, subjective norms, and self-efficiency sense to have a more positive attitude toward e-commerce and thus drive them to more actively engage in WOM.

Our paper offers three important contributions to literature. First, our study contributes to understanding the effect of quality of different services and consumers’ unique characteristics on consumers’ attitude toward e-commerce, which in turn is expected to positively impact the willingness of the consumers to engage in WOM. Second, this study explores the conditions under which such service quality and consumer characteristics contribute to a more positive attitude. In doing so, this study clarifies theoretically the positive influence of CSR on the contributions of logistics and training service quality, subjective norms, and self-efficiency sense to consumer attitude toward e-commerce. Furthermore, by explicitly considering the importance of consumers in rural areas in helping e-commerce firms better expand their business and by providing the first empirical test using both quality and quantity data, this paper proposes that the determinants of consumer attitude toward e-commerce may not be always universal. To sum up, our study conceptually and analytically links logistics and training service quality, subjective norms, and self-efficiency sense with consumer attitude and willingness to engage in WOM by incorporating the potential role of CSR in moderating such relationships, thus contributing to better understanding and having implications for successful e-commerce entrepreneurial practices in areas undergoing the process of transition to an important emerging e-commerce marketplace.

2. Theoretical Background and Hypothesis Development

Petty and Cacioppo first proposed the elaboration likelihood model (ELM) in 1986. It is regarded as a basic theoretical framework that deals with persuasive information to change original attitudes [11]. Central and peripheral routes trigger attitude change between individuals. In other words, people process information by receiving information and being persuaded to accept it. The attitude change caused by the central route is more lasting, stable, and predictable for long-term behavior. It is a choice made after thoughtful, rational thinking. On the contrary, the attitude change caused by the peripheral route is more unstable. It is just a short-term attitude change and is easily affected by opposite persuasion information.

For individuals, the difference may be that the receivers have a different understanding of the learning process and content [12]. The potential elaboration of the receivers is based on the information they receive to generate their ideas [13]. Individuals’ route choices depend on their elaboration degree of knowledge. High elaboration likelihood drives individuals to review carefully or deal with the quality of information deeply to accept a persuaded argument. The receivers evaluate the problems presented in the message to accept cognitive processing [14]. People with low elaboration likelihood are easily affected by peripheral clues. This influence leads to new cognition formation to change the original attitude [15].

The present study is largely motivated to explore how to persuade rural consumers in poverty-stricken areas to change their attitude toward e-commerce and recommend it to others. Given information persuasion theory, taking the ELM model as the typical example, this study uses ELM as the theoretical basis, combs the central and peripheral route factors in e-commerce poverty alleviation, and explores the influencing factors on the attitude change of rural consumers in poverty-stricken areas toward e-commerce use by combining with the characteristics of e-commerce poverty alleviation. This study investigates the attitude of rural consumers toward e-commerce poverty alleviation from the perspective of ELM. ELM can help us to understand how firms evaluate the attitude of rural consumers toward e-commerce poverty alleviation and their communication intention by using their features, and to change rural consumers’ dependence on donations.

Based on the ELM model, firm training and logistics services provided by e-commerce firms are used as the central route. In contrast, self-efficiency sense and subjective norms are used as the peripheral route. By further exploring the moderating effect of perceived CSR on the relationships between central and peripheral routes and attitude toward e-commerce use, this study holds that the contribution of central and peripheral routes to attitude of rural consumers toward e-commerce is likely to be dependent on the degree of perceived CSR. If the degree of perceived CSR of e-commerce is higher, rural consumers who perceive service quality highly are more likely to have a positive attitude toward e-commerce. Furthermore, such rural consumers are expected to be more likely to recommend the use of e-commerce use to others by actively engaging in WOM.

On the contrary, rural consumers are likely to prefer the peripheral route if perceived CSR is lower. Self-efficiency sense and subjective norms are higher, and attitude toward the ecommerce in use is easier to change. Although this adoption is a short and unstable change, it affects the intention to engage in WOM. On the basis of the ELM model, this study discusses in-depth the relationships between the perceived quality of services, the sense of CSR, and personal characteristics with changes in attitude toward the adopted e-commerce and intention to engage in WOM. On the basis of the characteristics of e-commerce poverty alleviation, the central and peripheral routes of e-commerce poverty alleviation are combed.

2.1. Perceived Quality of Training Service

One of the keys to successful firm operations is the quality of service [16]. High-quality service improves external customers’ positive recognition of the organization and constantly affects customers’ behavior to accept the organization’s service [17]. High quality of service can also increase the positive emotional responses of customers [18] and improve positive decision-making [19]. The quality of service that individuals perceive positively affects customers’ attitude and changes their original attitude of rejection [20,21,22]. Training service generally refers to the initial ability of the business operation of relevant technologies to teach skills or share knowledge for firm employees to maintain and upgrade their technical capabilities throughout their working career, promote relevant enterprises, and improve quality and efficiency [23]. To ensure that employees use the gained knowledge and skills, firms provide employees with training services [24] to enliven the service atmosphere [25] and improve the service quality [26]. Individuals must have professional knowledge in IT, upstream and downstream docking, and network marketing in e-commerce platforms. Rural consumers are unlikely to have such knowledge without receiving professional training. A large number of compound talents who are proficient in network technology and familiar with the operation law of the agricultural economy are required in the marketing of agricultural products. They need to spend a lot of time on learning and training. Furthermore, relevant cognition in training can positively impact employees’ commitment to the organization [27], affecting the service quality for rural consumers in the operation of e-commerce platforms [28]. Therefore, helping rural consumers with technology and improving their capabilities is the key to help them complete the operation of e-commerce platforms [29].

In poverty-stricken areas, e-commerce talents are scarce, and understanding e-commerce, information technology, and relevant knowledge is low. To make more rural customers adopt e-commerce platforms, firms must train them to improve their knowledge of e-commerce and functional skills to use e-commerce. Such training will enable rural consumers to take a positive attitude toward the use of e-commerce. If the perceived quality of e-commerce training services is higher, the attitudes toward adopting e-commerce platforms will be more positive. Thus, the following hypothesis is proposed:

Hypothesis 1 (H1).

