The threat of climate change and the depletion of non-renewable resources, particularly fossil fuels, has become a driving force for communities worldwide to seek solutions to this pressing challenge. The European Union has not remained passive in this regard. In 2014, the European Commission introduced A Policy Framework for Climate and Energy in the Period from 2020 to 2030 (European Commission, 2014a), which outlined key guidelines for the future development of energy policy. According to this document, the primary objectives for EU member states include ensuring the security of energy supply, enhancing competitiveness, and guaranteeing a high level of sustainability in the energy sector, all with the overarching aim of establishing a low-emission economy. Among the specific targets set for the EU, the framework includes a 40% reduction in greenhouse gas emissions by the year 2030 (European Parliament, 2009). Subsequently, the EU released the European Green Deal, a strategic roadmap for future economic development, reaffirming Europe’s ambition to become the first climate-neutral continent (European Commission, 2014b).

The European Energy Policy 2030 described in the document has served as a foundation for formulating the Polish Energy Policy 2050 (PEP 2050) (Wierzbowski et al., 2017), and more recently, Poland’s Energy Policy until 2040 (Ministerstwo Klimatu i Środowiska, 2021). As Poland is traditionally dependent on coal, the main concept of PEP 2050 is the decarbonization of the country’s energy supply, leading to a reduction in fossil-fuel-related emissions. The specific aims include increasing the energy efficiency (including reduction of grid losses), continuation of improving the renewable energy source development (along with energy storage technologies), and further deregulation of energy sector to enable effective resell of electricity to the public grid by the prosumers owning micro-installations and construction of a nuclear power plant of minimum 6000 MW (Wojtaszek et al., 2024; Paska et al., 2020). Moreover, the factor that may, despite its negative character, contribute to increasing the speed of the abovementioned changes is the global energy crisis related to the pandemic, which has intensified since the Russian invasion of Ukraine in 2022. The rising prices of fuels proved that green energy might be the long-term solution for whole societies (Gajdzik et al., 2024a).

The transition away from coal as a main energy source necessitates a swift shift in the Polish energy sector. As of 2024, coal’s contribution to the country’s electricity generation has decreased to approximately 62.4%, while the proportion of renewable energy has increased to approximately 30.5% (Fraunhofer Institute for Solar Energy Systems ISE, 2024). The most popular sources of green energy in Poland are the photovoltaic (PV) installations, both in the form of major industrial plants as well as individual household systems. In 2024, the installed power of photovoltaics in public power plants exceeded the record level of 18 GW, achieving a value of 18,826 MW in July (Agencja Rynku Energii, 2024a). At the same time, the installed power of prosumer micro-installations reached over 11,000 MW (with the number of individual prosumers nearing 1.44 million). It is also worth mentioning that Polish communities started engaging in energy cooperatives, local initiatives based on joint investments of communities in renewable energy sources (Gajdzik et al., 2024b); however, this form is still relatively new and in a development stage in the country.

Wind turbines are another green technology that plays a role in Polish electricity production. Onshore wind power plants provided 14.4% on average of public net electricity generation in 2024 (with 16% in July) (European Network of Transmission System Operators for Electricity, 2024), and their installed power equals almost 9500 MW (Zdonek et al., 2023). Currently, there are no offshore plants; however, there are projects underway on the Baltic Sea, assuming the creation of wind turbine farms of 8.4 GW total capacity in the first stages (Palmowski and Kwiatkowska, 2023). The challenge in the development of wind technology in Poland is connected to the natural conditions, which vary substantially across the country’s regions. According to average wind speeds, the best regions for wind turbine placement are Pomerania and Central Poland, with the rest of the country presenting mediocre possibilities in this aspect (Malska and Mazur, 2017). Due to the climate issues and lack of state support, prosumer wind installations are extremely rare in Poland. While the installation of small turbines for strictly private use has been largely deregulated for years, connecting to the grid posed a significant difficulty for individuals. In July 2024, there were only 66 prosumer wind installations of 0.41 MW installed power (Agencja Rynku Energii, 2024b). Currently, the Polish government has begun to encourage individuals to apply wind turbines in their household systems, launching the “My Wind Power Plant” state subsidy program (Augustowski and Kułyk, 2024).

