Abstract
This study endeavors to delve into the intricate study of public preferences surrounding green consumption, aiming to explore the underlying reasons of its low adoption using social media data. It employs the Elaboration Likelihood Model (ELM) and text data mining to examine how information strategies from government, businesses, and media influence consumer attitudes toward green consumption. The findings reveal that women and individuals in economically developed regions show more concerns for green consumption. The public responds positively to government policies and corporate actions but negatively to media campaigns. Engagement with information and emotional responses influence attitudes toward green consumption. Subsequently, this study offers strategies for policymakers and businesses to enhance consumer attitudes and behaviors toward green consumption, promoting its development. Moreover, the innovative aspect of this study is the combination of ELM theory and text data mining techniques to monitor public attitude change, applicable not only to green consumption but also to other fields.
Similar content being viewed by others
Introduction
2021 is the first year of “carbon neutrality”. Driven by the “dual-carbon goal”, the task of China’s low-carbon green transformation has become urgent. At present, consumption is the main motivation of China’s economic growth. Enhancing people’s awareness of saving and promoting consumption to achieve a green and low-carbon transformation is of great significance for reducing carbon emissions. If the public is promoted to realize green consumption from the demand side, it is inseparable from the collective efforts of all populations of society. It is not only necessary for the government to improve the system and mechanism of certification management and incentives but also it requires enterprises to exercise their influence on users and take advantage of technology and platforms, boosting users to make sustainable consumption decisions. At the same time, social organizations are also an important force in promoting public green and low-carbon consumption. They can play a role in publicity, advocacy, and supervision in promoting low-carbon consumption, and enhance the public’s awareness of participation, to stimulate the enthusiasm of the public for green and low-carbon consumption. Although the government, enterprises, social organizations, and other entities have taken some measures to promote green consumption and have tried to persuade the public to adopt green consumption behavior through information strategies, convincing the public to engage in altruistic green consumption is complicated. Does this information arouse the public’s attention to green consumption? Does it enhance the public’s awareness and understanding of green consumption? Can it effectively stimulate the formation of public green consumption attitudes? It has become the focus of relevant stakeholders such as the government and enterprises. Social media text data is really valued for businesses and governments. By analyzing consumer sentiment in social media posts and comments, businesses can gauge consumer attitudes toward their products and make necessary improvements. Consumers use social media to gather information about product quality, price, and environmental impact, influencing their purchasing decisions. The viral nature of social networks allows quality products to spread quickly, transforming consumers into information producers and resource sharers. This social influence can shift consumer behavior and attitudes, impacting green consumption. This study aims to explore how entities like governments and businesses can leverage information strategies to influence green consumption attitudes, using social media text data as a real-world source for large-scale testing and validation of the Elaboration Likelihood Model (ELM) theory. Based on this, it is of great practical significance for policymakers and business managers to explore the current situation of public attention to green consumption and then study the driving factors of their green consumption attitudes.
Current scholars attribute green consumption willingness to subjective motivation and the influence of the external environment (Xia et al., 2023; Toder et al., 2023; Lee and Haley, 2022). Subjective motivation includes self-consumption incentives consisting of behavioral perception and self-efficacy (Testa et al., 2021; Shehawy, 2023), and the external environment such as laws and regulations, and social and technological development (Yan et al., 2020). Existing studies emphasized that public attitude is an important antecedent of consumers’ behavioral willingness (Dhir et al., 2021). The ELM model points out that people’s cognition and attitude mainly depend on whether individuals are exposed to exhaustively processed information (Cacioppo and Petty, 1984). External environmental information can create opportunities and constraints for forming individual attitudes and behaviors. With this intricate dynamic, environmental behavioral willingness, such as green consumption, is the outcome of the interaction between individual attitudes and external environmental information. Hence, a deep understanding of external environmental information can explain consumers’ willingness more comprehensively (Canio et al., 2021). At the same time, mining the public green consumption attitudes can find the key factor affecting the change in attitude and enhance the public’s perceived effectiveness of green consumption. However, most previous studies have used questionnaire surveys or empirical research methods (Cheung and To, 2019; Zaremohzzabieh et al., 2021) to explore the relationship between individual green consumption attitudes and external environmental information, but the scale of their data samples and research audience range have limitations. With the development of computer technology, text data mining technology can extract valuable information from a large amount of text data, build a theoretical framework based on the content generated by netizens or media, combine traditional statistical analysis with text data mining, and analyze the impact of external environmental information on the public green consumption attitudes from more comprehensive data sources and a wider range of audiences. Various theories have been employed to understand consumer behavior, including the Theory of Planned Behavior (Gabisch, 2011; Eid et al., 2021), Self-Determination Theory (Huang et al., 2016), Technology Acceptance Model (Li and Chen, 2019), Use and Satisfaction Theory, and Diffusion of Innovations Theory (Kim et al., 2020). While these theories have effectively predicted consumer motivation, behavioral intention, and technology adoption, their explanatory scope falters when explaining how external information influences consumer attitude change and decision-making. Notably, they often concentrate on either high-level or low-level cognitive processing, leaving a gap between the two. In particular, the Self-Determination Theory excessively emphasizes intrinsic motivation while neglecting external factors; the Technology Acceptance Model lacks depth in explaining individual acceptance of new technology; the Use and Satisfaction Theory and Diffusion of Innovations Theory are relatively limited in considering the diversity of cognitive changes. Additionally, each of the aforementioned theories focuses on either advanced or basic cognitive processing. The ELM, however, addresses both aspects concurrently, earning it frequent citations in green consumption literature. The ELM’s primary aim is to consider all information elements, acknowledging that individuals vary in emotions, abilities, and motivations and may not fully comprehend a message before deciding or thoroughly examining all message elements. Regarded by many researchers as the most prevalent model in social psychology, consumer behavior, and decision-making research (Teng et al., 2015; Verweij et al., 2015), the ELM aids in crafting persuasive messages for consumers, and taking into account their capabilities and motivations. At the same time, ELM is commonly used to delineate the processes of information reception and attitude formation. It effectively explains how individuals process, select, and receive vast amounts of information and the subsequent impact on their attitudes. This article employs the ELM framework for quantitative analysis, offering insights into consumer attitude shifts post-social media information reception. Previously, there was no clear categorization of information and its associated interest subjects, which overlooked the public’s perception differences. Hence, it’s essential to use text mining technology to analyze the influence of information released by various subjects on public attitudes towards green consumption. This approach seamlessly merges the ELM framework with text mining technology, which is capable of processing extensive unstructured data and understanding consumer attitude changes from multiple angles and levels. As such, it provides valuable insights for businesses, governments, and media organizations.
The main contributions of this paper are as follows: (1) considering that the public in different regions can freely express their views on green consumption on social media, their emotion is active, interactive, and anonymous, which enables them to completely express their views. Based on this, this paper retrieves the news and online comments on daily green consumption on social platforms such as Baidu News and Toutiao after the proposal of a “dual-carbon goal”, and comprehensively studies the stakeholders’ efforts to promote residents’ green consumption behavior. (2) Since green consumption behavior is reflected in all aspects of the public’s daily life, only analyzing green consumption texts may lose some key factors. Therefore, this paper divides the text into government information, enterprise information, and media information according to different stakeholders. It is conducive to mining the key clues under diverse information sources, collecting and analyzing the public’s feedback on the actions of different subjects promptly, and discovering the existing problems, to reveal the driving factors of the different green consumption attitudes under that information. (3) In the “dual-carbon” policy context, we propose a novel methodological framework that integrates the ELM and text mining techniques. This approach addresses the limitations of previous green consumption research, where qualitative analyses were often constrained by small data samples, and text mining lacked theoretical backing. This study examines how external information influences consumer emotions and attitudes from a psychological perception standpoint, applying the framework to text mine online reviews of green consumption. Concurrently, the ELM framework is used to quantitatively analyze shifts in consumer attitudes under the impact of various factors, such as the external economic and policy environment. Thus, it effectively expands the current research methodology on the public’s willingness and attitude towards green consumption and provides new insights for predicting the public’s attitudinal tendencies.
