Abstract
With the advancement of the clean energy transition, residential photovoltaic (PV) power generation has become a key pathway for rural energy transformation and the promotion of rural revitalization. However, some members of the public have cognitive biases and a crisis of trust toward it. Much existing research on energy transition focuses on analyzing influencing factors and theoretical deliberation, with relatively limited in-depth investigation from the public perspective regarding their responses and attitudes. Given this gap, this paper intends to conduct sentiment analysis on online discussions concerning residential PV power generation in China. Objective: To reveal public sentiment orientations and key concerns, delve into underlying issues, and provide effective recommendations for promoting the development of residential PV systems. Method: Taking China as a case, the study conducted high-frequency keyword analysis, semantic network analysis, and sentiment analysis on 2371 online comments related to “Residential photovoltaic,” “Distributed photovoltaic,” or “Rural photovoltaic power generation.” Key Findings: The results of sentiment analysis showed that positive sentiment accounted for 65.71%, neutral sentiment for 15.08%, and negative sentiment for 19.20%. However, segmental statistics revealed that the combined proportion of high and moderate positive emotions was only 4.79%, while that of high and moderate negative emotions reached 5.63%. This shows public sentiment toward residential PV is a mixture of positive and negative feelings, with negative emotions dominating. These negative emotions focused on keywords such as “Contract,” “Being deceived,” “Radiation,” and “Roof.” Root Causes: The primary issues arise from four key areas: unilateral imbalance in rights and obligations during contract exchanges; exploitation of information asymmetry among the elderly; amplified costs due to limited or unknown technological literacy; and hidden cost deprivation related to rooftop rights. To address these concerns, this paper puts forward the following recommendations. First, standardize the contract exchange system. Second, launch targeted protection initiatives for the elderly population. Third, establish a multi-faceted, collaborative system for popularizing scientific knowledge about technology. Fourth, standardize rooftop resource exchange practices.
Keywords
Introduction
Against the backdrop of intensifying global warming trends and growing societal attention to climate issues, improving ecological environment quality has become a shared pursuit of all humanity. In this context, actively advancing the energy transition holds significant importance. Energy transition is recognized as a key strategy for addressing climate change and environmental degradation and is a critical mission for all nations (Doğan et al., 2025). Under the United Nations’ Sustainable Development Goal 7 (SDG 7), the objective is to ensure universal access to affordable, reliable, and sustainable modern energy. However, at the current pace, by 2023, approximately 600 million people will still lack access to electricity, and nearly 2 billion will continue to rely on polluting fuels and technologies for cooking (United Nations, 2022). Therefore, promoting the transformation of energy toward the goal of sustainable development still requires vigorous advancement. As one of the major energy powers, China has continuously forged a path of energy transformation suited to its national conditions under the guidance of the new energy security strategy of “Four Revolutions + One Cooperation.” Against the backdrop of the “dual carbon” goals and energy transition, solar PV power generation, particularly residential PV systems, is rapidly evolving from a supplementary source to large-scale centralized power plants into a vital component of China’s distributed energy system. This transformation stems from its advantages of widespread deployment, flexible installation, and proximity to end-users. Distributed PV systems generate electricity locally using nearby solar resources to minimize energy losses and alleviate supply shortages (Huang et al., 2025). In recent years, driven by multiple factors including national subsidy policies, county-wide pilot programs, and financial innovations, China’s residential PV market has experienced explosive growth. Cumulative installed capacity and the number of users have repeatedly reached new highs. According to the China Statistical Yearbook 2024, compiled by the National Bureau of Statistics of China (NBSC), as of the end of 2023, the installed capacity of solar power generation reached 610.48 million kilowatts (China Statistics Press, 2024), showing rapid growth compared to previous years (i.e. 2022 and 2021). China is advancing ambitious solar energy development initiatives, aiming to reach a capacity of over 2200–2800 gigawatts (GW) by 2030 (Lu et al., 2024). To further promote the development, construction, and utilization of distributed PV systems, China is actively advancing energy transformation in vast rural areas. In the 2025 notice issued by the National Energy Administration (NEA) on further implementing the “Thousands of Households Bathed in Light Initiative,” it explicitly stated the need to further advance the development and utilization of distributed renewable energy in rural areas. The notice encourages adopting models such as “company + village + farmer households” to utilize idle farmland and rural rooftops for constructing distributed wind and PV power generation. These systems should incorporate a certain proportion of energy storage, enabling self-generation and local consumption, with surplus electricity fed into the grid. Farmers can thereby secure stable rental income or electricity revenue (National Energy Administration, 2024). However, the industry’s rapid expansion has also brought with it a series of growing pains. Certain regions in China have encountered issues such as financial disputes over “solar loans,” inconsistent installation quality, lack of operation and maintenance services, and lower-than-expected user returns due to policy changes. These practical challenges not only impact the immediate interests of early adopters but have also sparked widespread discussion and a wait-and-see attitude among potential users. In the era of the internet and big data, various social media platforms have become crucial spaces for Chinese citizens to express opinions, share experiences, and shape public discourse. This vast volume of spontaneous online commentary data authentically captures users’ perceptions, attitudes, and emotional inclinations toward residential PV systems and their evolving trajectories. Compared to traditional surveys or interviews, this unstructured online commentary data offers a more dynamic and objective reflection of social sentiment and public opinion.