Even though the idea of green energy is, in the minds of many, equivalent to the photovoltaic technology, the generation of electricity is not the only area that must be taken into consideration. Another unit that can be more frequently observed in both private households and public buildings is the heat pump, which is the main factor in decarbonizing the heat supply of buildings (Lämmle et al., 2023). Heat pumps operate by a simple mechanism of extracting heat from the environment (e.g., air or ground) and delivering it to a target space (e.g., central heating system in houses) (Jelić et al., 2023). At the same time, it cannot be forgotten that the pumps need electricity to operate, especially in unfavorable atmospheric conditions. Even though all of them are based on the same idea of heat exchange, there are multiple variants of heat pumps depending on the heat source, heat transfer material, and the heat receiver, such as simple air-air pumps, often used air-water pumps, or more advanced and more costly ground-coupled brine-water pumps. Obviously, those variants differ in terms of efficiency and reductions of CO₂ emissions. The best results are obtained by combining the pump with photovoltaic panels. For example, a system containing an air-source heat pump and PV can decrease the emissions by 52%, while applying PV in combination with a geothermal ground pump equipped with a 110-m borehole heat exchanger can lower emissions by 63% (Chhugani et al., 2023).

Currently, in Poland, the heat pump market has experienced a temporary fluctuation and a decline in sales, which can be attributed to the broader crisis in the construction sector. According to the Polish Organization for Heat Pump Technology Development (PORT PC), a leading Polish organization dedicated to research and the promotion of heat pump technology, 2023 marked the first year since the organization’s founding in 2012 that a decline in heat pump sales was recorded–a drop of nearly 39% compared to 2022 (Polska Organizacja Rozwoju Technologii Pomp Ciepła PORT PC, 2024). Interestingly, the only category of heat pumps that did not experience a decline in sales was ground-coupled heat pumps, which saw a 12% increase compared to 2022, despite typically being the most expensive option and requiring the most complex installation procedures. Accessing the ground as a heat source involves either the shallow excavation of a large surface area or the drilling of deep boreholes, making the process more resource-intensive than other heat pump systems.

A crucial aspect of Polish decarbonization and the switch to green energy is the implementation of national financial incentives for those willing to install a renewable energy system in their households. Currently, among the most popular are the My Electricity program, thermal modernization tax relief, and the Clean Air Priority Program. The My Electricity program was launched in 2019 and aimed at increasing the number of prosumer PV micro-installations in the country by more than 200,000 (Olczak et al., 2021). It was based on providing partial reimbursement of costs incurred in the investment in the installation. Thermal modernization tax relief was implemented into the fiscal system in 2019 and allowed the PIT payers to deduct the costs of modernization investment (including installation of green energy and heating systems) from their incomes (Bednarz and Babiarz, 2023; Łożykowski et al., 2023). The Clean Air Priority Program (CAPP), launched in 2018, is also aimed at individual households and offers non-refundable grants to be spent on partial repayment of a bank loan taken in order to purchase and install a photovoltaic micro-installation (Mazurek-Czarnecka et al., 2022; Matczak et al., 2023). Such incentives effectively encourage many household owners to use renewable energy resources. In addition, scientists point out that prosumers show a general tendency to act in an eco-friendly way also in other consumption situations (Gajdzik et al., 2023).

Although appliances that utilize renewable energy sources have been available in the Polish market for a considerable period, they continue to be regarded as a relatively novel technology. Due to their high initial price, equipment such as photovoltaic panels or heat pumps has become available to the wider public only quite recently. Multiple models could be applied to investigate the level of adoption of new technology among the consumer group, leading to the identification of the Purchase Intention. Some of the most widely used, already well-established notions in this context are the Theory of Planned Behavior (TPB) (Hill et al., 1977) and the Theory of Reasoned Action (TRA) (Ajzen, 1991; Ajzen and Madden, 1986). Drawing from those two theories, Davis (1989) proposed a new concept—the technology acceptance model (TAM)—mostly employed to explain the general determinants influencing technology adoption (Musa et al., 2024; Marangunić and Granić, 2015). TAM offers the researchers insights into how the society views and embraces new technologies, studying their intentions based on its two major antecedents: Perceived Usefulness (PU), associated with benefits for the users, and Perceived Ease of Use (PEU), referring to seeing the technology or device as simple and convenient (Elangovan et al., 2024).