The structure of this paper is as follows: " “Literature review and grounded theory” is a literature review; “Proposed methodological framework” explores the current situation of public attention to green consumption with coverage and online comments; “Text mining of online comments of green consumption” analyzes the differences and changes in public green consumption attitudes under the measures released by different subjects by constructing dictionary semantics; “Analysis of public green consumption attitudes under different information contents” discusses the theoretical contributions and management implications of this paper; “Discussion and conclusion” is the summary and prospect of the full text.
Literature review and grounded theory
Consumer attitude, a key predictor of individual behavior (Joshi and Rahman, 2019), not only shapes how consumers perceive companies and their offerings but also influences their intentions (Khan et al., 2022; Shi and Jiang, 2022) and significantly impacts purchasing behavior (Li et al., 2021; Oliver and Christina, 2023; Kim and Lee, 2023). Thus, a thorough analysis of consumer attitudes can guide future green consumption strategies. This paper examines the literature in two areas: text mining of green consumption attitudes and factors influencing these attitudes.
Text mining of green consumption attitudes
Green consumer behavior research is a fusion of various theories and methods, each contributing to the complex task of defining green consumer behavior attributes (Barbu et al., 2022). Traditional studies have used questionnaires or interviews to investigate public motivation, buying willingness, and factors influencing green consumption across daily life aspects (see Table 1). However, these regional surveys, often limited by research questions and sample sizes, struggle to explore diverse groups’ perceptions and attitudes toward green consumption, raising questions about their objectivity and validity. Text mining methods offer a more objective insight into public attitudes towards green consumption, mining online comments for more realistic information. Thus, this paper leverages text mining techniques on large-scale online social platform data to analyze public focus, emotional tendencies, and understanding of green consumption.
Influencing factors of green consumption attitudes
According to the Technology Acceptance Theory, public concern, perception, and emotion toward green consumption are shaped by internal factors like subjective norms and perceived behavioral control. External environmental factors, such as economic incentives, environmental technologies, and social initiatives, can stimulate rational perceptions. Existing literature has explored the influence of public green consumption willingness or attitude in both individual internal and external contexts (see Table 2).
These studies focus on environments shaped by single external information sources like policy, enterprise, or social pressure. Within these environments, they construct models of public green consumption willingness or behavior based on classical behavioral theories, which are then validated through questionnaires or empirical research. The public is exposed to external environmental information that does not come from a specific or single content but receives information conveyed or created by various stakeholders. In addition, it can be seen from the definition of attitude, which includes cognition, emotion, and behavioral tendencies. When considering the public’s future wishes, the public’s cognition, characteristic perception, and emotional preference for green consumption should also be considered. Therefore, this paper amalgamates content from news reports encompassing information from the government, businesses, media, and other entities. It tracks the interaction between the public and these entities on social networks, mining multi-level content about their consumption cognition, emotional preferences, and behavioral tendencies from consumer-generated content. This approach enables a multi-dimensional exploration of public attitudes towards green consumption.
To sum up, the following are the premise for studying the public’s green consumption intentions: how the public understands the existing green consumption field, whether the external information conveyed by all sectors of society is widely understood and accepted by the public, what aspects the public pays attention to external information, and whether the public can change the green consumption attitudes by combining their own cognition and external information change. Moreover, mining the public green consumption attitudes from a large number of sample data can effectively avoid the limitations due to the demographic characteristics or selection factors of the research subjects. At the same time, it provides guiding suggestions for the government, enterprises, and other relevant stakeholders on how to promote the formation of public green consumption attitudes in the later stage. Based on this, this paper proposes a framework that mainly uses social media information to analyze the public green consumption attitudes. On the one hand, this paper analyzes the demographic characteristics of people actively searching for green consumption-related information. On the other hand, it deeply mines the reports related to public green consumption on news websites after the proposal of the “dual-carbon goal”, and traces their public online opinions on Weibo, divides Weibo post information from different relevant stakeholders, and conducts subject classification and emotional analysis on the comments under different post information. At the same time, it statistically analyzes the demographic characteristics of commenters and compares the differences with the people actively searching for green consumption-related information. Finally, based on ELM theory, by constructing a semantic dictionary, using semantic analysis and a generalized linear hybrid model, it discusses the changes in public green consumption attitudes, and puts forward relevant suggestions on the way to guide consumers to participate in green consumption.
Proposed methodological framework
This paper aims to develop a methodological framework, integrating the ELM and text mining, to quantify the impact of information strategies by governments, companies, and media on consumers’ attitudes towards green consumption. The model accounts for the diverse ways consumers process stimuli when exposed to information, altering their perceptions and understanding of green consumption and engaging in information diffusion. The model can categorize and quantify changes in consumers’ emotional, cognitive, and behavioral dimensions under varying policy, economic, and social media environments. It is also capable of addressing practical issues such as quantifying the geographical scope, gender, attractiveness, and focus content of a policy or product’s audience group. Furthermore, it can assess whether an information strategy can enhance consumption willingness, form a positive public opinion network, and contribute to positive consumer attitudes. This framework, while applicable to green consumption, can be extended to other fields by combining ELM and textual information to monitor public attitude changes, offering new perspectives and tools for understanding and predicting consumer behavior.
Based on our analysis, we present the following propositions for this study’s methodological framework:
Hypothesis 1: Most consumers perceive information strategies (e.g., governmental, corporate, and media messages) as influencing their attitudes toward green consumption.
Hypothesis 2: Consumers process stimuli differently when exposed to information, altering their perceptions and understandings of green consumption and engaging in information diffusion.
Hypothesis 3: Consumers’ attitudes toward green consumption may vary by geographic region and gender, and these attitudes may evolve over time.
Hypothesis 4: Changes in green consumption attitudes are analyzed at cognitive, behavioral, and affective levels by quantifying three factors: “information attention and understanding,” “information participation mode,” and “belief and attitude change.” We assume a positive linear relationship between these three factors.
Hypothesis 5: Under various policy and economic environments, the public’s attention, understanding, participation, and attitudes toward green consumption information strategies will fluctuate and adapt to the environment.
These propositions guide our framework’s design and operation, enabling a deeper exploration of the various influences on green consumption. The overall methodological framework is organized in four steps (as shown in Fig. 1):
The figure illustrates the implementation procedure of the proposed framework based on ELM and text analysis.
The first step is retrieving text data, which is mainly derived from official news and online comments.
The second step is preprocessing the collected data. Since text data collected on the Internet often contain noise, special characters, and other meaningless icons, which affect the effectiveness of text analysis, this step will remove stop words, nulls, symbols, special characters, affixes, etc., ignoring low-frequency words while dividing sentences.
