How higher‐order personal values affect the purchase of electricity storage—Evidence from the German photovoltaic market

Poier, Stefan; Nikodemska‐Wołowik, Anna Maria; Suchanek, Michał Article — Published Version How higher‐order personal values affect the purchase of electricity storage—Evidence from the German photovoltaic market Journal of Consumer Behaviour Provided in Cooperation with: John Wiley & Sons Suggested Citation: Poier, Stefan; Nikodemska‐Wołowik, Anna Maria; Suchanek, Michał (2022) : How higher‐order personal values affect the purchase of electricity storage—Evidence from the German photovoltaic market, Journal of Consumer Behaviour, ISSN 1479-1838, Wiley, Hoboken, NJ, Vol. 21, Iss. 4, pp. 909-926, https://doi.org/10.1002/cb.2048 This Version is available at: https://hdl.handle.net/10419/265068 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by/4.0/ ACADEMIC PAPER How higher-order personal values affect the purchase of electricity storage—Evidence from the German photovoltaic market Stefan Poier 1,2 | Anna Maria Nikodemska-Wołowik 1 | MichałSuchanek 2 1 Faculty of Economics, University of Gdansk, Sopot 2 FernUniversität in Hagen Correspondence Stefan Poier, Faculty of Economics, University of Gdansk, Armii Krajowej 119/121, 81-824 Sopot; FernUniversität in Hagen. Email: [email protected]du.pl Funding information This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Abstract Why do consumers make different decisions even when socioeconomic conditions are similar? The present article examines the effects of human values—as formulations of motivational goals—on the decision to purchase electricity storage for a photovoltaic system, a phenomenon hardly explored in prior research regarding highpriced household investments. About 50 percent of photovoltaic-system owners in Germany are also owners of an electricity storage. This study aims to explore the more deeply rooted motivational factors behind these different decisions to extend our understanding of consumers' decision-making processes regarding energyefficiency investments. It is based on an online survey of 460 owners of residential photovoltaic systems in Germany in 2019 and focuses on the interplay between higher-order values, purchase decisions, perceived risk, and environmental concern. The analysis of the higher-order values showed direct effects of conservation and self-transcendence and indirect effects of openness to change and self-enhancement, both mediated by perceived risk. 1|INTRODUCTION Over the past decade, electricity-storage batteries have become increasingly affordable for owners of residential PV systems, making a cost-neutral self-supply of electricity possible. However, approximately 50 percent of households that invest in a PV system decide against a storage solution (Figgener et al., 2018, p. 36). Given similar sociodemographic characteristics of PV owners, e.g., age, income, household size, and education level (Jacksohn et al., 2019, pp. 222– 223), investigating which other reasons, deeply rooted in users' value systems, determine these decisions is valuable. As research on bounded rationality and behavioral economics (Barros, 2010; Beck, 2014; Simon, 1955,1993) has shown, consumers do not evaluate only on the basis of pure facts since the core element of behavioral economics is the idea of biases and heuristics (Del Campo et al., 2016; Gigerenzer & Gaissmaier, 2011; Mousavi & Gigerenzer, 2014). Both lead to non-optimal results in decisionmaking. Heuristics are simplification strategies that serve to facilitate and expedite the decision-making process by ignoring some information (Tversky & Kahneman, 1974; Willman-Iivarinen, 2017,p.2; Zhang et al., 2020, p. 795). Biases, resulting from heuristics, can be referred to as “cases in which human cognition reliably produces representations that are systematically distorted compared to some aspect of objective reality”(Acciarini et al., 2021, p. 641). In other words, individuals interpret objective reality according to their values, biases, and perceptions (Acciarini et al., 2021, p. 644), which implies that their interpretations of reality vary depending on their individual value dispositions. Contrary to the utility maximization theory, which implies that there is one “right”decision that all individuals in a This research was based on part of the first author's dissertation, conducted under the guidance of the second author. Received: 3 June 2021 Revised: 27 February 2022 Accepted: 15 March 2022 DOI: 10.1002/cb.2048 This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2022 The Authors. Journal of Consumer Behaviour published by John Wiley & Sons Ltd. J Consumer Behav. 2022;21:909–926. wileyonlinelibrary.com/journal/cb 909 comparable situation would make, observed behavior is based intuitively on individual motivations, emotions, and cognitive biases (ElHaffar et al., 2020). Consumers attempt to imagine how they would feel using a product, how their environment would respond to the decision, and how the decision would align with their individual desires and needs, which are determined by superordinate values and goals (Liebel, 2007, p. 453). The role of values in the decision-making process has been thoroughly documented in prior studies (Anic et al., 2019; Berglund & Matti, 2006; Farah & Fawaz, 2016). Since values are to be understood as guiding principles in life, i.e., as a desirable state, they determine the direction of consumer decision-making (Parks & Guay, 2009, p. 676). Egoistic individuals with strong self-enhancement values are more likely to choose an