However, existing research primarily focuses on examining the influencing factors and theoretical deliberations of energy transition, such as using the system generalized method of moments estimation method to assess the impact of low-carbon energy transition on China’s common prosperity (Liu et al., 2023). Using panel data from 30 provinces in China spanning 2008–2019, this study examines the impact of energy transition on the urban-rural income gap (Gao et al., 2023). Collect relevant data from various countries to study the impact of geopolitical threats, geopolitical actions, and geopolitical risks on energy transition (Wang et al., 2024a). Examining Ghana’s National Energy Transition Framework through the Lens of Equitable Distribution in Decision-Making, Representation, and the Spatial and Temporal Framework of Energy Service Costs and Benefits (Sefa-Nyarko, 2024). Research methods focused on mediation analysis and case studies (e.g. using S&P 500 company data to test that capability-based energy transition strategies positively mediate the relationship between renewable energy industry performance and the financial performance of non-energy companies; Sirin and Yilmaz, 2024); or explore the role of community-level batteries in the energy transition through multiple business model cases and regulatory reviews in Australia’s National Electricity Market (Csereklyei et al., 2024).
Additionally, some studies discuss and compare a series of case studies from European countries to analyze key issues in the governance of energy transition (Rotondo et al., 2020). However, there has been limited in-depth research on public responses and attitudes toward residential PV systems, which excludes user needs and concerns from the energy transition process, which risks perpetuating energy injustice.
Therefore, this study aims to address this gap by posing the core research questions: What are the overall characteristics of sentiment toward residential PV systems in China’s current online discourse? How can the emergence of these sentiments be explained? What underlying issues in the industry’s development do these sentiment dynamics reveal?
This research contributes in the following ways: Theoretical contributions: 1. Expanding the application boundaries of sentiment analysis. As a vital branch of natural language processing (NLP), sentiment analysis provides new insights for social science research. Its application in energy public opinion analysis enables automated, large-scale sentiment computation across massive text volumes, overcoming limitations of traditional research methods such as small sample sizes, high subjectivity, and poor timeliness. 2. Broad data sourcing. Relevant commentary data is collected from multiple social platforms, avoiding reduced research persuasiveness due to reliance on a single source. 3. Promotion of interdisciplinary integration. This study spans multiple disciplines, including energy policy, computer science, sociology, and management, facilitating the convergence of diverse perspectives to address complex real-world challenges. Practical contributions: 1. Providing decision-making references for government departments. The research findings can clearly reveal public feedback on existing policies, helping policymakers evaluate policy effectiveness, optimize policy design, and achieve more precise and effective governance in industry management. 2. Offering market insights to PV enterprises and financial institutions. This helps companies identify gaps in their products and services, understand core user concerns, thereby optimizing marketing strategies, enhancing installation and O&M service quality, and innovating more attractive financial products—ultimately boosting corporate competitiveness and risk management capabilities. 3. Fostering rational public understanding and participation. By presenting a multifaceted view of public discourse, this research enables potential users to comprehensively and objectively understand the benefits and risks of residential PV systems. This reduces information asymmetry, guides more rational investment decisions, and promotes healthy, orderly market development.
The structure of this paper is arranged as follows: “Literature review” Section provides a review of relevant research on energy justice, distributed energy, and sentiment analysis methods. “Research design” Section introduces the research design of this paper. “Data analysis” Section conducts high-frequency word analysis, semantic network analysis, and sentiment analysis of online comments. “Dominant sentiment analysis of online comments based on social exchange theory” Section presents an analysis of dominant sentiments in online comments based on Social Exchange Theory. “Conclusion” Section concludes the study.
Literature review
Research on energy transition
Climate change is one of the major environmental challenges facing the world today. China has played an active and constructive role in addressing climate change, committing to peak carbon emissions by 2030 and achieving carbon neutrality by 2060, and a low-carbon energy transition is crucial for realizing these goals. Moreover, China’s global energy investments play a role in the global energy transition. By continuously investing in overseas renewable energy sectors, China not only helps achieve its own decarbonization goals through clean energy imports but also promotes global decarbonization by providing cleaner renewable energy to host countries, particularly low-income nations (Siyou et al., 2024). In previous energy transition theories, a relatively stable and predictable environment may have been assumed, but we are increasingly entering an era that urgently demands transformation, where unstable environments have become the norm. Thus, it is necessary to reconceptualize energy transition (Ping, 2024). The so-called energy transition primarily involves phasing out nuclear energy and promoting the development of renewable or sustainable energy sources. Based on current circumstances, China’s energy system should undergo a substantial low-carbon transformation (Zhou et al., 2021). In this regard, numerous scholars have examined the impacts of various factors on energy transition. From fossil fuels to renewable energy, energy efficiency and technological advancement are the most critical factors for an effective energy transition, with energy efficiency taking the lead by transforming the energy sector into green production and consumption in an effective manner (Hassan et al., 2022). Moreover, energy efficiency can be influenced by extreme climate events, which enhance energy efficiency by stimulating forced effects (Wang and Tang, 2023). Additionally, the generalized method of the moments estimation model reveals that the degree of resource endowment influences regional disparities in energy transition, with policy support significantly promoting the transformation (Su and Tan, 2023). The development of green finance significantly promotes energy transition (Lee et al., 2024) and exhibits notable spatial spillover effects, demonstrating the reciprocal characteristics of “local and neighboring regions” (Wan et al., 2023). Using the Stochastic Impacts by Regression on Population, Affluence, and Technology (STIRPAT) method and mediation effects, it has been demonstrated that Artificial Intelligence (AI) plays a positive role in promoting energy transition and carbon emissions reduction, with the potential to achieve carbon neutrality through AI (Wang et al., 2024b). However, some scholars argue that the higher a country’s level of urbanization, the lower its GDP tends to be. Factors such as growth and the use of foreign direct investment to drive development can increase emissions, hindering the energy transition (Nurcan, 2025). It is evident that research on energy transition predominantly focuses on exploring influencing factors, mostly employing quantitative methods. A decade-long study examining the impact of domestic conflicts on residential clean energy choices revealed that higher levels of conflict within families were associated with a lower likelihood of adopting clean energy solutions (Subedi et al., 2025). However, qualitative studies on energy transition remain relatively scarce.