The third step is conducting text mining on the preprocessed data. First, divide the relevant stakeholders of the text information through Convolutional Neural Network (CNN), to understand which subjects have taken corresponding measures to promote the formation of the public’s green consumption attitude (see “Topic analysis” below). Second, use Latent Dirichlet Allocation (LDA) to classify topics, to subdivide the focus of the topic content published by each type of subject. Third, construct an Long Short-Term Memory (LSTM) model to divide the public’s emotional orientation, and put the extracted emotional features into a fully connected network for visual analysis, to clarify the differences in the effects of topic information released by different subjects on improving the public green consumption attitudes (see “Emotional analysis and co-occurrence network analysis” below). Fourth, due to the differences in environmental and regional characteristics among the public, as well as the cognitive differences in the information content released by different subjects, a descriptive analysis of the gender and region of the public is carried out (see “User gender and region analysis” below).
The fourth step is analyzing the impact of information on the formation of public green consumption attitudes based on the ELM model and the above-mentioned text mining content. This includes constructing factors that affect public attitudes based on ELM, using the LIWC dictionary and statistical unsupervised algorithms to quantify the value of factors, and obtaining the fixed and random effects of each factor through the Generalized Linear Mixed Model (GLMM) model test, to clarify the changes in public green consumption attitudes (see “Analysis of public green consumption attitudes under different information contents” below). Using the proposed step-by-step method framework, this paper can identify the public’s true emotions toward green consumption and can compare the differences in public attitudes toward different subject initiatives. In addition, the analysis results are easily interpreted through the visual data presented at each stage.
Text mining of online comments of green consumption
Topic analysis
This section first crawls the news, official account, date, and other information on “green and low-carbon consumption” released from 2019 to 2022, and selected the content related to the public. A total of 6425 posts is divided into three categories: government, enterprise, and media. According to the content of news, the Weibo post and online comments under the news are mined, and 12,696 posts are randomly selected. There are three types of topics through the main coding: government policy, corporate measure, and media promotion. A tagged dataset is obtained through coding, among which 3177 posts are “government policy”, 5058 posts are “corporate measures”, and 3177 posts are “media promotion”. Finally, a classification model is trained through text CNN to automatically classify the original Weibo posts, and the classification accuracy of the final trained model is tested by a test set, and the accuracy rate of the training dataset reaches 66.7%. Based on the classification of topics in the original Weibo posts, to deeply mine the topic content discussed under each category of topic, LDA is used to classify the online comments under the three categories. The focus of the topic content published by each subject is subdivided, which is convenient for later analysis of the public’s emotional preference toward green consumption. The results are shown in Table 3.
LDA is used to cluster the classified online comments and government topics are divided into 6 categories according to feature words, mainly including “environmental protection and the plastic reduction”, “anti-food waste”, “garbage sorting”, “low-carbon travel”, “vehicle carbon reduction”, and “green clothing”. As evident in Table 3, the discourse surrounding “environmental protection and the plastic reduction” mostly refers to the restriction of using plastic products such as plastics and disposable products in supermarkets or hotels in daily life; “Anti-food waste” refers to regulating the phenomenon of food waste through rules and regulations, and advocating the public to implement the “Clean your plate” campaign; “Garbage sorting” refers to the government’s promotion of garbage classification in the context of carbon neutrality, and the use of garbage sorting stations; “Low-carbon travel” refers to the government’s innovation of the carbon-inclusive model, which includes launching individual carbon accounts, issuing green consumption coupons for daily subway or bus trips, and providing corresponding certificates to carbon reduction experts, and promoting the public to actively participate in low-carbon life; “Vehicle carbon reduction” mainly refers to cancel the purchase restriction policy of used cars and the purchase restriction of new energy vehicles; “Green clothing” mainly uses environmentally friendly renewable materials to improve the material of existing clothing, and recommends that institutional units purchase clothing moderately.
The online comments under corporate measure are divided into five categories, including “Express carbon reduction”, “Environmentally friendly food delivery”, “Environmentally friendly plastic reduction”, “Green home appliances”, and “Green office”. As evident in Table 3, “Express carbon reduction” is recycling cartons or establishing personal express to reduce environmental pollution caused by the express delivery industry; “Environmentally friendly food delivery” focuses on plastic pollution generated by food delivery industries such as ELEME Takeout and Meituan; “Environmentally friendly plastic reduction” mostly refers to advocating the use of electronic labels to reduce product packaging pollution and reduce the use of free plastic shopping bags provided by supermarkets and other retail companies. “Green home appliances” refers to upgrading and promoting existing energy-saving technologies for home appliance companies; “Green office” refers to reducing the use of consumables in daily work and strengthening environmental awareness.
There are seven categories of online comment topics under media promotion: advocate low-carbon life, environmentally friendly plastic reduction, express carbon reduction, takeaway tableware, food waste, energy-saving appliance, and green transportation. As evident in Table 3, “advocate low-carbon life” refers to promoting low-carbon behaviors in daily life to the public, enhancing environmental awareness, and making up for the public’s awareness of green and low-carbon; “Environmentally friendly plastic reduction” suggests that the public reduce the use of plastic products, enterprises reduce packaging waste, and the government controls mask pollution during the epidemic; “Express carbon reduction” reports the recycling activities of express delivery industries such as Cainiao Post Station, calling on the public to actively participate in express recycling; “Takeaway tableware” refers to reducing the use of disposable tableware in takeaways; “Food waste” describes the phenomenon of young people buying seasonal food, and popularizes the relationship between food waste and green and low carbon, to illustrate the harm of wasting food; “Energy-saving appliances” describes the way to use home appliances in life for realizing energy conservation and emission reduction; “Green travel” advocates the use of public transportation to reduce carbon emissions.
By extracting the feature words of the online comment content under the three categories, the topic classification is obtained. Among them, “environmental protection and plastic reduction”, “express delivery and carbon reduction”, “takeaway environmental protection”, and “food waste” are all mentioned frequently in policies, corporate measures, and media promotion reports. “Green travel” and “garbage sorting” have mainly received more attention from the government and the media, and “energy-saving appliances” are related to enterprises and media. In addition, celebrities such as “Zhang Yunlong” and “Chen Xuedong” have been mentioned frequently in the context of corporate topic comments, indicating that companies promote green and low-carbon consumption by using the celebrity effect, which has a strong leading role for fans. Fans tend to emotional consumption, promoting to build consumers’ trust in green and low-carbon consumption.
Emotional analysis and co-occurrence network analysis
The above-mentioned topic clustering of online comment content is carried out. To analyze the specific emotional preference of the public of each category under green consumption, an LSTM (Long Short Term Memory) deep learning model is constructed, and the public’s emotional preferences are divided into positive and negative, which can be expressed in percentages. The LSTM model can maximize the exploration of the relationship between different words and emotions in sentences, which can be used as a fitting tool for further feature extraction. Finally, extracted features are put into a fully connected network for visual analysis, and the emotional analysis results and the semantic co-occurrence network are obtained (as shown in Tables 4–6 and Figs. 2–7 below).
This network depicts how government policy topics influence the expression of positive emotions during the study period.
This network shows how government policy topics influence the expression of negative emotions throughout the research period.
This network depicts how corporate measures influence the expression of positive emotions during the study period.
This network represents how corporate measures influence the expression of negative emotions throughout the research duration.
This network demonstrates how media promotions influence the expression of positive emotions during the period of study.
This network portrays how media promotions influence the expression of negative emotions throughout the research period.