option that enhances their own advantage, while altruistic individuals with strong self-transcendence values are more likely to care about the common good (Bouman et al., 2018). Personal human values determine the way a situation is interpreted (Sagiv & Roccas, 2021, p. 10), for example, how strongly a risk is perceived that discourages a purchase. According to de T'Serclaes (2010), risk perception is a major barrier to energy-efficiency investments. In the case of proenvironmental behavior, the strength of the perceived threat to nature may influence whether or not such a behavior is carried out. This can be, for example, simply switching off an unneeded electrical device (Kastner & Stern, 2015), but it can also be a costly investment in energy-efficiency measures (Kastner & Matthies, 2016; Ramos et al., 2016), provided that this is interpreted as environmentally friendly behavior. Individuals attempt to align their behavior with their values and tend to purchase products that match these values (Voorn et al., 2018). Therefore, values can contribute to consumer loyalty and prevent the consumer from switching brands (Voorn et al., 2021). In addition to values, personality traits are among the key content aspects of personality (Sagiv & Roccas, 2021, p. 2). Although they are considered to be independent constructs, recent research has revealed moderate relationships (Fischer, 2018; Parks-Leduc et al., 2015). McCrae and Costa (1999) and McCrae et al. (2000) considered that personality traits are expressed as characteristic adaptations, which include human values (Fischer & Boer, 2015, p. 493). Since traits determine what individuals are like and values indicate what is important to them in life, it can be said that “traits shape values in interaction with the local environment”(Fischer & Boer, 2015, p. 493). In recent years, Busic-Sontic and Fuerst (2018) and Busic-Sontic et al. (2017) have conducted pioneering research regarding the effects of personality traits on making energy-efficiency investments. Based on this, Poier (2021) investigated the influence of traits on decisionmaking between household members in the context of the purchase of photovoltaic systems. However, relatively little is known about the contribution of human values to the purchase of high-priced energy-efficiency technology. Therefore, in general this research deepens and broadens prior studies on pro-environmental consumer behavior, as an extensive evaluation of sustainable consumer behavior is still underrepresented in the existing literature (Agrawal & Gupta, 2018; Kuanr et al., 2021; Marzouk & Mahrous, 2020). In particular this study narrows the gap in research regarding how consumer behavior based on individual dispositions like basic human values can be more accurately inferred. Regarding pro-environmental behavior, most studies have examined everyday behaviors such as saving electricity and water, utilizing reusable packaging, and buying organic food (Gatersleben et al., 2014; Honkanen et al., 2006; VegaZamora et al., 2020). This research proposes a conceptual framework grounded in existing interdisciplinary literature for studying how individual dispositions among consumers affect their behavior. To the best of our knowledge, the effects of consumers' individual value dispositions on their decision to purchase a battery storage for selfgenerated solar power have not been investigated in previous studies. To assess these products' features, extensive knowledge and processing capacity on the part of the consumer is required. Moreover, values that develop from early childhood to young adulthood may explain differences in consumer behavior. In the aforementioned context, the role of environmental concern in the purchase decision as a control variable and as a mediator was examined. The remainder of this article is organized in the following way. A review of the literature and a detailed description of the theoretical foundation of the model and the formulation of hypotheses is provided in Section 2. Section 3presents the source of the data and describes how the working sample is prepared. After that, the results are presented in Section 5. This is followed in Section 6by a discussion of the results including theoretical contributions and practical implications. The article ends with the conclusions, limitations of the study and an outlook for future research. 2|LITERATURE REVIEW AND HYPOTHESES The history of modern values research goes back to the work of Vernon and Allport (Cantril & Allport, 1933; Vernon & Allport, 1931), who developed a personality measurement instrument called “A Study of Values.”Since the 1970s, several value theories have been developed that can be roughly divided into two areas of research at the individual level and at the cultural level (Hanel et al., 2018). The three most prominent value theories at the cultural level were developed by Inglehart (1977), Hofstede (1980), and Schwartz (1999). For this research article, however, tools that measure individual values are of importance, for example, the Rokeach Value Survey (RVS) (Rokeach, 1973), the Values, Attitudes, and Lifestyles (VALS) framework, which builds a connection between values and lifestyle (Mitchell, 1984), and Kahle's List of Values (LOV) (Homer & Kahle, 1988; Kahle et al., 1986). The Schwartz Theory of Basic Human Values (Schwartz & Bilsky, 1987,1990) is among the most popular and researched models to date (Knafo et al., 2011, p. 181; Sagiv & Roccas, 2021, p. 3). In current psychological research, the term “values”is defined by Schwartz (1999, p. 24) as “(…) conceptions of the desirable that guide the way social actors (e.g., organizational 910 POIER ET AL. leaders, policy-makers, individual persons) select actions, evaluate people and events, and explain their actions and evaluations.”In the context of purchase situations, consumers seek to achieve certain goals through the acquisition of a commodity (Silberer, 1995, p. 2704). Although values are subjective for an individual consumer, they are generally shared within the social or cultural group. Values develop over the long term through individuals' experiences, education, and socialization (Bilsky et al., 2011). Schwartz expanded the individual value system of the earlier models to 10 and assumed that each type of value formulates a specific motivational goal. The 10 values, universal across cultures, were universalism, benevolence, conformity, tradition, security, power, achievement, hedonism, stimulation, and self-direction; they have been widely analyzed and confirmed by most studies (Bilsky et al., 2011; Schwartz, 1994; Schwartz & Boehnke, 2004). A revision of Schwartz's theory revealed a more differentiated structure with 19 values (Cieciuch et al., 2014; Cieciuch et al., 2015; Schwartz, 2017; Schwartz et al., 2012). This refinement was deemed necessary to better reflect the actual core idea of a circular continuum of values. Schwartz identified four higher-order values, each comprising several single values. Humility and face form their own values in the revised 19 values, which were not present in the original model. Like hedonism, each lies on the borders between two higher-order values. According to the Coding and Analysis Instructions (Schwartz, 2016b), humility aligns best with selftranscendence, while face should be assigned to self-enhancement. The first higher-order value, self-transcendence, comprises universalism nature, universalism concern, universalism tolerance, benevolence caring, benevolence dependability, and humility. Individuals with high values of self-transcendence appreciate the well-being of others and a sense of community and nature. The second higher-order value is openness to change, comprising self-direction thought, selfdirection action, and stimulation. People scoring high on openness to change seek excitement and new experiences, and they demonstrate independent thinking and self-realization. The third value cluster is called self-enhancement and includes achievement, power dominance, power resources, and face. Individuals with a high self-enhancement level seek prestige, personal success, and social status; they want to dominate others. Hedonism can be assigned to both openness to change and self-enhancement (Bilsky & Schwartz, 1994). The fourth higher-order value, conservation, comprises security personal, security societal, tradition, conformity rules, and conformity interpersonal. People with high conservation scores prefer safety and traditional values, desire security, and dislike improper behavior. While selfenhancement is about one's own advantage and dominance over others, self-transcendence focuses on the well-being of others. In openness to change, the focus is the view beyond one's own nose and the expansion of one's horizon that counts; in contrast, with conservation, it is the preservation of the status quo and adherence to rules (Schwartz, 2012). Figure 1illustrates the circular arrangement of the 19 revised values. Several researchers investigated the influence of values on consumers' purchase decisions (Kaže, 2010; Kostelijk, 2015; Krystallis et al., 2012;Le˜ao et al., 2007). Furthermore, regarding a general examination of the relationship between values and behavior, Cieciuch (2017) emphasized the simultaneous role of values as both drivers of and obstacles to behavior. Values affect the attractiveness of choice options as well as the perception and evaluation of alternatives (Honkanen et al., 2006; Schwartz, 2005). The connection between values and involvement was examined by O'Cass (2001) with the result that the involvement with an object is higher when it is strongly connected to the consumer's values. As for comparing values with personality traits, a study by Voorn et al. (2018, p. 537) found that values were the stronger factors in a consumer–brand relationship. They played a greater role in future decisions than traits, which is an indication of the importance of values for purchasing decisions. 2.1 |Conceptual framework Aride and Pàmies-Pallisé (2019) and Wilson and Dowlatabadi (2007) used different approaches to introduce two integrated models of consumer decision-making. Both models proposed that it should be possible to explain differences in behavior, at least in part, through differences in consumers' individual value dispositions. A synthesis of these individual theories, models, and studies used in the present investigation is presented in Figure 2. Human values form the starting point of the value-belief-norm (VBN) theory, with the new ecological FIGURE 1 Modified version of the motivational continuum of values. In the innermost circle, the circumplex structure of the 19 revised human values is shown. Surrounding this, the four higherorder values along the two dimensions openness versus conservation and self-transcendence versus self-enhancement are shown. Moving outward, the next circle divides the values into social and personal focus. The outermost circle distinguishes between the dimensions anxiety-avoidance and anxiety-free. Source: Own visualization, based on Schwartz et al. (2012, p. 669) POIER ET AL.911 paradigm (NEP) in second