Research on energy justice
Since the concept of energy justice was proposed, numerous scholars have conducted extensive research on it. Current studies on energy justice mainly focus on the following aspects: Firstly, elaborating on and developing the concept of energy justice within the disciplinary contexts of philosophy, ethics, and other related fields. The concept of energy justice refers to a conceptual framework aimed at identifying when and where injustices occur and how laws and policies can best respond to such injustices (Heffron and McCauley, 2014); alternatively, it can be described as a global energy system that equitably distributes the benefits and costs of energy services, thereby enabling representative and fair energy decision-making (Sovacool and Dworkin, 2015). In summary, the concept of energy justice serves more as an indicator to measure fairness and justice in aspects such as energy distribution and decision-making processes and is one of the bases for governments to adjust and implement energy-related policies. The key elements encompassed by energy justice initially included distributive justice, procedural justice, and recognition justice (Heffron et al., 2015). However, as energy justice has been increasingly applied in practice, discussions on energy justice and human rights-related issues have garnered significant attention, and restorative justice (McCauley and Heffron, 2018) and cosmopolitan justice (Healy et al., 2019) have been added to the key elements of energy justice. The former refers to any injustices caused by the energy sector that should be rectified, while the latter indicates that in the field of energy, everyone is a citizen of a unified world. Accordingly, it is necessary to consider the cross-border impacts brought about by various energy activities while also calling for countries to collaborate in addressing energy issues (Heffron, 2022).
Secondly, investigate the issue of distributive injustice caused by spatial injustice. A case study of a project in a certain region of India, dubbed the world’s largest solar park, reveals the more contentious land and power politics surrounding renewable energy, arguing that such projects exacerbate the instability of vulnerable groups (Yenneti et al., 2016). Additionally, some studies have applied a spatial perspective to examine energy poverty research, arguing that energy poverty is a form of injustice that requires broader interventions (Bouzarovski and Simcock, 2017).
Thirdly, the issue of energy transition. Field research in the Democratic Republic of Congo has revealed the existence of a “low-carbon divide” during energy transitions. For instance, sectors like transportation and electricity are significantly linked to toxic pollution and biodiversity loss, and it is argued that the sustainability criteria and analytical parameters used to evaluate energy or low-carbon transitions must be expanded (Sovacool et al., 2020) to safeguard the legitimate interests of vulnerable groups. However, other studies have found that Renewable Energy Communities possess the potential to engage citizens in energy transition and shape a just transition. Low-income and energy-poor households can benefit from affordable energy prices and energy efficiency measures when participating in renewable energy communities. Renewable energy communities are emerging as actors in a democratic, transformative, and equity-enhancing just transition (Hanke et al., 2021).
As mentioned above, existing studies have focused more on theoretical discussions of energy justice and individual aspects of energy justice issues while neglecting research on the systemic energy justice problems arising from actual energy transitions.
Research design
Data collection and processing
This paper primarily utilized Octopus software to legally collect online comments related to topics such as “residential photovoltaic,” “distributed photovoltaic,” or “rural photovoltaic power generation” from four types of Chinese social media platforms, namely Douyin, Kuaishou, Xiaohongshu, and Bilibili. The data collection period spanned from May 20 to May 23, 2025, with a total sample size of 3200 entries. The collected sample data underwent cleaning, including calculating the word count for each comment, plotting a box plot, and then selecting a word count threshold of ⩽Q1 (10 characters). Comments with 10 characters or fewer, such as “I want to install one,” “Must regret it,” and “What kind of news could it generate?” were deleted. After removing excessively brief comment texts, 2379 comment texts remained. For some useless or irrelevant comment texts, manual step-by-step verification and screening were conducted. Advertisement-type comments like “Self-investment? Contact us—self-investment won’t have these issues,” and ambiguous references like “Is it resolved? We’re working on it—what if he won’t dismantle it?” and “Folks, is this a problem?” “I’m struggling with this too—thanks for the heads-up,” totaling eight comments. After deletion, 2371 comment texts remained. The time span for data collection was from March 2023 to May 2025. The reason for selecting 2023 as the starting point was that in March 2023, China’s National Energy Administration, in collaboration with other departments, issued a notice on establishing pilot counties for the rural energy revolution. This notice explicitly incorporated residential PV systems into the core projects of rural revitalization, encouraging farmers to build power stations on their own rooftops. Additionally, it proposed a series of policy guidelines, including the implementation of the “Sunshine for Every Home Initiative.”