The perceived quality of service provided by e-commerce firms positively impacts rural consumers’ attitudes toward e-commerce use.

2.2. Perceived Quality of Logistics Service

A logistics distribution system must match e-commerce activities. Poor rural areas are remote, with inconvenient transportation and low total logistics, resulting in high distribution costs, which often restricts the development of e-commerce and affects e-commerce poverty alleviation. Logistics service affects the profit and income of firms [30] and is the reason for various marketing products to compete [31]. Therefore, how to conduct the development and improvement of e-commerce logistics systems in poverty-stricken areas is one of the main factors affecting the development of projects for e-commerce poverty alleviation.

Logistics service quality generally refers to the service that allows firms to send goods to the correct places following accurate information regarding quantity and price, state, and time to fulfill customer needs and ensure customer satisfaction [32,33]. The key to improving logistics service quality is the interaction between firms and customers, with the customer at the center [34]. When a firm has high communication and information technology, improving logistics service quality positively impacts customer satisfaction [35,36,37]. Customer satisfaction is a critical factor for customers to adopt e-commerce. If the quality of logistics service is higher, the attitude of rural consumers toward e-commerce platforms will be more positive. Thus, the following hypothesis is proposed:

Hypothesis 2 (H2).

The perceived quality of logistics service provided by e-commerce firms positively impacts rural consumers’ attitudes toward e-commerce use.

2.3. Subjective Norms

Subjective norms reflect individuals’ social pressure perceived from the behaviors performed or not performed. Subjective norms often refer to the suggestions of individuals or groups, motivating them to take a specific action [38,39]. Individuals may adopt and act if their family, friends, relatives, and colleagues hope that they do so [40]. Subjective norms can affect use behavior. When individuals receive hopes, suggestions, and opinions from others, they will exert effort to perform [41,42]. In cross-cultural research, subjective norms also directly affect attitudes and directly and indirectly affect behavioral intentions [43,44,45]. In poverty-stricken areas, young adults go out to work, and most of the rural consumers are the elderly, women, and children. Given the constant attack of e-commerce on the real economy, online shopping and sales have become necessary. Subjective norms from the pressure of the social environment and the people around them are expected to impact rural consumers’ attitude toward ecommerce positively. To sum up, the following hypothesis is proposed:

Hypothesis 3 (H3).

The subjective norms of rural consumers positively impact rural consumers’ attitudes toward e-commerce use.

2.4. Self-Efficiency Sense

Self-efficiency sense is the belief that an individual can mobilize the motivation, cognitive resources, and action process required to meet the requirements of specific situations [46]. The core faith is that individuals can make a change by taking action [47]. People with high self-efficiency sense believe that they are more effective, successful, and healthy than those with low self-efficiency sense in goal achievement [48]. Self-efficiency sense will affect individuals to deal with the possible things in the future based on their capacities. E-commerce is new for rural consumers in poverty-stricken areas. Having a self-efficiency sense is particularly important to master the operation ability of e-commerce through learning. In other words, if the rural consumers have a stronger self-efficiency sense to e-commerce, the attitude toward ecommerce will be more positive. Thus, the following hypothesis is proposed:

Hypothesis 4 (H4).

The self-efficiency sense positively impacts rural consumers’ attitudes toward e-commerce use.

2.5. The Moderating Effect of Corporate Social Responsibility

No consensus has been made concerning the definition of CSR [49]. However, it is generally referred to as discretionary activities or actions of a firm that protects the interests of stakeholder groups or a larger societal collective by voluntarily contributing public welfare value to society to achieve the firm’s sustainable development [50,51]. Firms’ engagement in CSR activities can improve the reputation of firms and increase their profits [52], which positively impacts consumer buying behavior [53,54] and satisfaction [55].

CSR involves being responsible to shareholders and stakeholders [56]. CSR means that the firm’s stakeholders are treated in good ethics or in a socially responsible way. The stakeholders who have direct (board of directors, shareholders, and management) or indirect (employees, customers, and competitors) interest relations with the firm are groups or individuals that can affect or be affected by the achievement of organizational objectives [57]. Rural consumers in poverty-stricken areas are the largest potential customer group in ecommerce platforms. To change the rural consumers’ attitude toward e-commerce use, firms should enhance their CSR and fully meet the needs of consumers for quality services [58,59]. Thus, the engagement in CSR activities is expected to play an important moderating role in explaining the perceived quality of training and logistics services offered by e-commerce firms to the attitude of real personnel toward e-commerce use. On this basis, the following hypotheses are proposed:

Hypothesis 5a (H5a).

The degree of CSR positively moderates the relationship between perceived quality of training service and rural consumers’ attitudes toward e-commerce use.

Hypothesis 5b (H5b).

The degree of CSR positively moderates the relationship between perceived quality of logistics service and rural consumers’ attitudes toward e-commerce use.

Based on behavioral theory, when an individual’s subjective norms for a specific behavior are better, this person will strongly demonstrate the behavior [38]. In other words, the individual wants to develop the e-commerce platform in poverty-stricken areas and change rural consumers’ attitudes toward e-commerce use. A person’s attitude changes when firms influence stakeholders by persuading surrounding people to pressure the person concerned [60]. Firms should improve their social responsibility to encourage customers to take the initiative to convince the people surrounding rural consumers. If CSR is higher, firms’ image in society will be better [61], and consumers will be more willing to deal with firms with stronger CSR [62,63]. Furthermore, higher CSR causes rural consumers to have a positive attitude toward the firm’s e-commerce platform. Therefore, when CSR is higher, it can better adjust individuals’ subjective norms and make them more willing to take positive attitudes toward e-commerce use. In other words, CSR is predicted to play a positive moderating role in shaping the relationship between subjective norms and the attitude of rural consumers toward e-commerce use.

Self-efficiency sense is the confidence in the belief that an individual can mobilize the motivation, cognitive resources, and action process required to meet the requirements of specific situations [46]. Furthermore, self-efficiency sense plays an important role when facing challenges and goals [64]. When dealing with new things and new challenges, if self-efficiency sense is not enough, they do not invest too much energy and effort to try something new [65]. Generally, people can be motivated to behave under the external influence by observing and learning others’ behaviors to guide their subsequent behaviors [66].