To analyze the impact of government policies, corporate measures, and media promotional information on public green and low-carbon consumption, this section conducts emotional analysis and semantic analysis on the comment information under the content of government, enterprise, and media. The aim is to study whether their measures can positively contribute to public’s recognition of green and low-carbon consumption. It is evident in Tables 4–6 that the public holds positive emotions toward government policies and corporate measures as a whole, and they highly support the government’s policies to promote green consumption. The public has expressed their willingness to actively participate in cultivating a green environment and low-carbon home and enhancing self-environmental awareness. However, they hold negative emotions toward the content of media reports as a whole. Among them, the government policy recommends the use of environmentally friendly recycled materials to produce clothing and advocates schools, governments, and other units to reduce the purchase of clothing. The improvements and effects of relevant clothing companies have not been reported by social media. The public’s attention to such information is also less. In addition, as is evident in Figs. 1–6, environmental protection and plastic reduction are mentioned on the topic of government policies, corporate measures, and media promotion. After the government proposed policies related to environmental protection and plastic reduction, enterprises upgraded shopping bags to degradable materials and simplified product packaging. The media advocates that residents bring their shopping bags by popularizing the effective use of popular science environmental protection bags. Compared with the publicity of policies and corporate measures, media promotion can positively enhance the public’s awareness of green and low-carbon consumption. Enterprises make efforts to simplify packaging for environmental protection, which has received strong public support. The public holds positive emotions toward the green and low-carbon improvement measures proposed by the express delivery, takeaway, home appliances, and office supply industries. The anti-food waste law issued by the government did not positively enhance the public’s recognition of green and low-carbon consumption. It indicates from the comments that the public acquire little information about specific implementation rules, and the policy content information is too broad and cannot be associated with green and low-carbon consumption. The media conveyed the practical significance of the policy to the public by releasing topics such as the relationship between food waste and low carbon, and the way to reduce food waste in daily life, thereby enhancing the public’s positive emotions towards green and low-carbon consumption. It shows that the effect of topic information released by different subjects on raising public awareness of green and low-carbon consumption is diverse.
User gender and region analysis
While the public may access pertinent information on the online platform, the information content of government policies, corporate measures, and media promotion actually perceived by the public varies due to geographical restrictions. Factors such as environmental differences, regional differences, and cognitive differences among the public also contribute to this divergence. In light of these considerations, this section analyzes the gender and region of users under different topics according to the commenters’ information. Due to the wide range of participating public areas, the region is divided into seven categories as Northeast-NE (Heilongjiang, Jilin, Liaoning), North China-N (Beijing, Tianjin, Hebei, Shanxi, Inner Mongolia), Central China-C (Henan, Hunan, Hubei), East China-E (Shandong, Jiangsu, Anhui, Shanghai, Zhejiang, Jiangxi, Fujian, Taiwan), South China-S (Guangdong, Guangxi, Hainan, Hong Kong, Macau), Northwest-NW (Shaanxi, Gansu, Ningxia, Qinghai, Xinjiang), Southwest-SW (Sichuan, Guizhou, Yunnan, Chongqing, Tibet), the results obtained are shown in Figs. 8–10.
a Represents the geographical distribution of commenters on government policy topics, b represents the gender distribution of commenters on government policy topics.
a Represents the geographical distribution of commenters on corporate measure topics, b represents the gender distribution of commenters on corporate measure topics.
a Represents the geographical distribution of commenters on media campaign topics, b represents the gender distribution of commenters on media campaign topics.
By analyzing the basic information of commenters on the topic of green and low-carbon consumption in two aspects: gender and region, it is evident in Figs. 8–10 that women participate more in green and low-carbon consumption on the Internet. The public in East China pays the highest attention to government policies, corporate improvement measures, and media promotion, while the public in Northwest China pays the lowest attention to that information. In addition, the public in North China, followed by South China pays attention to green consumption policies. Enterprises promote consumers’ green and low-carbon behavior by optimizing green technology, strengthening green publicity, and promoting environmental recycling, attracting the attention of the public in North China, Central China, and Southwest China in turn. The media popularizes green and low-carbon knowledge to the public, and guides and promotes the public to reduce carbon emissions from daily clothing, food, housing, and transportation, which improves their awareness of green consumption. The public with high awareness of this information comes from central and northern China. The limited scope of information receivers and the different levels of public attention to information will affect the judgment of whether their behavior is green and low-carbon and then weaken the awareness of green, low-carbon, and environmental protection.
Analysis of public green consumption attitudes under different information contents
This section selects ELM as the theoretical basis to analyze the differences in public green consumption attitudes under different topics and uses the dependent variable “change in beliefs and attitudes” to express it. Combined with the above analysis and the level of public involvement and the policy and economic environment, the factors that affect the public green consumption attitudes are divided into 8 categories (as shown in Table 7): information content category, individual attention and understanding of information, way of participating in information, policy environment, economic environment, zone, sex, and time.
Dictionary preparation and analysis
The construction of the dictionary primarily relies on the content of existing online comments. By using LIWC dictionaries (Linguistic Inquiry and Word Count) and statistical unsupervised learning algorithms, the process aims to extract insights expressing mental states from people’s language. There are a total of 1255 words expressing the concept of “attention and understanding of information” in a dictionary (e.g.,: understand, question, good, have a law to follow, choose, realize, experience.). There are a total of 1145 words expressing the concept of “ways to participate in information” (e.g.,: support, participate, start with me, like, forward, refuse to waste, join, promote, act, unquestionable, guarantee, advocate), and a total of 1645 words expressing “changes in belief and attitude” (e.g.,: anger, disappointment, confidence, worry, suspicion, idea, environmental awareness, consideration, everyone’s responsibility, attention to environmental protection, acceptance). To show the positive linear relationship between “attention and understanding of information”, “ways to participate in information”, and “changes in beliefs and attitudes”, a 2D density map is used to represent the granularity of the percentages of three types of data. The results are shown in Figs. 11–13.
a Represents the distribution of “Certaion-Cogmech”, b represents the distribution of “Certain-Affect”, c represents the distribution of “Cogmech-Affect”.
a Represents the distribution of “Certain-Cogmech”, b represents the distribution of “Certain-Affect”, c represents the distribution of “Cogmech-Affect”.
a Represents the distribution of “Certain-Cogmech”, b represents the distribution of “Certain-Affect”, c represents the distribution of “Cogmech-Affect”.
Figures 10–12 show a positive linear relationship between “attention and understanding of information”, “ways to participate in information” and “changes in beliefs and attitudes” with logarithmic data densities clustered around the origin. Figures 11a, 12a, and 13a indicate that the linear relationship between “attention and understanding of information” and “ways to participate in information” is the steepest. Among them, the public’s “changes in belief and attitude” and “attention and understanding of information” under policy and media topics have a stronger positive linear correlation, compared with the weaker linear correlation between “ways to participate in information” and “changes in belief and attitude”. This also shows that the public’s attention and understanding of green and low-carbon consumption information will directly affect the way they participate in information and the change of green consumption attitude.
Policy environment and economic environment
Policy effectiveness measure
In addition to the role of the above-mentioned policy-related reports, the impact of policies on public awareness and behavior toward green consumption also depends on the policy content itself, including the intensity of policy implementation, specific measures, and implementation goals. This paper selects consumer-related policy content from 2019 to 2022 and uses existing literature to quantify the impact of the policy environment (Li et al., 2020), as shown in Table 8.