It is precisely this policy support that has generated significant momentum for residential solar installations and material sales. Platforms like Douyin and Kuaishou have seen increased views on residential solar topics compared to previous periods, with a notable rise in user-generated content such as installation case studies, yield calculations, and usage experiences. Collecting this data can meet the demands of big data analysis. Additionally, the effective comment samples were segmented, and stop words were removed using a custom list of special terms and the Harbin Institute of Technology Chinese stop word list to achieve the research objectives.
Methods
Sentiment analysis was employed as the research methodology. Sentiment analysis, also known as opinion mining, is the task of extracting and analyzing people’s opinions, emotions, attitudes, perceptions, etc., toward different entities such as topics, products, and services (Birjali et al., 2021). It belongs to the field of natural language processing and aims to interpret our underlying attitudes toward entities (Soleymani et al., 2017). Sentiment analysis is widely applied in qualitative research across various fields and scenarios, such as analyzing product-related reviews, public opinion analysis, and risk assessment; however, these sentiment analyses tend to be basic, typically extracting emotional information by inferring semantic relationships within sentences (Zhu et al., 2023). After years of in-depth research, scholars have gradually proposed multimodal sentiment analysis, which incorporates rich visual and auditory information to more accurately assist researchers in inferring the emotional intensity expressed in texts, speech, and other forms. This can be categorized into positive, neutral, and negative sentiments.
With the popularization of the internet, big data, and social media, the group evaluating a particular topic is no longer limited to a small segment of highly educated individuals but extends to a broad demographic encompassing all ages. This leads to richer commentary content and a significant increase in volume. Due to the virtual and non-contact nature of the internet, it also encourages the general public to express their genuine opinions more freely. Therefore, this study on people’s attitudes toward residential PV systems was based on online comment text mining. Utilizing the ROST Content Mining 6.0 (ROST CM 6.0) software as the data analysis tool to conduct high-frequency word analysis, semantic network analysis, and sentiment analysis on the collected texts. ROST CM 6.0 is a large-scale free social computing platform for assisting humanities and social sciences research, developed and coded by Professor Shen Yang from Wuhan University, China. This software enables a series of text analysis functions, including Weibo analysis, chat analysis, web-wide analysis, website analysis, browsing analysis, word segmentation, word frequency statistics, English word frequency statistics, traffic analysis, and cluster analysis. In short, it can essentially meet the requirements of this study.
Data analysis
Analysis of high-frequency words in online comments
A total of 2371 online review texts were imported into ROST CM 6.0. Based on the software’s default word segmentation table and the word segmentation criteria of this study, I manually removed words without specific representational meaning and merged synonyms. In addition, the minimum word count was set to two characters. Finally, 300 high-frequency words were obtained (see Table 1). Additionally, based on the obtained high-frequency words, a high-frequency word cloud diagram (see Figure 1) was generated using the Micro Word Cloud software. The principle of the word cloud is to intuitively present the core contradictions in the current online comments through the displayed font size. Therefore, in a word cloud, the larger the font size of a word, the greater the attention the word receives, and the words adjacent to it often imply logical associations.
Displaying high-frequency words.
Source: Author.
Cloud map of high-frequency words in online comments.
Through the high-frequency word statistics table and the generated word cloud, the intuitive perceptions of netizens can be roughly observed. The word frequencies of “Photovoltaic,” “Contract,” “Installation,” and “Roof” ranked among the top four. The term “PV company” also garnered significant attention, with a total frequency of 189 mentions. The term “Photovoltaic” highlights the research theme of this paper, while “Contract,” “Installation,” and “PV company” focus more on the market transactions and service processes related to the PV industry. As the physical carrier required for residential PV systems, “Roof” was associated with issues such as spatial property rights and asset stability, while also implying innovations in subsequent models like roof leasing. “Loan,” “Cost recovery,” and “Rent” ranked sixth, tenth, and thirteenth, respectively, reflecting netizens’ focus on the financing and returns brought by residential PV systems. At the same time, netizens also expressed concerns about issues arising from the investment, construction, and usage processes of PV. Based on the above, a cost-return logic framework can be established. The terms “PV panels,” “Power generation,” and “Grid connection” reflected netizens’ attention to the technology of PV construction and the smoothness of its operation. The terms “Country,” “Radiation,” and “Problems” can be comprehensively summarized as the impacts brought about by policies and societal perceptions. “Country” reflected the “guiding force” of policies concerning investment, construction, and operation in the PV industry. The frequency of “Radiation,” appearing 66 times, directly indicated public concern over the potential hazards of PV, while also revealing a certain degree of social cognitive bias. This reflected a governance gap resulting from inadequate public science communication efforts. The term “Problems,” as a high-frequency negative word, encapsulated netizens’ complaints regarding various aspects of residential PV systems and serves as the pivotal point of this sentiment analysis.