At present, the pursuit of profit maximization is not the only criterion for business success. More and more firms use CSR practices to influence their main stakeholders’ cognition, emotion, and behavior [67,68]. Investment of firms in CSR activities can trigger consumers’ positive reactions [69]. In other words, whether e-commerce firms engage in CSR activities in poverty-stricken areas has been perceived by rural consumers plays a vital role in affecting the subsequent behaviors and reactions of these rural consumers. E-commerce firms can devote more resources to CSR activities by engaging in more ‘love and help’ activities, caring for socially vulnerable groups, and supporting rural consumers. This will make rural consumers more confident in using e-commerce, positively influencing the importance of self-efficiency in enhancing the real personnel attitude toward e-commerce use.

On the contrary, a negative perception of e-commerce firms’ CSR activities may exert a negative external influence on rural consumers. Rural consumers will probably abandon themselves, disconnect from society, and lose their confidence in e-commerce use, making them more reluctant to take a positive attitude toward e-commerce use. In short, rural consumers’ perceived degree of CSR positively moderates the contribution of self-efficiency to rural consumers’ attitudes toward e-commerce use. Therefore, the following relationships are proposed:

Hypothesis 5c (H5c).

The degree of CSR positively moderates the relationship between the subjective norms of rural consumers and their attitude toward e-commerce use.

Hypothesis 5d (H5d).

The degree of CSR positively moderates the relationship between the self-efficiency sense of rural consumers and their attitude toward e-commerce use.

2.6. Attitude and WOM Engagement Intention

Attitude refers to evaluating the likes and dislikes of perceived substances through the emotional response to facts [70]. It has a positive effect on behavior intention [38]. Especially when the attitude toward one thing is single, attitude will strongly impact behavior [71]. A good attitude can predict behavior intention [72,73,74,75].

According to Festinger’s cognitive dissonance theory, people’s cognitive system tends to maintain consistency. When cognitive dissonance happens, people will feel certain psychological pressure, driving people to reduce or eliminate the dissonance [76]. Rural consumers who have a positive attitude toward e-commerce use may regulate their cognitive system to be consistent to lower their psychological pressure. The intention to engage in WOM of rural consumers is predicted to be stronger when the cognition is more consonant; the intention to engage in WOM is expected to become weaker when cognition is more dissonant [77]. In other words, if the rural consumers’ attitude toward e-commerce use is positive, their intention to engage in WOM may become strong. Thus, the following hypothesis is proposed:

Hypothesis 6 (H6).

Rural consumers’ attitudes toward e-commerce use positively impact their intention to engage in WOM.

The research framework of this research is presented in Figure 1.

3. Methodology

3.1. Research Setting

The present study aims to understand the experimental achievements and specifically identify the problems of poverty alleviation in China’s e-commerce. This study explores what determines the attitude of consumers toward e-commerce, which in turn drives their willingness to engage in WOM in rural areas, especially those living in the poor areas of China. To empirically test the hypotheses, survey data on a sample of Chinese consumers from rural areas were collected. China provides an appropriate setting for our study. China has launched an incremental market liberalization through reform and opening-up to boost its economic development over the past four decades. Given the substantial progress in economic development of China in urban areas, the country has shifted its focus on balancing development and reducing economic disparities between urban and rural areas by specifically advancing and promoting rural revitalization and speeding up the modernization of rural areas. To achieve this, the Chinese government is rapidly putting in place more sound mechanisms and breaking barriers that restrict the free flow of resources such as capital, talents, and information between urban and rural areas. The Chinese government has exerted considerable efforts to develop the rural areas by improving the delivery logistics system services and encouraging e-commerce and logistics firms to better expand their business and support the development of local e-commerce business sectors in rural areas. In addition, as the e-commerce giants in China face slower growth in the saturated urban markets in recent years, they are increasingly looking for gold in rural areas. Despite the ongoing rapid urbanization over the past few years, around 510 million people are still living in rural regions, which makes up approximately 36.11% of the total population. The number of internet users in rural areas accounts for more than 26% of the total internet user population in China. Rural consumers have been quickly emerging as a lucrative new market opportunity for e-commerce giants. For example, Alibaba, the largest e-commerce giant in China, launched an initiative called rural Taobao in 2014 and invested USD 717 million on rural supply chain logistics, warehousing, and technology in 2018 to expand e-commerce business to villagers living in rural areas. This initiative has benefited the e-commerce giant and people living in rural areas by lifting more people out of poverty and improving their livelihoods. Similarly, the second largest e-commerce giant, JD.com, has committed to investing heavily in improving logistics services in rural areas of China by hiring more delivery personnel to ensure adequate dedicated delivery men in every village to receive parcels and take them to different houses, and even testing the delivering service with drones to rural areas since 2016. With the efforts made by the Chinese government to facilitate the flow of consumer products to rural areas and encourage e-commence giants to expand business in rural areas, the country has achieved remarkable development in revitalizing rural areas and alleviating poverty. In fact, China has made a huge contribution to global poverty reduction by lifting hundreds of millions of Chinese people out of poverty and transferring labor from rural areas to urban ones. Roughly 860 million people were living in urban areas by the end of 2020, and nearly 160 million of them were rural migrant workers. Against the background of the continuous progress of IT and digital technology, the e-commerce platforms in China have rapidly become a new means and key channel to revitalize rural areas and alleviate poverty. According to the statistics of the Chinese Ministry of Commerce (MOFCOM), there were 989 million internet users, with the internet penetration rate reaching more than 70% by the end of December 2020. Approximately 309 million of them were rural internet users. Various forms of poverty alleviation by e-commerce have been rapidly emerging in rural areas across the country. Currently, e-commerce has expanded to the countryside to achieve full coverage of 832 state-level poor counties. The online retail sales in these areas have reached more than 301 billion yuan (approximately USD 43.1 billion using the official exchange rate) in 2020, up by 20% on a yearly basis. More than 3 million e-commerce business operators were in these Chinese state-level poor counties by the end of 2020, which increased by 366,000 from 2019 (up by 13.7%). The e-commerce poverty alleviation has entered a stage of sophisticated poverty relief. The practical experience gained by China e-commerce poverty alleviation has also provided reference experience for the industry. The present rural consumers are largely different from urban consumers. As more than 160 million rural migrant workers were transferred from rural to urban areas, the consumers currently living in rural areas are relatively old people and may not be very familiar with online shopping. This study intends to understand the experimental achievements and specifically identify the problems of poverty alleviation in China’s e-commerce. The data were collected through surveys of rural consumers (liu shou ren yuan or rear personnel) in Leizhou City of Guangdong province.