place (Stern, 2000, p. 412). According to the VBN theory, personal norms affect not only behavior; they are the link between the VBN theory and the norm activation model (NAM), according to Schwartz and Howard (1981). The prerequisite for both theories is that personal norms are influenced by the consumer's ecological worldview. Concretely, this means they allow human values to influence purchase decisions only if consumers see the purchase as contributing to environmental protection or consider environmental concern as the reason for it. For the structural model, this meant that environmental concern, measured by the NEP, was included as a mediator variable in the model, in the case where consumers see the purchase of a battery storage system as a contribution to environmental protection. Although the protection of nature as a facet of the value “universalism nature”on the one hand and environmental concern on the other seem to say something similar at first glance, they differ from each other. Basic human values are, by definition, “conceptions of the desirable that guide the way social actors (…)selectactions”(Schwartz, 1999). Values, therefore, indicate what is important to the individual (Schwartz, 2016a, p. 63). In contrast, the new ecological paradigm measures the individual's beliefs about nature and environmental issues, and thus, his or her level of environmental concern (Bagozzi et al., 2002). Therefore, values (in this case, universalism nature) and beliefs are two different constructs of consumer behavior (Steg & de Groot, 2012). This also becomes obvious in the value-belief-norm theory, which proposes that values influence beliefs about the environment, and thus, human values are mediated by environmental concern (Stern, 2000, p. 412). The NAM has subjective norms and perceived behavioral control in common with the theory of planned behavior (TPB), which ends with the individual's behavior (Ajzen, 1991). Thus, both the NAM and VBN theory identify behavior as a consequence of personal norms. In addition, with subjective norms and perceived behavioral control, the NAM forms two of the three prerequisites for the TPB. The final requirement for the TPB, attitudes, is influenced by values via worldviews and is the missing link established by the value-attitudebehavior (VAB) hierarchy between human values and attitudes (Homer & Kahle, 1988). Thus, values have an effect on the purchase decision through attitudes even without the mediating function of environmental concern (Schwartz, 2016a, p. 72). 2.2 |Development of hypotheses Following Wolske et al. (2017), renewable energy technology can be seen in different ways as a consumer good, an investment, a possibility to increase one's well-being, an innovative technology, and a contribution to environmental protection. Its perspectives as a consumer good and an investment justify viewing the subject of the study as a consumer or household decision. Accordingly, sociodemographic control variables were derived. The installation of novel energy-efficiency measures, such as electricity storage batteries, implies uncertainty and risk for consumers (Busic-Sontic & Brick, 2018, p. 2). Perceived risk has been one of the most important determinants of consumer behavior research since the 1960s (Li et al., 2020, p. 77). Dowling (1986) defined it as the uncertainty consumers face when making a purchase and it is commonly understood as a “cognitive evaluation of outcome probability and outcome severity”(Loewenstein et al., 2001; Slovic, 2011, pp. 364–371; Wolff et al., 2019, p. 3). In this context, the risk assessment, that is, the strength of risk perception, is influenced by the consumers' basic human values, in addition to other variables (Rundmo et al., 2011). Equipping one's household with energy-efficiency measures such as solar-power systems or battery storage requires a substantial financial investment that is often financed through loans. Future savings are uncertain and, like the service life of such systems, can only be estimated, resulting in a significant financial risk for the consumer (Rockstuhl et al., 2021). Moreover, FIGURE 2 Conceptual framework. Figure 2illustrates different models of consumer behavior with their interfaces, starting from human values and moving to intention, norms, and behavior. AC, awareness of consequences; AR, ascription of responsibility; NAM, norm activation model; NEP, new ecological paradigm; TPB, theory of planned behavior; VAB, value-attitude-behavior; VBN, value-belief-norm theory. Source: Own visualization, based on Poier (2021) 912 POIER ET AL. the potential for fire or defects in the battery storage system pose a risk to consumers' health and safety, as well as to their property. Many installations in one's own home depend on the companies and the people entrusted with them; therefore, trust in the installer is highly significant in relation to the purchase of a PV or storage system. Rai, Reeves, and Margolis (2016, p. 504) found that more than 50% of prospective PV customers considered the installer's opinion very important or even extremely important. Since trust presupposes a risk situation (Fjaeran & Aven, 2021; Luhmann, 1988), the study also inquired about perceived risk in the context of trust in people involved. A meta-analysis conducted by Li et al. (2020, p. 90) found a direct negative relationship between perceived risk and purchase behavior. Self-enhancement is located in the value circle in a segment that combines fear avoidance and focus on the self. Individuals who score high in power and achievement strive to