Semantic network analysis of online comments
Semantic network analysis enables scattered comments to reveal the interconnected network of related topics while also delineating the thematic boundaries of sentiment analysis for this study. Therefore, this paper conducted a semantic network analysis of online comments using ROST CM 6.0 software, combining the calculation of degree centrality to generate Figure 2. From the figure, it can be visually observed that the terms “Photovoltaic,” “Loan,” and “Installation” had the greatest number of connections, indicating their higher degree centrality. The entire network diagram radiated outward from these three terms, demonstrating that they were the most critical factors attracting public attention to residential PV systems. These terms serve as the “anchors” of online discussions, each carrying controversies from different dimensions.
Semantic network analysis diagram of online comments.
“Photovoltaic,” the central theme of this paper, occupied a super-core position in the semantic network analysis, connecting with “Country-Grid connection-Power grid” and “Installation—Return on Investment—Investment—One Year—House,” which covers the chain of demands from PV policy support to livelihood benefits. The term “loan” can be regarded as the core risk factor, primarily linking to elements such as “solar panels—ID card—property—company—rent—equipment” within the network. This reflected situations where, due to insufficient existing financial resources to cover initial investments, users took out loans for solar panel installation and maintenance. Concurrently, there existed a trust crisis concerning companies involved in the solar industry. The term “installation” served as the core service within the network, linking elements such as “farmers—income—profits—subsidies—rent” and “photovoltaic companies—contracts—relocation.” This reflected netizens’ significant interest in monetization and value appreciation following PV installations, while also highlighting post-installation conflicts in PV construction projects. It was important to note that each node mentioned above can become a secondary center or subtopic, reflecting the interconnected and mutually influential nature of netizens’ concerns and demonstrating the complex risk ecosystem of residential PV systems.
Sentiment analysis of online comments
Analyzing text data with emotional tendencies in user comments on social media is a valuable approach for gaining an in-depth understanding of the public’s views on a specific topic. Analysis of online review texts using ROST CM 6.0 revealed that positive sentiment accounted for 65.71%, neutral sentiment for 15.08%, and negative sentiment for 19.20% (see Figure 3). From the segmented statistical results, high and moderate positive sentiments collectively accounted for 4.79%, while high and moderate negative sentiments together constituted 5.63% (see Figure 4). Therefore, overall, netizens’ attitudes toward residential PV systems contained some positive elements but were predominantly negative.
Statistical chart of emotional distribution in online comments.
Statistical chart of sentiment segmentation in online comments.
Through comparative analysis of negative sentiment values in the comment texts, it was found that negative emotions primarily revolve around issues such as “Contract,” “Being deceived,” “Radiation,” and “Roof.” Regarding “Contracts,” 14.3% of comments expressed concerns about discrepancies between paper and electronic versions, as well as contract coercion issues. Examples include: “They want to bind you with a contract for at least 20 years. If you breach it, you must pay full compensation . . .” and “The elderly in our family were also tricked into signing a contract, and after we found out, it turned out that the paper version and the electronic version had completely different contents.” Among the statements describing “Being deceived,” the group under discussion primarily consists of elderly individuals in their households, accounting for 25%, exemplified by phrases such as “Someone tried to install it for my mother-in-law too, but both her son and I objected, fearing they might be scammed.” In the “Radiation” discussion, 30% of comments expressed concerns that installing solar panels at home could adversely affect health and pollute the living environment. Examples include “The random disposal of waste battery panels severely pollutes the environment” and “Photovoltaic solar panels generate electromagnetic radiation during operation, and excessive radiation may harm human health.” However, some netizens have offered counter-explanations, such as “No significant impact; it is less than the radiation level of mobile phones.” In the “Roof” statements, 13.2% of comments expressed concerns about the aesthetics of roofs and difficulties in enjoying natural light, for example: “The house already had poor lighting, and now even the roof doesn’t get any sunlight.” and “The house already had poor natural lighting, and now even the roof cannot get sunlight.” However, among the comments about “roofs,” there were also positive sentiments like “Actually, when my family installed it, it was to fix the roof. The house was old, and the roof leaked everywhere, but since installing the solar panels, the south-facing side with the panels hasn’t leaked at all.”
In summary, due to differences in netizens’ educational backgrounds, cognitive levels, and regional disparities, the resulting sentiment toward residential PV evaluations was multifaceted and multidimensional. Consequently, it is necessary to comprehensively adjust the construction and marketization of residential PV based on actual conditions.
Dominant sentiment analysis of online comments based on social exchange theory
Social Exchange Theory (SET) stands as one of the most influential conceptual paradigms in explaining interorganizational behavior, defining social behavior as the exchange of activities between individuals that may yield tangible or intangible resources (Ren et al., 2025). Moreover, individuals assess the specific allocation ratio of costs and benefits, as well as investment and profits during exchanges, and seek to ensure that their profits are proportional to their investments, primarily viewing issues from an economic perspective. One of the core principles of SET is reciprocity, which expects individuals to return favors or benefits in kind, thereby fostering trust and cooperation. However, reciprocity is not always symmetrical, and the balance between costs and rewards determines satisfaction and the longevity of relationships (Borige Vijay and Shailaja, 2024). The essence of human social behavior lies in pursuing mutually beneficial and equitable exchange relationships. When individuals perceive an exchange as fair and anticipate returns exceeding their inputs, positive emotions arise; conversely, when perceived unfairness causes costs to outweigh expected gains, negative emotions become the dominant attitude. Based on the analysis results from the previous section—namely, that negative emotions outweigh positive emotions—the specific reasons are analyzed as follows.