According to the measurement criteria in Table 8, through formula (1), the PMC (Policy Modeling Consistency) indicators of different policy documents are calculated.
where i is the publication year of the analyzed policy, n is the summary of green consumption policies released by different agencies in the i year, j is the jth policy promulgated by the regulatory agency in the i year; mj, gj, and pj refer to the goals, measures, and intensity scores of the jth policy respectively; PMCi refers to the overall policy intensity of the policy in the i year. Combined with the policy document on the official website, the following results can be obtained in Table 9:
Consumer confidence index
The active engagement of the public in green consumption and the elevation of their self-awareness of environmental protection depends not only on the guidance of multi-subject measures of the public’s awareness and behavior but also on the level of external social and economic development. The Consumer Confidence Index (CCI) comprehensively reflects and quantifies consumers’ evaluation of the current economic situation and their subjective feelings about economic prospects, income situations, income expectations, and consumption psychology. Therefore, to measure the impact of the external economic environment on the public green consumption attitudes, the consumer confidence index from 2019 to 2022 is selected for measurement, as shown in Table 10.
Generalized linear mixed model analysis
To analyze the public’s green consumption attitudes, this paper uses a multi-level hybrid model to study the changes in the public’s green consumption attitudes. Since the measurements of Cogmech, Certain, and Affect are based on count data in the dictionary, instead of the variables based on continuous data, the traditional linear hybrid model is unsuitable. The Figs. 11–13 depict that the variables and the data there are correlated and are concentrated at the origin. Therefore, the Generalized Linear Mixed Model (GLMM) is used to study the public’s attitude toward green consumption. In GLMM, the linear predictor η is given by a combination of fixed and random effects: η = Xβ + Zγ.
In the above formula, X is the matrix of N*p, which represents the fixed effect independent variable; β is the vector of p*1, which represents the fixed effect parameter vector; Z is the matrix of N*q, which represents the random effect variable; γ is the vector of q*1, which represents the effect value of the random factor. g(.) represents the link function, which connects the predictor variable and the result variable. h(.) is the inverse function of g(.), then there are: g(E(y))=η, E(y)=H(η)=u, therefore, y = h(η)+ε.
y = Xβ + Zγ + ε, ε is the vector of N*1, representing residual. In addition, to eliminate the differences between the characteristics of each variable and make the results more reliable, the consumer confidence index and policy effectiveness index are processed and the subject content under the three topics of government, enterprise, and media are recorded as Z1–Z6, Q1–Q5, and M1–M7, respectively.
Model construction
In this article, the “incremental increase” strategy proposed by Raudenbush and Bryk (2002) is adopted: step 1: test unconditionally or only the first-level model, which contains only the first-level fixed intercept and random effects related to the second-level unit. This fundamental model allows to evaluate changes in longitudinal data response values before introducing covariates. Step 2: add the first-level covariates to the first-level model, and add the random effects of the first-level covariates to the second-level model, the first-level covariates are used to capture the growth effect. Moreover, an extension is made to the second-level model, the random changes in first-level covariates are focused. Step 3: add second-level covariates and fixed effects to the third-level model. If any covariates are not supported by the data, they will be removed, and the action model will be tested. Compare the fitting indices of the different models at each step to ensure that the final research model has the lowest index. Utilizing SAS to conduct “gradual increase” training on data under government, enterprise, and media topics. According to the covariance index, t test for parameter estimation of fixed effects (t-test: fixed effects), t test for random effect estimation (t test: random effects), F test for all fixed effect parameter estimates (F test), residual logarithmic approximation of the estimated model (-2Res Log Pseudo-Likelihood), residual sum of squares of the final model (Generalized Chi-Square), average as variability observed by the model (Gener. Chi-Square/DF), a mixed model related to “attention and understanding of information”, “ways to participate in information” and “changes in beliefs and attitudes” are elaborated below.
Analysis of public green consumption attitudes under government topics
According to the above analysis, the related models of variables Cogmech, Certain, and Affect are gradually constructed. The results of the model and related indicators are shown in Table 11 below:
According to the statistical results of the Cogmech model, with the time change, individuals’ attention and understanding of policy information have changed significantly, especially in 2021 (β10 = 0.022, P = 0.014 < 0.05). The model outcomes reveal notable insights, indicating that the main effects of time and CCI are not statistically significant, while the Cogmech effect is significant. When individuals’ attention and understanding of policy information increases, the public’s awareness of green consumption has also improved, and it has prompted people to actively participate in the spread of information on the network platform. The spread of policy information also shows that the public’s attention and understanding of green consumption positively regulates how they participate in information. The results of the Affect model show that the intercept coefficient of the model is not significant. Without external stimuli such as social platform network information, the public’s belief and attitude toward green consumption will not change actively. Among them, the main effects of time, CCI, and Certain are not statistically significant, and the Cogmech effect is significant, indicating that the public’s attention and understanding of green consumption information positively affects the public’s attitude toward green consumption. To sum up, changes in the policy and economic environment did not change the public’s attitude toward green consumption. The way the public actively participates in information does not affect the attitude toward green consumption. In other words, the public only spreads relevant information on social network platforms. They will not change the existing consumption attitude or put it into practical action. The pivotal aspect to change the public’s attitude toward green consumption by disseminating government policy information to the public is enhancing the public’s attention and understanding of the policy content. At present, the public’s attention and understanding of green consumption policies are relatively weak. In addition, the public exhibits a tendency not to actively understand or spread green consumption policy information, or change existing green consumption attitudes. To address this, the popularization and publicity of its policies need to be continuously strengthened.
Analysis of public green consumption attitudes under corporate topics
The relevant models for Cogmech, Certain, and Affect under corporate topics are shown in Table 12 below:
According to the statistical results of the Cogmech model, compared with the information released by government policies, the public pays more attention to the measures of enterprises to promote green consumption, and understands more clearly about the measures’ purpose (β00 = 0.035, P < 0.001). The main effects of gender, region, time, and PMG are not statistically significant; CCI significantly affects the public’s attention and understanding of enterprise-related measures. According to β40 = −0.104, it is evident that the public’s consumer confidence index negatively regulates the attention and understanding of enterprise measures. The model outcomes reveal notable insights, indicating that gender and region among individuals do not affect the way of participating in information; for every unit increase in the consumer confidence index, the enthusiasm of the public to participate in corporate topics decreases. At the same time, the public’s attention and understanding of information are positively correlated with the way of participating in information. The results of the Affect model show that the public’s attitude toward green consumption changed significantly in 2021. The increases in public consumption confidence, attention, and understanding of corporate initiatives have not promoted changes in public green consumption attitudes. The way of actively participating in corporate topic information is conducive to improving the public’s attitude toward green consumption.
Under corporate-related topics, the consumer confidence index negatively affects the public’s attention and understanding of green consumption, the way of participating in information, and the green consumption attitudes. When the public is optimistic about the economy, subjective feelings do not change beliefs and attitudes toward green consumption. In addition, public attention and understanding of corporate measures positively regulate how information is involved but negatively affect changes in public green consumption attitudes. The public pay corresponding green costs and changes their inherent consumption habits to put the green consumption measures mentioned by corporate into practice, so most people are only willing to spread information through social networks to express their support but unwilling to change their attitude toward green consumption. Under the topic of corporate promoting green consumption measures, the public’s self-green consumption attitude will improve with time, but the public will not take the initiative to understand the improvement of enterprises’ green technology or improve self-green consumption awareness.