Since its formulation, TAM has been applied in multiple study areas related to new technologies, including, for example, Internet used for diverse purposes (Shih, 2004; Liaw and Huang, 2003; Castaneda et al., 2007), mobile technologies (Ngubelanga and Duffett, 2021; Lu et al., 2005; An et al., 2023), tourism (Morosan, 2012; Go et al., 2020), e-learning (Zhang et al., 2008; Cheung and Vogel, 2013), electric mobility (Handarujati, 2024; Zdonek and Melnarowicz, 2023), or robots used for diverse purposes (Parvez et al., 2022; Hogan et al., 2020; Chatzopoulos et al., 2022). The technology acceptance model has also been successfully used in multiple studies regarding various aspects of renewable energy sources or other clean technologies. TAM and its elements have been a basis for research tools in investigations related to photovoltaic panels (Zdonek et al., 2023; Khalid et al., 2021; Bouaguel and Alsulimani, 2022; Alam et al., 2021), solar panels used for water heating systems (Elmustapha et al., 2018; Liu et al., 2022), heat pumps (Yilmaz et al., 2024; Bai et al., 2022), and renewable energy in general (Gârdan et al., 2023; Piselli et al., 2021; Fatoki, 2022).

Despite its widespread use, TAM is not the only model that can be used to examine consumers’ intentions related to renewable energy sources. Other concepts include Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) (Apfel and Herbes, 2021; Aggarwal et al., 2019), social cognitive theory (SCT) (Li and Dai, 2024; Bandura, 2001), or diffusion of innovation (DOI) theory (Franceschinis et al., 2017; Rogers, 2003). DOI introduces five significant innovation features: compatibility, complexity, trialability, observability, and relative advantage, which are believed to be among the best predictors of innovation adoption. There is a theoretical discussion whether relative advantage can be identified with Perceived Usefulness, as both variables concentrate, directly or indirectly, on the superiority of one technology over others. However, relative advantage can be defined as the degree to which the innovation is perceived to be better than current solutions (Franceschinis et al., 2017; Bjørnstad, 2012); therefore, it is strongly connected with comparison to other technologies, while Perceived Usefulness concentrates on the benefits for the user. It could be concluded that these factors might be similar at first, but are not equal to each other, with PU showing a strong direct positive effect on the acceptance of new technologies (Venkatesh et al., 2003).

TAM can be adjusted to a given context as it may be extended in various directions by adding new constructs suitable for the topic (King and He, 2006). Already-used models can be merged to create a tool fitted specifically for the study theme. For example, studies mix TAM with DOI (Tran and Cheng, 2017). Researchers applying TAM often add other factors to PU and PEU, such as PC (also called Perceived Reasonable Cost) (Ali and Naushad, 2023; Alam and Rashid, 2012), social norms or social influence (Grębosz-Krawczyk et al., 2021b; Lee, 2010), financial incentives, frequently used in case of green technologies (Chel and Kaushik, 2018; Yang and Park, 2020), and Environmental Concern (Malik et al., 2020; Gadenne et al., 2011). As some scientists believe basic TAM to be too parsimonious (Bagozzi, 2007; Schierz et al., 2010), the authors decided to use TAM as a foundation and supplement the original factors with the Perceived Cost (PC) and Environmental Concern (EC).

The TAM was selected as the analytical framework due to its proven effectiveness in explaining user acceptance of technological innovations across various contexts, including energy-related behaviors. Compared to more complex models such as the Unified Theory of Acceptance and Use of Technology (UTAUT), TAM offers a more parsimonious structure while maintaining strong explanatory power. Its core constructs—Perceived Ease of Use and Perceived Usefulness—are well established in the literature and have demonstrated reliability in predicting behavioral intentions. Additionally, TAM is highly adaptable, allowing for the incorporation of context-specific variables, which in this study include EC and PC. This flexibility makes TAM particularly suitable for investigating consumer attitudes toward renewable energy technologies in the Polish context.