Theoretical explanations for the dominance of negative emotions
There exists a unilateral imbalance in rights and obligations during the exchange of contracts. The high-frequency term “Contract” ranked second with 438 occurrences. In negative reviews, typical cases mentioning “dual contracts,” “20-year binding clauses,” and “monopoly over repair pricing” included instances such as “demanding verification codes under the guise of transferring money to secretly sign electronic contracts” and “requiring full compensation for breach of contract, with repair prices determined solely by the company.” In semantic network analysis, “contract” formed strong associations with “scammer,” “electronic contract,” and “photovoltaic company,” directly linking “Contract” to the negative perception of a “fraud tool.” According to the “fairness principle” of social exchange theory, when users participate in residential PV systems, the expected exchange in the bilateral contract is mutually beneficial. Users provide rooftop usage rights and cooperate with installation, while enterprises offer transparent rent or subsidies, standardized operation and maintenance, and clearly defined rights and responsibilities. However, data analysis revealed that the actual exchange outcomes exhibited a certain degree of unilateral bias. On one hand, enterprises evade obligations through dual contracts—where commitments in paper contracts significantly diverge from those in electronic contracts—which may force users to bear unexpected costs such as high breach-of-contract penalties and additional repair fees. On the other hand, users’ core rights are being undermined. For instance, crucial contract terms—such as power generation calculations, operational and maintenance standards, and revenue payout schedules—are typically determined unilaterally by the company, leaving users with no room for negotiation. This unfair exchange, where businesses benefit while users suffer, directly undermines the principle of mutual benefit and fairness, leading to negative sentiment becoming the dominant attitude in contract-related reviews.
Exploitation of Information Vulnerabilities Among the Elderly Population. In negative comments, 25% of “Being deceived” scenarios targeted the elderly, with typical expressions like, “What if my elderly parents were tricked into installing panels for 30 years?” and “Someone installed panels at my mother-in-law’s place—I’m worried they got scammed.” In the semantic network, “Farmers” are associated with “scammers” and “signing contracts.” Considering that rural elderly individuals are the primary group with weak PV-related knowledge among farmers, this confirms that the elderly are the core target of exploitative transactions. The rational actor assumption in social exchange theory presupposes information symmetry. However, elderly groups in residential PV exchanges find themselves at an informational disadvantage due to factors such as low digital literacy and limited access to information channels. On one hand, companies exploit information asymmetries to establish exploitative exchanges—luring elderly individuals into contracts through personal connections and false promises of high returns, while concealing core risks such as 25-year lock-in periods, hidden loans, and difficulties in property demolition. On the other hand, the transaction costs borne by the elderly are significantly underestimated: they not only relinquish the right to use their rooftops, but they may also incur additional debt due to dual contracts. Meanwhile, their anticipated returns often remain unfulfilled due to corporate defaults. This one-sided exploitation of the disadvantaged by those with informational advantages generates intense negative emotions, such as anger and anxiety, among the elderly population and their family members, making them a significant source of negative reviews.
The unknown or underestimated costs of technological cognition amplification. The term “radiation” appeared 66 times among high-frequency words. Negative comments such as “solar panels cause cancer” and “solar panels pollute the environment” accounted for 30% of the discourse. Even when netizens countered that “radiation is lower than that from mobile phones,” it failed to reverse the prevailing negative perception. Negative sentiments surrounding “radiation” reflect certain societal cognitive biases and reveal a governance gap stemming from insufficient science communication efforts. The core of social exchange theory lies in risk-benefit tradeoffs. When participating in PV exchanges, users factor health risks into the cost of the exchange. However, due to gaps in public science education, users remain unaware or minimally informed about potential radiation-related risks. On one hand, governments and enterprises have failed to provide risk mitigation mechanisms—neither presenting authoritative PV radiation testing reports nor offering insurance safeguards, such as compensation clauses for radiation-induced health issues, to reduce users’ perceived risk. On the other hand, users’ risk costs are amplified due to a lack of knowledge or limited awareness. According to the loss aversion tendency in social exchange theory, users perceive health losses as far more significant than economic gains. Even though claims about “radiation causing cancer” lack a scientific basis, users still perceive them as high-probability costs. This leads to perceived risk costs outweighing anticipated power generation benefits, thereby generating negative sentiment.
Implicit Cost: Deprivation of Roof Rights. The high-frequency term “Roof” appeared 245 times, ranking fourth. Negative comments focused on aesthetic damage, such as red roofs being painted jet black, and obstructed natural light (e.g. homes with poor lighting where roofs cannot receive sunlight). Only a few positive comments related to “repairing roofs to prevent leaks.” Within the semantic network, “Roof” exhibited strong associations with “installation” and “photovoltaic companies,” indicating that roofs served as the core spatial resource for PV exchanges. However, users’ expectations for compensation when relinquishing roof rights remained unmet. Roofs represent users’ private spatial resources; their participation in PV exchanges essentially involves transferring roof usage rights in exchange for economic or functional benefits. However, data analysis revealed that this exchange involved hidden cost deprivation. On one hand, businesses compensate only for visible benefits while overlooking hidden costs—namely, the aesthetic appeal and natural lighting of rooftops. On the other hand, users consider the aesthetic and lighting value of rooftops nearly as important as the required income, yet businesses fail to provide additional compensation for the resulting loss of visual appeal and natural light. After installing rooftop solar panels, users effectively cede permanent rights to part of their rooftop space without receiving long-term compensation for these rights. This creates an imbalanced exchange where the transfer of resources outweighs the benefits gained, thereby exacerbating negative sentiment.