3.2. In-Depth Field Interviews

To gain insight into the focal phenomenon, i.e., the latest rural consumers’ attitudes toward the adoption of e-commerce and important sources of their intention to engage in e-commerce WOM, a series of onsite in-depth interviews with the rural consumers in the area were conducted before administering the questionnaires. An e-commerce poverty alleviation project in a rural area located in Leizhou City in Guangdong Province of China was taken as a survey sample and conducted 104 onsite field interviews with rural consumers in the region to identify major problems or challenges that are faced by rural consumers in the early stage of e-commerce development and thus need to be urgently solved in e-commerce poverty alleviation. Our in-depth field interviews with rural consumers in the rural county located in Leizhou City of Guangdong Province in China suggest that, in general, the current e-commerce platform project was not carried out smoothly among the rural consumers in such poverty-stricken areas. One of the important reasons is that the construction of communication and other infrastructure is relatively weak in poverty-stricken areas. In terms of transportation, in the region where onsite field interviews were conducted, there was only one bus road there, which goes to the town directly, which is argued by the interviewers to significantly influence the batch transportation of goods. In terms of internet coverage, the region is relatively far and remote and has poor network signal. Nearly 70% of the rural consumers in the region did not have a computer at home, and approximately 40% of the rural consumers did not have a TV at home. Therefore, this region cannot reach the level of e-commerce development. Our interviews clearly demonstrate that the development of infrastructure construction dramatically affects the development of e-commerce in such poverty-stricken areas.

In addition, our interviews also revealed the fact that rural consumers in such poverty-stricken areas tended to lack awareness of the concept of e-commerce. Generally speaking, young people aged 20–35 should be the ‘principal force’ of online shopping. However, the young rural consumers aged 20–35 prefer to shop directly in a store. The reason may be that a perfect e-logistics system had not been established yet. The development of e-commerce must be based on perfect traffic conditions, a high-quality network environment, and some supporting infrastructure construction. Therefore, building an ideal network space and a sound logistics system is believed to be the foundation for e-commerce poverty alleviation to have a stable development platform under the background of ‘internet + strategy’. Given the lack of an e-logistics system, a small quantity of local agricultural products is sold. Therefore, they can only cooperate with the third-party logistics companies, but no fixed companies in cooperation virtually increase the costs of buyers and sellers. Another problem is the lack of professional e-commerce talent training. The results of our field interviews are summarized in

Table 1. Results of onsite filed interviews with rural consumers.

Selected Interview Questions	Major Concerns/Issues	%
Please list the main issues that you are most concerned about for e-commerce.	▪ training service	24.04
	▪ after-sales service	7.45
	▪ speed of logistics	18.27
	▪ convenient payment	4.57
	▪ CSR	20.67
	▪ subjective norms	14.42
	▪ self-efficiency sense	10.58
Have you heard of e-commerce?	▪ yes	72.12
	▪ no	27.88
Please list the primary reasons why you don’t use e-commerce.	▪ slow logistics	27.88
	▪ no home delivery service	17.31
	▪ poor after-sales service	14.74
	▪ unfamiliar to e-commerce	32.69
	▪ inconvenient online payment	7.38
Have you ever seen advertisements promoting e-commerce?	▪ often	9.62
	▪ occasionally	31.73
	▪ never	58.65
How often do you visit e-commerce platforms?	▪ once a week or more	16.35
	▪ every two weeks	8.65
	▪ once a month	30.77
	▪ only when in demand	40.38
	▪ hardly visits	3.85

and

Table 2. Major factor inhibiting the use of e-commerce for rural consumers.

Would You Please Tell Me Why You Do Not Choose E-Commerce?
“We are relatively poor here, with an underdeveloped network. E-commerce cannot go on well, even if it comes in. There were firms coming in before. They worked well at first. Later, they were basically of little use as soon as they left” (male, 52 years old, 16 January 2021).
“We do not know what e-commerce is and how to work or operate it. No one teaches us. It is too difficult for women who do housework at home all year round” (female, 35 years old, January 16, 2021).
“I have heard a little about e-commerce poverty alleviation. After all, I saw it on TV sometimes. However, there is a lot of online fraud at present. If no professionals teach us, or the firm here is not responsible, I am unwilling to try e-commerce. After all, it seems that the risk is more significant than the benefit. At least that’s how I feel” (male, 45 years old, 16 January 2021).
“A lot of farm work at home cannot be finished. I have to take care of the elderly and children. I do not have time to pay attention to this. In addition, I find that no neighbor is using it. I do not know what it can bring. Therefore, we usually like to go to the store to buy what we want” (female, 42 years old, 16 January 2021).
“I have to ride a bike to pick up the goods bought online in a town several kilometers away, let doing e-commerce alone. In addition, even if we do, the traffic here is inconvenient. There is only one cement road. It is a little difficult for a big truck to pass through” (male, 32 years old, 16 January 2021).

.

3.3. Sampling and Data Collection

To develop the survey instrument for the study, a careful process was followed by first developing an English-language version of the survey instrument and then translating it into Chinese with the help of two independent native speakers. To ensure conceptual equivalence and check for form and meaning accuracy, the Chinese version of the survey instrument was back-translated into English by two additional independent translators who are competent in Chinese and English. To ensure the content and validity of the measures, our survey instruments were pre-tested with 30 rural consumers (i.e., rear personnel) in the region. Basing on the feedback from the pre-tests, a few questionnaire items were further modified to ensure their relevance and clarity of the questionnaire items. Given the potential challenges faced in collecting sufficient primary data in China and the importance of the development of trust increasing high-quality responses, the survey was administered with the help of a renewed research company in the Chinese local market to encourage survey participation and enhance the high-quality response rate. Finally, a total of 466 questionnaires were collected. After excluding 32 irrelevant or incomplete responses, a total of 434 completed and usable questionnaires were retained for analysis. Among the respondents, 39.6% were male, 71.3% were married, around 75.6% had completed either a college degree or a high school or less education, and the average age was 41.3 years.