Analysis of public green consumption attitudes under media topics
The relevant models for Cogmech, Certain, and Affect under media topics are available, as shown in Table 13 below:
The result derived from the Cogmech model shows that the public’s attention and understanding of media promotion information has increased significantly in 2019 and 2020. However, it is noteworthy that the reinforcement of consumer confidence has not promoted the public’s attention and understanding of the necessity of daily green consumption. The intercept term coefficient of Certain models is not significant, indicating that the public will not actively participate in the daily green consumption promoted by the media, which also shows that the media must popularize the knowledge of green consumption to the public. When the public has stronger attention and understanding of media publicity information, they are easier to participate in the spread of media information. The results of the Affect model show that the public’s attitude toward green consumption has changed significantly in 2019 and 2021. Surprisingly, despite heightened attention and understanding of green consumption, this increased awareness has not correlated positively with the observed changes in attitude; the public’s participation in forwarding information, supporting and other behaviors positively regulates attitude toward green consumption. Under media-related topics, the public improves attention and understanding, the way of to participate in information, and the attitude toward green consumption with the expanded popularization of green consumption knowledge by the media over time. The consumer confidence index negatively affects the public’s interest in green consumption, attention and understanding, and ways to participate in information. In addition, when the public pays more attention, understands more content of media publicity, and more actively participates in the discussion on online platforms, the information will spread more widely in social networks, but it does not significantly change the public’s attitude toward green consumption. The way the public participates in information in different regions is diverse. Men and women have obvious differences in the change of green consumption attitudes. In terms of the geographical attributes of commenters, the public in Central China, East China, and Northeast China actively participated in forwarding, supporting, and liking after the release of media information, strengthening the coverage of their information dissemination.
Discussion and conclusion
This article introduces a methodological framework, leveraging the ELM and text mining, to examine how information strategies from entities like governments and corporations shape public attitudes towards green consumption via social media data. CNN and LDA models are employed to categorize social media data’s subjects and topics, and LSTM models to analyze variations in public sentiment, focus content, and demographic attributes under different subject topics. The ELM serves as the theoretical foundation to extract public green consumption attitudes and drivers from online comments, in conjunction with the LIWC lexicon.
Our framework offers an effective alternative to traditional questionnaires and interviews by utilizing online user-generated text, developing sentiment dictionaries, extracting count data, and longitudinally testing the ELM’s theoretical framework with real-world data, enhancing the study’s innovation. Notably, the application of ELM to analyze information’s impact on attitude change in green consumption has been confined to experimental methods for theory testing and validation. The integration of this theory with big data for analysis and validation marks a novel and original aspect of this study.
This study finds that women and individuals in economically advanced regions are more attentive to green consumption, aligning with prior empirical studies (Huang et al., 2022). Public emotions indirectly sway green consumption attitudes. Huffaker (2010) empirically showed that information’s emotional dimension boosts attention, feedback, and participation. Liao and Huang (2020) confirmed that information content and emotional traits positively affect users’ behavioral tendencies when combined with ELM and emotional cognition theory. Promoting “Internet + ” is crucial for enhancing public attention. Through network push, information on government policy, corporate green technology, and social activities can be promptly disseminated to the public. Regular news events can subtly incorporate green narratives, gradually instilling a green consumption concept in the public. With the rapid advancement of information technology, online consumption is progressively replacing traditional shopping. E-commerce enterprises can be actively promoted to directly sell or collaborate with brick-and-mortar enterprises to offer green products and services. The celebrity effect can be utilized to enhance brand awareness or disseminate social responsibility, thereby fostering consumer trust in green consumption and promoting emotional consumption by the public.
This study reveals a positive correlation between public knowledge and attitudes towards green consumption under policy topics but a negative one under corporate and media topics. As per the Explanation Level Theory and Social Cognition Theory (Bandura, 2001; Trope and Liberman, 2010), individual behavior is shaped by the interplay of cognition, personal traits, and the external environment. The persuasiveness of information is enhanced when the explanation level induced by external information matches the individual’s explanation level. Thus, the public’s attention to and understanding of information serves as a positive regulator for their green consumption attitudes. This regulation, however, is contingent on the content released by different subjects (Yue, 2022). The study’s results indicate that knowledge influences attitudes, and a person who is more informed about green consumption is more likely to adopt green consumption behaviors (Michel et al., 2022). Furthermore, rectifying public cognitive bias towards green consumption aids in forming green consumption attitudes. However, positive cognition, emotions, and attitudes towards green consumption don’t necessarily translate to positive actions. Public green consumption attitudes require deeper contemplation combined with external information. Consequently, during publicity and promotion, it is imperative for enterprises and media to provide detailed information about the production process and green products, which is pivotal in cultivating consumer behavior.
In this research, a notable discovery emerged, indicating found that the economic environment negatively impacts public attitudes toward green consumption. This finding aligns seamlessly with existing literature. Ellen et al. (1991) suggested that stronger consumer confidence leads to greener product purchases and environmental activities. However, the market’s green products often lack utility, sufficient functional attributes, and effective information, deterring public participation in environmental protection. Lower perceived efficacy of ecological protection from green consumption behaviors makes it harder for individuals to feel their contribution, negatively reinforcing individual incentives and suppressing altruistic motives, leading to abandonment of green standards (Nguyen et al., 2019). The study also found that the policy environment had no significant effect, aligning with Habich-Sobiegalla et al. (2019). China’s green consumption is policy-led and top-down-driven. Despite annual introductions of relevant policies, a noteworthy absence prevails in the form of any specific green consumption law. There is no precise regulations on residents’ green consumption, with guidance primarily in an advisory level. This deficiency results in a failure to create a conducive social atmosphere for green consumption. The policy system should focus on providing society with more efficient, eco-friendly, and low-carbon green products and services. For instance, relying on big data platforms to establish a green consumption point system or personal carbon accounts; rewarding green consumption behaviors like garbage sorting and shared bicycle use with points; and designing a green consumption point system that integrates products, scenarios, applications, and other aspects.
This study’s methodological framework, based on ELM and text mining, comes with some limitations. Firstly, the study focuses on Chinese online users, necessitating future expansion to other countries for broader model applicability. Secondly, social media users may not represent all age groups, such limitations in data collection may result in incomplete datasets. To address this, future studies should strive for larger, more diverse samples. Lastly, this study identifies eight categories of factors influencing public attitudes using the LIWC Dictionary and Statistics’ unsupervised algorithm. However, this approach may overlook certain factors. Future research should expand factor consideration, refining the LIWC lexicon and corresponding algorithms to achieve more accurate factor quantification. This will deepen our understanding of consumer attitude shifts post-information receipt, offering valuable insights for businesses, governments, and the media.
Data availability
The author confirms that all data generated or analyzed during this study are included in this published article. Furthermore, secondary sources and data supporting the findings of this study were all publicly available at the time of submission. Additional data related to this study can be found in the Supplementary Information submitted with this article.