The essence of negative emotions taking hold is the failure of the exchange system
Based on the empirical data and SET-based analysis presented in this paper, the prevalence of negative sentiment in residential PV network discussions is not the result of user irrationality but stems from the triple failure of market-technology-resource exchanges. First, market exchange mechanisms have failed. Contract fraud and scams targeting the elderly reveal a lack of transparent and fair exchange constraints within the market, such as contract filing reviews and oversight of PV companies’ qualifications. Second, the technology exchange system has failed. Misconceptions about radiation and gaps in science literacy reveal a lack of information symmetry mechanisms regarding technological risks, such as insufficient authoritative science communication and limited availability of risk hedging tools. Finally, the resource exchange system has failed. The loss of rooftop rights reflects the absence of a compensation mechanism for spatial resources, such as accounting for the implicit costs of rooftops and safeguarding reversibility. These systemic failures collectively lead users to persistently perceive unfairness in PV exchanges, where costs outweigh benefits. This ultimately allows negative sentiment to dominate public discourse. Moving forward, a fair, transparent, and predictable exchange system should be established to restore the balance between costs and benefits.
Conclusion
This study conducted high-frequency keyword analysis, semantic network analysis, and sentiment analysis on 2371 online comments related to “Residential photovoltaic,” “Distributed photovoltaic,” or “Rural photovoltaic power generation.” Sentiment analysis revealed that public sentiment toward residential PV systems was predominantly negative, primarily focusing on keywords such as “Contract,” “Being deceived,” “Radiation,” and “Roof.” Through an analysis of causes using social exchange theory, the root causes of the problem were identified as stemming primarily from four aspects: unilateral imbalance of rights and obligations in contractual exchanges, exploitation of the elderly’s information disadvantage, amplified costs due to unknown or limited technological knowledge, and hidden cost deprivation of rooftop rights. To address this, this paper puts forward the following recommendations. First, standardize the contract exchange system. Second, launch targeted protection initiatives for the elderly population. Third, establish a multi-faceted, collaborative system for popularizing scientific knowledge about technology. Fourth, standardize rooftop resource exchange practices. Specific details are as follows:
Standardize the contract exchange system. In response to issues such as the proliferation of dual contracts, the lack of safeguards for electronic contract signing, and the deprivation of users’ core rights, local government energy authorities must improve property rights systems and benefit distribution mechanisms while strengthening market oversight and contract governance. First, expedite the implementation of standardized provincial contracts to clarify essential legal relationships such as rooftop leasing and ownership of PV facilities, thereby safeguarding the legitimate rights and interests of farmers. Second, establish a third-party electronic signing platform in collaboration with government authorities. After executing a paper contract, parties may use this platform to sign an electronic contract, which incorporates a dual-verification step: facial recognition combined with mandatory aloud reading of the contract terms to confirm the signing. The above measures can significantly reduce profit disputes arising from dual contracts and protect users’ legitimate rights and interests. Third, establish a contract filing and accountability mechanism to increase the cost of non-compliance. After successfully signing contracts with users, enterprises must upload them to the municipal-level energy regulatory platform for filing within a specified timeframe. The platform uses AI to identify abnormal clauses deviating from standard templates, automatically triggering manual verification. Enterprises using non-standard contracts will receive a warning and be required to rectify violations upon the first offense. Two cumulative violations will result in suspension of grid connection privileges, while three violations will lead to inclusion on the industry blacklist. Under asymmetric information, suppliers must bear a certain degree of risk, thereby reducing the buyer’s risk (Lian and Dong, 2024). These measures will proactively regulate corporate behavior, encouraging enterprises to voluntarily comply with industry regulations.
Launch targeted protection initiatives for the elderly population. To address the issue of elderly individuals being vulnerable to fraud, it is necessary to collaborate with grassroots governments, village committees, third-party notary agencies, and other relevant entities to jointly resolve this problem. On one hand, leverage the role of grassroots governments and village committees in promoting safety awareness and fraud prevention. For rural seniors, governments and village committees should organize specialized activities—such as utilizing visually engaging formats like comic-themed mural walls for outreach—and convene respected elders within the community to conduct ideological guidance and safety training. These individuals can then be entrusted to assist in outreach efforts, thereby enhancing the effectiveness of awareness campaigns. On the other hand, introduce third-party notary services where feasible and enhance contract guidance for vulnerable groups such as the elderly in rural areas. The above measures will help improve elderly users’ awareness of residential PV systems and enhance their vigilance.