3.4. Bias Testing

To verify the presence of nonresponse bias in the data that may possibly influence our statistical results, the differences between responding and nonresponding consumers as well as the consumers that responded to our survey early against those that responded late were compared on key constructs and consumer demographic characteristics. To check for this potential threat to validity, the early-responding consumers were considered as the proxy for responding consumers and the late-responding consumers as the proxy for nonresponding consumers under the assumption that the consumers responded late are more similar to nonresponding consumers than those responded early to those nonresponding ones [78]. To further assess potential nonresponse bias, the differences between complete and partially complete questionnaires were also compared. Our results of a t-test show that there was so statistically significant difference between the consumers that responded to our survey early and those that responded late as well as the complete and partially complete questionnaires, suggesting that no evidence of a serious concern on response bias in our data [78].

In addition, as the use of self-report survey data may suffer from potential common method variance (CMV), appropriate questionnaire design and sample procedures were implemented to reduce potential CMV. First, when designing our survey instrument, the measures of our constructs were separated into several subsections and used different formats to mitigate a simple “straight line” pattern of response [79,80]. Second, to avoid the potential presence of CMV-biased response patterns, the order of the questions on the questionnaire were randomized using survey software and reversed the scaling on several questions. More importantly, all respondents were promised the strict anonymity and confidentiality of their responses in the survey cover letter. In particular, to reduce potential social desirability bias on self-report survey data, all respondents were informed that that there were no right or wrong answers to the questions, that they should simply respond to each question as honestly as possible, and that their responses would be only used for the sake of academic research. and which reduced the potential social desirability bias. Nevertheless, the possibility of CMV were examined by performing Harman’s one-factor test [81]. In doing so, exploratory factor analysis is performed by entering the variables in the study into a nonrotated factor analysis and the results of the one-factor analysis indicate that several factors, as opposed to one single factor, emerged and the first factor did not account for the majority of the variance.

Moreover, the presence of CMV was also assessed using the marker-variable technique [82,83]. A marker variable, i.e., attitude of the respondent toward blue color, was included in the model and significant relationships were not found between the marker variable and all latent variables in the model, and any significant change in the variance explained in the dependent variables. While CMV in the analysis cannot be completely ruled out, the results of a variety of techniques to empirically assess CMV suggest that CMV did not seem to be a significant concern for our analysis.

3.5. Variables and Measurement

In this study, well-established scales derived from prior research were used and modified specifically for this study to measure consumers’ attitude toward the adoption of e-commerce, their willingness to promote the use of e-commerce by engaging in WOM, their perceived logistics and training service quality, and the degree of their subjective norms and self-efficiency sense. Unless noted otherwise, all the dependent, independent, and moderating variables in the study were assessed using multiple-item, seven-point Likert scales ranging from “strongly disagree” (1) to “strongly agree” (7).

Consistent with prior research [84,85], three items were used to measure rural consumers’ willingness to engage in WOM for promoting the adoption of e-commerce. To measure consumers’ attitude toward the use of e-commerce, five items derived from prior research were used [86,87,88]. To measure the perceived quality of training service, ten items were adopted from the literature [89,90] and modified specifically for the study. As two items were dropped from the scale since they did not have a loading of at least 0.40 in the factor analysis, eight items were used to measure the quality of training service in the analysis. To capture the perceived quality of logistics service, nine items were adopted from Mentzer, Flint, and Hult [31]. As two items were dropped from the scale since they did not load onto the constructs of interest in the factor analysis, the quality of logistics service were thus measured using seven items. To measure the degree of subjective norms, eight items derived from prior studies were adopted [91,92,93]. To capture the degree of self-efficiency sense, eight items derived from the literature were adopted [94,95]. Lastly, to measure rural consumers’ perceived degree of the engagement in corporate social responsibility (CSR) of e-commerce firms, 10 items were adopted from Luo [96] and Isabelle [97] and modified them specifically for this study. As three items were dropped from the scale since they did not have a loading of at least 0.40 in the factor analysis, seven items were used to measure the perceived degree of CSR in the study.

4. Analyses and Results

In this study, partial least squares (PLS) structural equation modeling (SEM) was used to empirically test our research hypotheses [98]. Before empirically testing the hypotheses, the reliability and validity of the constructs were first assessed by checking the measurement model.

4.1. Measure Reliability and Validity

Table 3. Descriptive statistics and validity assessments of the constructs.