References
Ali M, Ullah S, Ahmad MS et al. (2022) Assessing the impact of green consumption behavior and green purchase intention among millennials toward sustainable environment. Environ Sci Pollut Res 23:1–13. https://doi.org/10.1007/s11356-022-23811-1
Amandeep D, Mohd S, Shalini T et al. (2021) Why do retail consumers buy green apparel? a knowledge-attitude-behaviour-context perspective. J Retail Consum Serv 59:1–11. https://doi.org/10.1016/j.jretconser.2020.102398
Bamberg S, Möser G (2007) Twenty years after Hines, Hungerford, and Tomera: a new meta-analysis of psycho-social determinants of pro-environmental behaviour. J Environ Psychol 27(1):14–25. https://doi.org/10.1016/j.jenvp.2006.12.002
Bandura A (2001) Social cognitive theory: an agentic perspective. Annu Rev Psychol 52:1–26. https://doi.org/10.1146/annurev.psych.52.1.1
Barbarossa C, Pastore A (2015) Why environmentally conscious consumers do not purchase green products: a cognitive mapping approach. QMR 18(2):188–209
Barbu A, Catană SA, Deselnicu DC et al. (2022) Factors influencing consumer behavior toward green products: a systematic literature review. Int J Environ Res Public Health 19(24):1–16. https://doi.org/10.3390/ijerph192416568
Cacioppo JT, Petty RE (1984) The elaboration likelihood model of persuasion. Adv Consum Res 19(4):123–205. https://doi.org/10.1016/S0065-2601(08)60214-2
Canio FD, Martinelli E, Endrighi E (2021) Enhancing consumers’ pro-environmental purchase intentions: the moderating role of environmental concern. Int J Retail Distrib 49(9):1312–1329. https://doi.org/10.1108/IJRDM-08-2020-0301
Cheng JB, Mao C, Huang Z et al. (2022) Implementation strategies for sustainable renewal at the neighborhood level with the goal of reducing carbon emission. Sustain Cities Soc 85(4):1–18. https://doi.org/10.1016/j.scs.2022.104047
Cheung MF, To WM (2019) An extended model of value-attitude-behavior to explain Chinese consumers’ green purchase behavior. J Retail Consum Serv 50:145–153. https://doi.org/10.1016/j.jretconser.2019.04.006
Chu W, Im M, Song MR et al. (2019) Psychological and behavioral factors affecting electric vehicle adoption and satisfaction: a comparative study of early adopters in China and Korea. Transp Res D Transp Environ 76:1–18. https://doi.org/10.1016/j.trd.2019.09.009
Danila V (2021) Environmental attitudes and contextual stimuli in emerging environmental culture: an empirical study from Russia. Sustain Prod Consumpt 27:2075–2089. https://doi.org/10.1016/j.spc.2021.05.008
Dhir A, Sadiq M, Talwar S et al. (2021) Why do retail consumers buy green apparel? a knowledge-attitude-behaviour-context perspective. J Retail Consum Serv 59:1–11. https://doi.org/10.1016/j.jretconser.2020.102398
Eid R, Agag G, Shehawy YM (2021) Understanding guests’ intention to visit green hotels. J Hosp Tour Res 45(3):494–528. https://doi.org/10.1177/1096348020947800
Ellen PS, Wiener JL, Cobb-Walgren C (1991) The role of perceived consumer effectiveness in motivating environmentally conscious behaviors. J Public Policy Mark 10(2):102–117
Filimonau V, Matute J, Czerwińska MK et al. (2020) The determinants of consumer engagement in restaurant food waste mitigation in Poland: an exploratory study. J Clean Prod 247(20):1–12. https://doi.org/10.1016/j.jclepro.2019.119105
Foroughi B, Nhan PV, Iranmanesh M et al. (2023) Determinants of intention to use autonomous vehicles: findings from PLS-SEM and ANFIS. J Retail Consum Serv 70:1–14. https://doi.org/10.1016/j.jretconser.2022.103158
Gabisch JA (2011) Virtual world brand experience and its impact on real world purchasing behavior. J Brand Manag 19(1):18–32. https://doi.org/10.1057/bm.2011.29
Habich-Sobiegalla S, Kostka G, Anzinger N (2019) Citizens’ electric vehicle purchase intentions in China: an analysis of micro-level and macro-level factors. Transp Policy 79:223–233
Haj-Salem N, Ishaq MI, Raza A (2022) How anticipated pride and guilt influence green consumption in the Middle East: the moderating role of environmental consciousness. J Retail Consum Serv 68:1–11. https://doi.org/10.1016/j.jretconser.2022.103062
Hao H, Guo J, Xin Z et al. (2021) Research on e-commerce distribution optimization of rice agricultural products based on consumer satisfaction. IEEE Access 9:304–315. https://doi.org/10.1109/ACCESS.2021.3114409
Hidalgo CJ, Coello PS, Reyes VH et al. (2022) Understanding citizens’ environmental concern and their pro-environmental behaviors and attitudes and their influence on energy use. Energy Rep. 8(3):103–109. https://doi.org/10.1016/j.egyr.2022.01.116
Hoffmann E, Mohaupt F (2020) Joint Storage: a mixed-method analysis of consumer perspectives on community energy storage in Germany. Energies 30(13):1–22. https://doi.org/10.3390/en13113025
Huang H, Long RY, Chen H et al. (2022) Exploring public attention about green consumption on Sina Weibo: using text mining and deep learning. Sustain Prod Consum 30:674–685. https://doi.org/10.1016/j.spc.2021.12.017
Huang YC, Backman KF, Backman SJ et al. (2016) Exploring the implications of virtual reality technology in tourism marketing: an integrated research framework. Int J Tour Res 18(2):116–128. https://doi.org/10.1002/jtr.2038
Huffaker D (2010) Dimension of leadership and social influence in online communities. Hum Commun Res 36(4):593–617
Joshi Y, Rahman Z (2019) Consumers’ sustainable purchase behaviour: modeling the impact of psychological factors. Ecol Econ 159:235–243. https://doi.org/10.1016/j.ecolecon.2019.01.025
Khan Y, Hameed I, Akram U (2022) What drives attitude, purchase intention and consumer buying behavior toward organic food? A self-determination theory and theory of planned behavior perspective. Br Food J 125(7):1–10. https://doi.org/10.1108/bfj-07-2022-0564
Kim MJ, Lee CK, Preis MW (2020) The impact of innovation and gratification on authentic experience, subjective well-being, and behavioral intention in tourism virtual reality: the moderating role of technology readiness. Telemat. Inform 49:1–10. https://doi.org/10.1016/j.tele.2020.101349
Kim N, Lee K (2023) Environmental consciousness, purchase intention, and actual purchase behavior of eco-friendly products: the moderating impact of situational context. Int J Environ Res Public Health 20(7):1–12. https://doi.org/10.3390/ijerph20075312
Kumar N, Mohan D (2021) Sustainable apparel purchase intention: collectivist cultural orientation and price sensitivity in extended TPB model. J Revenue Pricing Manag 20(9):149–161. https://doi.org/10.1057/s41272-021-00297-z
Laura S, Antonio AM, Martín N et al. (2020) Exploring preferences and sustainable attitudes of Airbnb green users in the review comments and ratings: a text mining approach. J Sustain Tour 10:1–19. https://doi.org/10.1080/09669582.2020.1838529
Lee H, Cheon HJ (2018) Exploring Korean consumers’ attitudes toward ethical consumption behavior in the light of affect and cognition. J Int Mark 30(2):98–114. https://doi.org/10.1080/08961530.2017.1376241
Lee JH, Haley E (2022) Green consumer segmentation: consumer motivations for purchasing pro-environmental products. Int J Advert 41(8):1477–1501. https://doi.org/10.1080/02650487.2022.2038431
Li JQ, Han Z, Ren H (2020) A quantitative evaluation model for the synergy of policy objectives of photovoltaic industry in China. J Tongji Univ (Nat Sci) 48(6):913–921