In the proposed model, the dependent variable is the Purchase Intention (PI) of green energy technology among Polish consumers. The PI can be defined as the likelihood that a consumer actually intends to purchase a specific product (Younus et al., 2015) or the preference of the respondent to buy the product or service in question (Peña-García et al., 2020). The research also includes four control variables in the form of basic demographic factors describing the respondents: gender, age, education, and place of residence.

The following subsections present the independent variables that are potential antecedents of green energy Purchase Intentions.

The survey was carried out on a sample of Polish consumers using a self-administered questionnaire, employing a random selection method (Groves et al., 2011). The survey utilized a database provided by a data collection and market research company, which maintains online panels of 3,000,000 consumers across 19 European countries. The survey was conducted in June 2024, a period marked by intensified public discussions in Poland about the national energy transition, EU environmental policies, and high energy prices. These factors may have influenced consumer perceptions of renewable energy technologies, particularly regarding their usefulness and economic feasibility.

In the initial step, we randomly selected 20 potential respondents to conduct a pretest. These respondents evaluated the survey’s content and relevance, offering feedback that led to minor adjustments for improved readability and clarity. After the trial stage, the final questionnaire was developed. In the second step, a sample of 1,000 consumers was drawn from a database of over 40,000 individuals aged 18 and older. The data collection process yielded a final sample size of 580 consumers, resulting in a response rate of 58%, which is acceptable for this type of survey (Baruch, 1999). The initial sample consisted of 1,000 consumers selected from a database as a representative sample. We received 580 complete responses. While the final sample is not fully representative of the entire Polish population, it is relatively large compared to similar studies conducted in this field, particularly considering the context of a country with 37.7 million inhabitants. This sample size ensures a solid foundation for meaningful analysis and identifying relevant trends. Although certain limitations related to demographic representativeness remain, the diversity and scale of the responses strengthen the robustness and validity of our findings. The sample allows for the exploration of key behaviors and attitudes related to renewable energy, which are likely to reflect patterns present in specific subgroups of the population. To reduce biases, respondents were randomly approached at various times and on different days (Sakshaug et al., 2010). The main characteristics of the respondents are detailed in Table 1.

Most were women aged 35 to 44 with secondary education, and participants were primarily city dwellers from urban areas with populations exceeding 500,000 residents.

The questionnaire comprised 24 questions of three different types. First, respondents were asked about their behavior regarding green energy sources. They could select from three possible answers: (1) “Yes, I use them,” (2) “Yes, I plan to use them within the next 4 years,” or (3) “No.” We explained what green energy sources are before asking the question.

Subsequently, we developed a set of five items to measure Perceived Ease of Use of green energy sources and six items to assess Perceived Usefulness, divided into two dimensions: three items addressing economic usefulness and three items reflecting socially influenced usefulness. Additionally, three items were constructed to evaluate Perceived (Reasonable) Cost, and three items were dedicated to measuring Environmental Concern among consumers. A five-point Likert scale was employed, ranging from 1 (“strongly disagree”) to 5 (“strongly agree”). This scale was chosen due to its suitability for simplifying survey responses while minimizing systematic measurement error (Lee and Turban, 2001). The initial pool of measurement items was generated through an extensive literature review on pro-environmental consumer attitudes and behaviors. These items were subsequently adapted and refined to align with the context and specificity of the green energy market. In constructing the questionnaire, we considered insights from numerous studies examining green consumer behavior. Ultimately, measurement items from Garay et al. (2019), Grębosz-Krawczyk et al. (2021a), Hamari et al. (2016), Malik et al. (2020), Parasuraman (2000), Paul et al. (2016), and Wang et al. (2017) were selected and, where necessary, modified to suit the research objectives. The final set of items is presented in Table 2.

Finally, demographic information about respondents, including gender, age, education level, and place of living, was gathered using a multiple-choice response scale.