Construct and Indicators	Mean	STD	SFL	Cronbach’s Alpha	CR	AVE
Perceived logistics service quality (LSQ)				0.945	0.955	0.751
LSQ1	5.491	1.533	0.865
LSQ2	5.583	1.367	0.863
LSQ3	5.643	1.328	0.862
LSQ4	5.548	1.284	0.864
LSQ5	5.620	1.254	0.869
LSQ6	5.597	1.375	0.869
LSQ7	5.599	1.437	0.874
Perceived training service quality (TSQ)				0.932	0.944	0.678
TSQ1	5.295	1.563	0.837
TSQ2	5.290	1.490	0.827
TSQ3	5.129	1.436	0.840
TSQ4	5.281	1.433	0.835
TSQ5	5.698	1.457	0.787
TSQ6	5.023	1.696	0.804
TSQ7	4.843	1.789	0.806
TSQ8	5.350	1.520	0.848
Subjective norms (SN)				0.949	0.957	0.736
SN1	5.479	1.406	0.849
SN2	5.535	1.404	0.867
SN3	5.475	1.401	0.865
SN4	5.505	1.274	0.864
SN5	5.521	1.311	0.857
SN6	5.507	1.361	0.863
SN7	5.525	1.356	0.866
SN8	5.304	1.537	0.834
Self-efficiency sense (SES)				0.946	0.955	0.727
SES1	5.438	1.431	0.859
SES2	5.512	1.388	0.858
SES3	5.459	1.437	0.861
SES4	5.546	1.394	0.859
SES5	5.530	1.385	0.860
SES6	5.753	1.358	0.857
SES7	5.438	1.431	0.856
SES8	5.594	1.329	0.809
Perceived corporate social responsibility (CSR)				0.867	0.898	0.557
CSR1	6.366	1.046	0.782
CSR2	6.359	0.786	0.714
CSR3	6.410	0.802	0.717
CSR4	6.295	1.215	0.705
CSR5	6.286	1.122	0.704
CSR6	6.309	1.095	0.789
CSR7	6.353	1.022	0.806
Attitude toward the use of e-commerce (AU)				0.903	0.928	0.721
AU1	5.177	1.671	0.816
AU2	5.719	1.389	0.869
AU3	5.664	1.395	0.870
AU4	5.539	1.344	0.840
AU5	5.753	1.431	0.849
Intention of WOM engagement (WOM)				0.896	0.935	0.828
WOM1	5.399	1.353	0.912
WOM2	5.415	1.436	0.911
WOM3	5.371	1.392	0.907

Note: SFL = standardized factor loading, AVE = average variance extracted, CR = composite reliability, STD = standard deviation. Due to space limitations, detailed measurement items are omitted, but they are available from the corresponding author upon request.

presents the results of the measurement assessment, which summarizes the means, standard deviations, factor loadings, construct reliabilities, and the average variances extracted (AVEs). Since all the established scales were used to measure the variables in this study, all measures exhibit strong reliability and validity. In PLS, item reliability is assessed by examining item loadings. As shown in Table 3, all item loadings are statistically significant, with values greater than 0.80, and thus clearly in excess of the recommended threshold of 0.7 [99], providing evidence of high item reliability and strong reliability of our measurement model [100,101]. To assess the internal reliability, the Cronbach’s alpha of constructs and composite reliabilities were examined. As reported in Table 3, all the Cronbach’s alpha values, ranging from 0.867 to 0.949, and composite reliabilities, ranging from 0.898 to 0.957, are greater than 0.80, exceeding the 0.70 benchmark [102,103]. Therefore, our constructs exhibit strong internal reliability and convergent validity. The AVE values for the constructs were also calculated and the results showed that all values are above the recommended threshold of 0.50, indicating convergent validity and reliability of the measures in the study [102]. To assess discriminant validity, the square root of AVE of each construct were compared with correlation between the construct and other constructs in the model. As shown in

Table 4. Correlations and discriminant validity among the constructs.

Variables	1	2	3	4	5	6	7
1. Perceived logistics service quality	0.867
2. Perceived training service quality	0.536	0.823
3. Subjective norms	0.522	0.552	0.858
4. Self-efficiency sense	0.389	0.502	0.532	0.853
5. Perceived corporate social responsibility	0.448	0.334	0.435	0.363	0.746
6. Attitude	0.622	0.674	0.649	0.586	0.474	0.849
7. Intention of WOM engagement	0.488	0.590	0.629	0.531	0.343	0.753	0.910

Note: Values in italicized bold denote the square root of the AVE of each construct.

, the results confirmed that the square root of the AVE values calculated for each of the constructs along the diagonal is higher than the correlations between different respective constructs in the corresponding off-diagonal elements of the matrix, providing an adequate discriminant validity of the measures in our sample [102]. To further verify the discriminant validity of the measures, the loading values of each single indicator were also compared with the cross-loadings with other indicators and the results show that each indicator loading is higher than the respective cross loadings, again suggesting the discriminant validity of the measures is adequate. Moreover, the heterotrait-monotrait ratio (HTMT) of the correlations was checked [104] and the results indicated that all HTMT correlations values are not greater than 0.85, suggesting adequate discriminant validity for all constructs in the model. Lastly, following prior work [105,106], Stone–Geisser’s Q² was used to assess the predictive validity of the latent constructs in the model using and the results showed that the cross-validated communality and redundancy values are higher than zero, ranging from 0.436 to 0.448, indicating adequate predictive validity in the model [100,107]. Overall, all the constructs and their respective indicators therefore exhibit strong reliability and validity in the context of this study.

4.2. Hypothesis Testing

Following the measurement model estimation, the theoretical model was empirically tested and the results of hypothesis testing and the amounts of explained variance in consumer attitude toward the use of e-commerce and the willingness to engage in WOM were presented in Figure 2. The overall structural model fit was assessed by checking for the standardized root mean square residual (SRMR) value. Our model had an SRMR value of 0.048, thereby indicating an excellent fit [102]. To test the statistical significance of path estimates, Chin’s [96] approach was followed using a bootstrap approach. More specifically, the coefficient of determination R² and the path coefficient with their respective t-values were estimated. The R² values for the two endogenous variables (e.g., attitude toward the use of e-commerce and willingness to engage in WOM) indicate satisfactory explanatory power for our model (ranging from 0.567 to 0.640). Overall, the results presented in Figure 2 indicate that the constructs are largely related in the theoretically predicted manner. More specifically, the results show a significant positive relationship between all of the four determinants of rural consumers’ attitude toward the use of e-commerce, i.e., perceived quality of training service (b = 0.302, p < 0.01), perceived quality of logistics service (b = 0.253, p < 0.01), the degree of subjective norms (b = 0.238, p < 0.01), the degree of self-efficiency sense (b = 0.209, p < 0.01), and the attitude of rural consumers toward the use of e-commerce in China. Therefore, these results indicate that the perceived quality of logistics and training services, and the degree of both subjective norms and self-efficiency sense of rural consumers, as hypothesized, are key determinants of their attitude toward the use of e-commerce. These results thus provide strong support for Hypotheses 1–4. The effect size (f²) of each independent variable was also examined and the results demonstrated that the effect sizes of the four significant determinants of attitude range from 0.08 to 0.15, suggesting that the effect sizes in this study are moderate.