Li L, Liu X, Zhang X (2021) Public attention and sentiment of recycled water: evidence from social media text mining in China. J Clean Prod 30:1–10. https://doi.org/10.1016/j.jclepro.2021.126814
Li T, Chen Y (2019) Will virtual reality be a double-edged sword? Exploring the moderation effects of the expected enjoyment of a destination on travel intention. J Dest Mark Manag 12:15–26. https://doi.org/10.1016/j.jdmm.2019.02.00
Liao L, Huang T (2020) Research on the influence of source, information content and emotional characteristics on Weibo reposting. J Inf Technol 40(9):42–52
Liu H, Ma RL, He GY et al. (2023) The impact of blockchain technology on the online purchase behavior of green agricultural products. J Retail Consum Serv 74:1–10. https://doi.org/10.1016/j.jretconser.2023.103387
Luo YY, Yang Z, Liang Y et al. (2022) Exploring energy-saving refrigerators through online e-commerce reviews: an augmented mining model based on machine learning methods. Kybernetes 51(9):2768–2794. https://doi.org/10.1108/K-11-2020-0788
Margariti A (2021) White” space and organic claims on food packaging: communicating sustainability values and affecting young adults’ attitudes and purchase intentions. Sustainability 13(19):1–16. https://doi.org/10.3390/su131911101
Mezger A, Cabanelas P, Cabiddu F et al. (2020) What does it matter for trust of green consumers? an application to German electricity market. J Clean Prod 242:1–11. https://doi.org/10.1016/j.jclepro.2019.118484
Michel JF, Mombeuil C, Diunugala HP (2022) Antecedents of green consumption intention: a focus on generation Z consumers of a developing country. Environ Dev Sustain 8:1–22. https://doi.org/10.1007/s10668-022-02678-9
Mu Z, Zheng Y, Sun H (2021) Cooperative green technology innovation of an e-commerce sales channel in a two-stage supply chain. Sustainability 13:1–28. https://doi.org/10.3390/SU13137499
Nguyen HV, Nguyen CH, Hoang TT (2019) Green consumption: closing the intention-behavior gap. Sustain Dev 27(1):118–129
Ogiemwonyi O, Muhammad TJ (2023) The correlative influence of consumer ethical beliefs, environmental ethics, and moral obligation on green consumption behavior. Resour Conserv 19:1–12. https://doi.org/10.1016/j.rcradv.2023.200171
Oliver AG, Christina SR (2023) Influence of the New Ecological Paradigm (NEP) and environmental concerns on pro-environmental behavioral intention based on the Theory of Planned Behavior (TPB). J Clean Prod 382(1):1–13. https://doi.org/10.1016/j.jclepro.2022.134629
Raudenbush SW, Bryk AS (2002) Hierarchical linear models: applications and data analysis methods. Adv Quantum Technol 36(1):116–117. https://doi.org/10.1080/00401706.1994.10485413
Roh T, Seok J, Kim Y (2022) Unveiling ways to reach organic purchase: green perceived value, perceived knowledge, attitude, subjective norm, and trust. J Retail Consum Serv 67:1–13. https://doi.org/10.1016/j.jretconser.2022.102988
Rustam A, Wang Y, Zameer H (2020) Environmental awareness, firm sustainability exposure and green consumption behaviors. J Clean Prod 268:1–12. https://doi.org/10.1016/j.jclepro.2020.122016
Shehawy YM (2023) In green consumption, why consumers do not walk their talk: a cross cultural examination from Saudi Arabia and UK. J Retail Consum Serv 75:1–11. https://doi.org/10.1016/j.jretconser.2023.103499
Shi J, Jiang Z (2022) Willingness to pay a premium price for green products: does a reference group matter? Environ Dev Sustain 25(8):8699–8727. https://doi.org/10.1007/s10668-022-02419-y
Sovacool BK, Kester J, Noel L et al. (2019) Are electric vehicles masculinized? gender, identity, and environmental values in Nordic transport practices and vehicle-to-grid (V2G) preferences. Transp Res D Transp Environ 72:87–202. https://doi.org/10.1016/j.trd.2019.04.013
Teng S, Khong KW, Goh WW (2015) Persuasive communication: a study of major attitude-behavior theories in a social media context. J Internet Commer 14(1):42–64. https://doi.org/10.1080/15332861.2015.1006515
Testa F, Pretner G, Lovino R et al. (2021) Drivers to green consumption: a systematic review. Environ Dev Sustain 23(10):4826–4880
Toder EV, Fullerton S, De Beer LT et al. (2023) Social and personal factors influencing green customer citizenship behaviors: the role of subjective norm, internal values and attitudes. J Retail Consum Serv 71:1–10. https://doi.org/10.1016/j.jretconser.2022.103190
Trope Y, Liberman N (2010) Construal-level theory of psychological distance. Psychol Inq 117(2):440–463
Verweij M, Senior TJ, Domínguez D et al. (2015) Emotion, rationality, and decision-making: How to link affective and social neuroscience with social theory. FRONT NEUROSCI-SWITZ 9:1–3. https://doi.org/10.3389/fnins.2015.00332
Wiśniewska M, Czernyszewicz E (2023) Survey of young consumer’s attitudes using food sharing attitudes and behaviors model. Br Food J 125(1):242–261. https://doi.org/10.1108/BFJ-09-2021-1025
Wu Z, Zhang Y, Chen Q et al. (2021) Attitude of Chinese public towards municipal solid waste sorting policy: a text mining study. Sci Total Environ 756:1–10. https://doi.org/10.1016/j.scitotenv.2020.142674
Xia L, Wu X, Chai LL (2023) A review on the influencing factors of residents’ low-carbon consumption intention and behavior. J N China Electr Power Univ (Soc Sci) 2:33–42. https://doi.org/10.14092/j.cnki.cn11-3956/c.2023.02.005
Yan LI, Tat KH, Chen J (2020) Assimilating and differentiating: the curvilinear effect of social class on green consumption. J Consum Res 47(6):914–936. https://doi.org/10.1093/jcr/ucaa041
Yue X (2022) Research on the influence of comment information on consumers’ contradictory attitudes from the perspective of Hofflander’s attitude change theory. J Ind Eng Manag 36(4):36–49
Zaremohzzabieh Z, Ismail N, Ahrari S et al. (2021) The effects of consumer attitude on green purchase intention: a meta-analytic path analysis. J Bus Res 132:732–743. https://doi.org/10.1016/J.JBUSRES.2020.10.053
Acknowledgements
This research is supported by the Key Program of the National Social Science Fund of China (Grant No. 20AGL019), Zhejiang Provincial Natural Science Foundation of China (Grant No. LY22G010004), as well as Zhejiang Gongshang University “digital +” discipline construction key project (Grant No. SZJ2022B019).
Author information
Authors and Affiliations
Contributions
Conceptualization: Jun Fan, Lijuan Peng, and Tinggui Chen; methodology: Lijuan Peng, and Tinggui Chen; software: Lijuan Peng; validation: Lijuan Peng; formal analysis: Lijuan Peng; data curation: Lijuan Peng; writing—original draft: Lijuan Peng, Tinggui Chen, and Guodong Cong. All authors have read and agreed to the published version of the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Ethical approval
This article does not contain any studies with human participants performed by any of the authors.
Informed consent
This article does not contain any studies with human participants performed by any of the authors.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Fan, J., Peng, L., Chen, T. et al. Mining the impact of social media information on public green consumption attitudes: a framework based on ELM and text data mining. Humanit Soc Sci Commun 11, 184 (2024). https://doi.org/10.1057/s41599-024-02649-7
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1057/s41599-024-02649-7