To evaluate the validity of our measurements, we performed a confirmatory factor analysis (CFA) using maximum likelihood estimation. This analysis was conducted to determine if the observed measures align with our conceptualization of the construct and to assess how well the data fit a proposed measurement model (Kline, 2015; Nguyen et al., 2020; Podsakoff et al., 2003). Convergent validity of the scale was confirmed by calculating the standardized loadings. The results of this convergent validity, as indicated by the factor loadings, are detailed in Table 3.

Conducting CFA, one item was dropped because of the undesirable values of standardized loadings associated with it (i.e., the values were lower than 0.5) (Kline, 2015; Nguyen et al., 2020; Podsakoff et al., 2003). As a result, we were unable to establish the convergent validity of Environmental Concern.

We also assessed the scale’s convergent validity using average variance extracted (AVE). AVE exceeded 0.5 for all variables except Environmental Concern (EC), which had an AVE of 0.197, indicating that this variable should be excluded. Reliability was evaluated by calculating both standardized and unstandardized Cronbach’s α, as well as composite reliability (CR). The results demonstrate good reliability, with both Cronbach’s α and CR exceeding the 0.7 threshold for all variables except EC. Additionally, in the case of PUS, Cronbach’s α and CR are smaller than 0.7, and AVE is smaller than 0.5. However, according to Hair et al. (2009), Cronbach’s α and CR can be higher than 0.6, and AVE can be higher than 0.4 in the case of new constructs. Table 4 provides details on convergent validity based on AVE, as well as reliability as measured by Cronbach’s α and CR.

In the case of PC, PEU, PI, PUE, and PUS, all statistics are at an acceptable level. Thus, good internal consistency among the items was confirmed.

The Fornell–Larcker criterion suggests that the root square of each construct’s AVE should be higher than the correlation with any other construct (Fornell and Larcker, 1981), as presented in Table 5.

We used structural equation modeling (SEM) with maximum likelihood (ML) estimation, utilizing the covariance matrix as input to evaluate our conceptual model and hypotheses. The ML estimation method, which is widely recommended for theory testing and development (Gerbing and Anderson, 1988), was employed to uncover causal relationships between variables. The model demonstrated a good fit, as indicated by standardized residuals below 0.05. Model fit statistics further support this: χ²/df = 4.121, RMSEA = 0.078, GFI = 0.911, CFI = 0.941, and TLI = 0.917 (Thakkar, 2020). Hypotheses were evaluated based on standardized beta (β) values, t-values, and significance levels. Detailed results from the SEM are presented in Table 6 and depicted in Figure 2.

Only two of the four hypotheses proposed were supported by the data. Hypothesis H1, considering a positive impact of Perceived Ease of Use on Purchase Intention of green energy technology, was confirmed. Similarly, hypothesis H2A, which suggested a positive impact of Perceived Usefulness influenced by Social Factors on Purchase Intention of green energy technology, was also supported. Among these two variables, Perceived Usefulness Influenced by Social Factors (PUS) had a more substantial effect.

However, hypotheses H2B and H3 were not supported. The impact of economic Perceived Usefulness and Perceived Reasonable Cost was found to be statistically insignificant (p > 0.05). Moreover, their impact was negative. Additionally, we could not test hypothesis H4 due to the inability to construct the latent variable for Environmental Concern required for this hypothesis.

When considering the control variables—respondents’ gender, education level, and place of residence—additional relationships were observed. In case of women, we statistically more often observe the impact of PU influenced by social factors on PI of green energy technology (β = 0.314; t = 2.780; p < 0.05). In the case of men, we observe a statistically significant impact of PEU on PI of green energy technology (β = 0.112; t = 2.185; p < 0.05). In the case of people with secondary education, we statistically more often observe the impact of PUS on PI of green energy technology (β = 0.271; t = 2.865; p < 0.05). However, in the case of people with primary education, we observe a statistically significant impact of economic PUE on PI of green energy technology (β = 0.870; t = 5.211; p < 0.05). Age and place of residence do not have a statistically significant impact on the PI of green energy technology.