Furthermore, Hypotheses 5a–5d were tested by examining the possible role of the perceived degree of the engagement in CSR by e-commerce firms in moderating the relationship between the four drives of consumer attitude (i.e., perceived quality of training and logistics services, subjective norms, and self-efficiency sense) and rural consumers’ attitude toward the use of e-commerce. As reported in Figure 2, the path coefficients of the interaction terms between the perceived quality of training service (b = 0.040, p < 0.05), logistics service (b = 0.062, p < 0.01), the degree of consumer subjective norms (b = 0.031, p < 0.05), and the perceived level of the engagement in CSR by e-commerce players are positive and statistically significant, respectively. These results indicate that rural consumers who perceive higher quality of training or logistics service and display a tendency of a relatively larger degree of subjective norms are likely to have a positive attitude toward the use of e-commerce when e-commerce players are perceived to engage more in CSR activities. On the basis of these results, Hypotheses 5a–5c are strongly supported. However, as shown in Figure 2, the coefficient for the interaction of rural consumers’ self-efficiency sense and the degree of perceived CSR by e-commerce players is not significant. Therefore, Hypothesis 5d is not supported.

Finally, this study examined the contribution of rural consumers’ attitude toward the use of e-commerce to their willingness to engage in WOM. As shown in Figure 2, the path coefficients from consumer attitude to the willingness to engage in WOM is strongly significant and in the predicted positive direction (b = 0.753, p < 0.001), thereby providing strong support for Hypothesis 6. This result implies that rural consumers in China who have a positive attitude toward the use of e-commerce are more willing to promote the adoption of e-commerce by engaging in WOM. In the following section, these results and their implications were discussed.

As a check on the robustness of the SEM results, both regression analyses and SEM analyses with Amos approach were performed to verify the hypothetical relationships of the study. The findings of the regression analyses and the SEM analyses using Amos are effectively equivalent to our results achieved from the use of the PLS SEM estimation, providing strong support for our main results.

5. Discussion and Implications

5.1. Theoretical Implications

This study, which takes ELM as its theoretical basis, combines the characteristics of e-commerce poverty alleviation. In poverty-stricken areas, many local villagers choose to go out to work. Thus, this study focuses on a unique group of rural consumers: rear personnel, young and middle-aged women, children, and the elderly who stay in poverty-stricken areas. At first, through a series of in-depth qualitative interviews employed before the quantitative analyses of the impact of various factors on the attitude of rural consumers toward e-commerce, this study identifies the major factors predicting these rural consumers’ attitudes toward e-commerce platforms. Then, the central and peripheral routes affecting their attitudes were combed through a questionnaire survey. The central route includes training and logistics services. The peripheral route has subjective norms and self-efficiency sense.

This study explores the factors driving the positive attitude of rural consumers toward ecommerce use in poverty-stricken areas, which enhances these consumers’ willingness to help e-commerce firms promote e-commerce services and platforms by engaging in WOM. This study reveals that the improvement of the perceived quality of training and logistics by ecommerce firms positively affects rural consumers’ attitudes toward e-commerce use. Furthermore, subjective norms and the self-efficiency sense of rural consumers positively impact consumers’ attitudes toward e-commerce use. Our study demonstrates that CSR plays a positive moderating role in shaping the relationships between quality training and logistics, subjective norms, and rural consumers’ attitudes toward e-commerce use. Finally, the study finds that rural consumers who have a positive attitude toward e-commerce use become more willing to engage more in WOM by recommending e-commerce platforms to their relatives, friends, and surrounding people.

This study offers important contributions to the literature by extending the application scope of ELM theory. ELM theory can be applied to e-commerce poverty alleviation to persuade rural consumers to use e-commerce. Building on the original framework, this study adds two crucial factors: CSR and the willingness to engage in WOM. More specifically, this study considers the intention of rural consumers to engage in WOM, the dependent variable of ELM, proving that positive attitudes toward e-commerce use will encourage rural consumers to engage in more WOM. In other words, rural consumers who have a positive attitude toward e-commerce are more willing to engage in WOM by recommending other consumers to use e-commerce. This finding contradicts the result found in the literature that WOM plays a vital role in predicting attitude and behaviors [108,109]. Thus, it provides a theoretical basis for the formulation and implementation of e-commerce poverty alleviation policies and extends the application of the theoretical framework of ELM, which provides important implications for future studies.

Furthermore, there is a growing consensus that firms should undertake CSR activities for a long time. However, the questions of why firms should engage more in CSR activities and, more importantly, how they undertake CSR activities, are relatively unclear. CSR has been argued to influence firm reputation and performance [52], customers’ behavioral intention [54], and satisfaction [55]. This study finds that CSR can play a critical role in positively moderating training and logistics contributions and subjective norms to rural consumers’ attitudes toward e-commerce use. In doing so, this study contributes to the CSR literature by demonstrating the urgency and necessity of firms to engage more in CSR activities and understanding better the specific direction for e-commerce firms to promote additional investment in CSR. Our study extends the logic of ELM and combining the recent development of CSR to derive theoretically driven contingencies on rural consumers’ attitudes toward e-commerce use. Thus, our study provides a fine-grained analysis and framework that can be used to determine rural consumers’ attitudes toward e-commerce use and precisely assess the effect of such attitudes on the intention of rural consumers to engage more in WOM.

5.2. Practical Implications

The findings of our study also offer several important implications for practitioners. First, the perceived quality of the training and logistics services of e-commerce is found to positively predict the attitude of rural consumers toward e-commerce use. Therefore, to encourage rural consumers to form positive attitudes and readily accept e-commerce use, ecommerce firms should provide good and practical training opportunities in poverty-stricken areas. Such a training program can significantly enhance rural consumers’ understanding of ecommerce and improve rural consumers’ e-commerce operation capability. In other words, the perceived quality of e-commerce training service offered by e-commerce firms can be considered one of the key factors determining their success in expanding their e-commerce business to rural areas and participating in e-commerce poverty alleviation [29,30]. This result is generally consistent with the prior literature, demonstrating that improving the training service quality can facilitate customers to produce more positive emotional responses [18,19]. Once the rural consumers form a positive attitude toward e-commerce use, they will become an important group of stakeholders for e-commerce firms and play a critical role in helping ecommerce firms expand their business in the region [57].