Establish a multi-faceted, collaborative system for popularizing scientific knowledge about technology. Addressing public concerns about radiation from residential PV systems hinges on the lack of authoritative science communication. Therefore, local science and technology bureaus, agriculture and rural affairs bureaus, research institutions, and media centers must collaborate to provide scientifically grounded answers to public queries regarding residential PV systems. First, local government departments should collaborate with research institutions, universities, and industry associations to produce accessible science communication content, such as short videos and illustrated guides, to vigorously debunk myths about “radiation hazards” and “battery pollution.” Second, utilize targeted dissemination through both online and offline channels such as rural broadcasts, bulletin boards, and media centers. Real-time media response tools can optimize public outreach and education, enhance public awareness, and strengthen public mobilization (Radjab Achmad and Achmad Ezra, 2025). Third, PV knowledge can be incorporated into rural community education curricula, with regular outreach activities to promote PV science among households. These measures will help enhance public understanding of the technology’s principles and environmental benefits.
Standardize rooftop resource exchange practices. Regarding the failure of enterprises to fully account for hidden costs, such as the aesthetic damage to farmers’ homes and reduced natural lighting, the following measures are proposed. First, relevant government departments must establish and standardize accounting standards for compensating hidden costs. These standards should clarify that compensation should primarily come from PV enterprises with appropriate subsidies from county-level finances. This will address losses such as aesthetic damage to rooftops and reduced natural light caused by PV installations. Second, establish a dedicated task force comprising village committees, grassroots governments, judicial bodies, and industry associations to provide free mediation services for resolving PV-related disputes. Third, the Market Regulation Bureau should establish a standardized traceability system covering the entire lifecycle of PV equipment. It requires enterprises to publicly disclose the entities responsible for maintenance and recycling of such equipment. It demands the imposition of administrative penalties on violations involving the improper disposal of waste equipment. It mandates that offenders bear the costs of environmental remediation. These measures collectively help mitigate economic losses and disputes arising from hidden costs associated with residential PV installations.
Overall, as the world’s largest PV market and consumer, China’s explosive growth in residential solar installations and the accompanying shifts in public discourse have created an invaluable “natural laboratory.” A review of the literature reveals that most global renewable energy policies are built upon a “technology-economic” paradigm, which assumes that public acceptance will naturally follow through improvements in technological efficiency and the provision of economic subsidies. However, China’s current case clearly demonstrates that one of the greatest bottlenecks for large-scale application may not be technological immaturity but a “social trust deficit.” As deployment shifts from large-scale power plants to distributed systems—residential installations that become deeply intertwined with people’s living spaces—the focus of governance must simultaneously transition from “hardware” investment to “software” development. This entails building and maintaining a fair, transparent market environment alongside a credible regulatory framework. China’s experience serves as a global cautionary tale: neglecting effective oversight of PV market practices, contract standards, and financial products can rapidly erode public trust through fraud and unfair contracts. This not only triggers intense social resistance and delays deployment but may also precipitate financial risks and social conflicts. For regions like Southeast Asia and Latin America planning large-scale distributed energy rollouts, this represents a critical forward-looking warning.
The main innovations of this study are as follows. In terms of theoretical contributions, first, we innovatively introduced sentiment analysis methods into the field of clean energy public discourse research, expanding their application scenarios in assessing the social impacts of energy transitions. Second, we promoted the integration and convergence of multidisciplinary perspectives in addressing complex real-world challenges. In terms of practical applications, first, by deeply analyzing public sentiment and core concerns, this research establishes an empirical foundation. This foundation supports governments in refining policy design, businesses in enhancing service quality, and society in addressing cognitive biases. In turn, it promotes the standardized development within the residential solar industry and builds public trust. Second, it addresses societal concerns, enabling potential users to form a comprehensive, objective understanding of the benefits and risks of residential solar systems. This reduces information asymmetry and guides them toward making more rational investment decisions.
It must be acknowledged that this paper still has some limitations. Firstly, due to the difficulty in data collection, the final sample size is smaller than expected. The limitation makes it impossible to employ more advanced methods such as deep learning for research. Second, there is a lack of discussion on typical case studies. Third, when analyzing the primary causes of negative emotions, existing data cannot verify whether such sentiments are influenced by selective reporting or exaggerated risk perceptions. However, from an evolutionary psychology perspective, perceptions of unfairness or deception—which are linked to survival threats—can trigger more intense emotional responses. Therefore, individuals who have suffered losses due to PV installations or revenue issues are more motivated and emotionally driven to speak out and issue warnings. Their actions aim to seek justice or prevent others from repeating the same mistakes. In contrast, those who have benefited significantly from PV installations tend to be calm and less aroused, with a relatively lower desire to share their experiences. This dynamic results in negative voices gaining greater visibility in the public discourse. Future research could expand the data sample size and employ deep learning models to further refine sentiment analysis. In this way, it can provide more in-depth theoretical support for the refinement and inclusivity of energy governance. Additionally, field visits could be conducted to select one or two representative cases for in-depth analysis. This analysis helps explore the underlying challenges in the development of distributed PV systems, thus providing specific actionable recommendations to support local energy transitions.
Footnotes
Acknowledgements
I thank Dr. Yang Rui for her assistance in providing ideas on data analysis methods, and also thank the reviewers for their constructive suggestions.
Ethical considerations
This article does not contain any studies with human or animal participants.
Author contributions
Min Wei’s contributions are collecting relevant data, analyzing data, and writing the article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data availability statement
Data will be made available on request, except for the public data already indicated in the text.