The perceived quality of logistics service is also found to positively affect the attitude of rural consumers toward e-commerce use. This finding naturally dovetails with previous studies that argue for logistics services’ importance in directly affecting performance [30] and customer satisfaction [36,37]. In other words, when rural consumers form attitudes toward the use of e-commerce platforms, they tend to consider whether the logistics service or infrastructures in their areas can keep up or not. Only when the logistics corridor is unobstructed in poverty-stricken areas can rural consumers understand and predict the possibility and potential of e-commerce development in the region. Building on this logic and empirical finding, this study suggests that e-commerce firms continuously accelerate their efforts to improve and enhance logistics service in rural areas, especially in poverty-stricken areas. Our field interviews and empirical results indicate that the overall development of China’s logistics industry is still relatively underdeveloped. Therefore, problems of unbalanced and insufficient logistics service development and improvement are still prominent. As a result, the logistics service quality needs to be further improved and optimized, particularly logistics infrastructure in poverty-stricken areas.

Furthermore, subjective norms and rural customers’ self-efficiency sense positively impact the attitudes toward e-commerce use. Rural consumers, more specifically, rear personnel, are a special, yet vulnerable, group of consumers. Their attitudes toward a business or service are usually more affected by their neighbors, relatives, and friends. They prioritize survival and thus tend to ignore new things. E-commerce is still new to many rural consumers. Thus, e-commerce firms must change the attitude of the people surrounding rural consumers toward e-commerce use if they plan to change the negative attitudes of their target rural consumers toward e-commerce use. This argument is based on the assumption that changing the attitudes of people surrounding the target rural consumers can indirectly exert substantial influence on the target rural consumers’ attitude toward e-commerce use.

In addition, the empirical finding from this study also verifies the importance of rural consumers’ self-efficiency sense in shaping their attitudes toward e-commerce use. Just as everything ultimately depends on oneself, self-efficiency sense refers to a degree of confidence of a person that he can complete a task. For example, even if rural consumers with low self-efficiency sense know how to use e-commerce, they may not be competent and confident in their e-commerce operating skills, resulting in negative attitudes toward e-commerce use. On the contrary, consumers with a high degree of self-efficiency sense are likely to have a positive attitude toward e-commerce use when they receive guidance in using e-commerce. Such guidance makes them believe they can be competent in doing e-commerce. As a result, rural consumers with a high self-efficiency sense are expected to understand how to deal with their pressures or challenges in e-commerce use, form positive attitudes, and finally take great efforts to learn and try new things, such as e-commerce [110,111,112].

Lastly, the results of our study offer additional evidence to confirm the assumption that e-commerce firms’ engagement in CSR activities plays a positive moderating effect on the relationships between training and logistics services, subjective norms, and the attitude of rural consumers toward e-commerce use. Our findings suggest that e-commerce firms should engage in more CSR activities when they aim to expand their business or service in relatively poverty-stricken areas. Such CSR engagement can meet their practical requirements and improve ecommerce service quality. Subsequently, rural consumers would realize that some pressure stemmed from their subjective norms. They would easily form positive attitudes or change their negative ones and actively engage in e-commerce use.

This study did not find a significant moderating effect of the CSR engagement on the contribution of rural consumers’ self-efficiency sense to their attitude toward e-commerce use. The reason may be that many rural consumers are elderly, women, and children. They generally have less contact with the outside world and are distant from e-commerce firms. E-commerce firms may be willing to engage in more CSR activities. However, such efforts may not significantly moderate the contribution of rural consumers’ self-efficiency sense to the attitude of these rural consumers toward e-commerce use given their low self-confidence in using new things. Ecommerce firms should strengthen cooperation with respective local governmental agencies or institutions and make substantial efforts to exert their influence in poverty-stricken areas.

More importantly, local governments in these areas should also actively guide e-commerce firms to engage more in various programs of poverty alleviation and revitalization to encourage e-commerce firms to devote more resources to CSR activities in poverty-stricken areas. E-commerce firms in poverty-stricken areas will enhance their reputation when rural consumers in these regions have positive attitudes toward e-commerce use. Such attitudes help e-commerce firms achieve a better promotion effect through the relevant and effective impact of WOM, bring many rural migrant workers back home, and encourage them to start new businesses.

5.3. Limitations and Future Research Avenues

This study is not without its limitations, which we believe may provide fruitful avenues for future research. First, given the data constraints, this study examined the framework using a sample of rural consumers in a limited number of poor counties. Our study thus could not empirically verify our conceptual framework in all poverty-stricken areas. Researchers are encouraged to extend our study by examining our research questions and framework using a sample of rural consumers in other poverty-stricken areas. This study focused on the factors determining the attitude and willingness to engage in WOM of rural consumers in China, the largest emerging economy in the world. Future research may fruitfully extend our study to other countries and examine what shapes consumer attitudes toward specific service or products, such as local food, or even different distribution channels for purchasing or promoting such services or products [113,114], as well as WOM engagement in other economies. For example, emerging economies may be vastly different from advanced economies in terms of industrial policies, poverty alleviation efforts, poverty alleviation objects, and poverty alleviation conditions within or across various regions in the country. Employing comparative research, examining the determinants of attitude and WOM engagement between emerging and advanced economies or across emerging economies, and linking such differences to the differential levels of factors of service and consumer-specific characteristics in these economies would be interesting [113,114,115]. Lastly, given that some rural areas are still in an early stage of e-commerce poverty alleviation, many efforts of e-commerce firms need some time to be effectively implemented and put into action. The predominance of e-commerce poverty alleviation would also need some time. Future studies should adopt longitudinal data or experimental designs to consider the dynamics between resources and efforts devoted by ecommerce firms, government involvement, and consumer attitude and behavior change, which would be a fruitful avenue for future research. Overall, we believe that this study may provide new insights into how to predict consumer attitude and behaviors in e-commerce firms’ capability to improve training and logistics services and better understand their rural consumers in rapidly emerging economies.