preserve or increase their own advantage and react more strongly to threats to their own interests (Schwartz, 2016a; Schwartz et al., 2000, pp. 318–319). Thus, the following hypotheses were developed: H_01: Perceived risk has a negative influence on the battery purchase. H_02: Self-enhancement has a positive effect on perceived risk. Regarding environmental concern, Kollmuss and Agyeman (2002) proposed a model according to which values were predecessors of proenvironmental behavior. Remarkably, most PV-storage owners cited a hedge against rising electricity prices as their primary motive for the purchase or said they wanted to contribute to energy-system transformation (Figgeneretal.,2018, p. 55). Therefore, at least the intention to make a contribution to the energy-system transformation—and thus to the environment as a common good—is about as rigid as the selfish motivation to save money. Individuals scoring high on openness seek challenges in life and want to implement their ideas and plans (Cieciuch & Schwartz, 2017). Together with persons with a high self-transcendence level, they are located in the sphere of anxiety-free values (Schwartz, 2017). They are interested not only in their own concerns but also in the environment. Primc, Ogorevc, Slabe-Erker, Bartolj, and Murovec (2021, p. 287) found that self-direction had a positive influence on environmental concern. Therefore, we expect a negative influence of openness on perceived risk and a positive effect on environmental concern, and we develop the following hypotheses: H_03: Openness to change has a negative influence on perceived risk. H_04: Openness to change has a positive effect on environmental concern. Individuals with pronounced self-transcendence have a social rather than a selfish focus. They care about the well-being of others and the protection of nature (Schwartz, 2016a). Further research on pro-environmental consumer behavior was conducted by Pinto et al. (2011). In a study on responsible water consumption in Southern Brazil, they concluded that socially oriented participants used water more responsibly than self-oriented individuals. In a study concerning the applicability of a new environmental-values scale, Bouman et al. demonstrated an effect of values on beliefs, norms, and behavior. For this purpose, they used a scale based on the measurement of values underlying environmental beliefs and behaviors. Of the 17 items used, 15 were from the Schwartz values continuum, rearranged to measure biospheric and altruistic values (representing self-transcendence values in the original values circle) and egoistic and hedonic values (representing the original self-enhancement values). The main findings were that biospheric and altruistic values lead to stronger climatechange beliefs, stronger pro-environmental personal norms, a greater willingness to engage in energy-saving behavior, and support for governmental sustainability investments (Bouman et al., 2018, pp. 9–11). Gatersleben et al. investigated the relationship between values, identity, and pro-environmental behavior among UK residents; they also used a modified scale to measure biospheric, altruistic, and egoistic values. Concerning values, they found that biospheric values most strongly predicted pro-environmental behavior (Gatersleben et al., 2014, p. 388). Primc et al. (2021, p. 287) argued that the opposite is true for egoistic, or self-enhancement, values, thus leading us to develop the following hypotheses: H_05: Self-enhancement has a negative effect on environmental concern. H_06: Self-transcendence affects environmental concern positively. H_07: Self-transcendence has a positive effect on the battery purchase. Individuals with high conservation scores are eager to go along with the majority opinion instead of going their own way. They strive for security and stability and want to maintain the status quo without experimenting. Thus, they are unlikely to invest in a new and uncertain technology. Since their main interest lies in protecting themselves and their close associates, interest in global issues and protecting the environment plays a secondary role (Schwartz, 2016a). Therefore, we considered the following hypotheses: H_08: Conservation has a negative effect on the battery purchase. H_09: Conservation affects environmental concern negatively. With increasing electricity demand, investing in renewable forms of energy becomes a growing need. Yet only every second residential PV system in the German market is sold with battery storage (Figgener et al., 2018, p. 36). Environmental concern and perceived risk are among POIER ET AL.913 the possible influencing factors in energy-efficiency investments (BusicSontic et al., 2017, pp. 314–315). Consumers' varying value dispositions influence how they evaluate the different dimensions of the construct of perceived risk (Sutalaksana et al., 2019, p. 919), which in turn acts as an obstacle to purchase decisions. Therefore, perceived risk was included in the model as a mediator variable. Figure 3summarizes the hypotheses of the study and shows the direct and indirect relationships considering the control variables. Whether environmental concern plays a mediating role in the purchase decision due to the influence of human values should be investigated. This leads to the question of the reason for the heterogeneity of consumer decisions and to the major research question: How do the differences in human values contribute to the decision to buy a battery storage device? 3|MATERIALS AND METHODS 3.1 |Survey design To answer the research question and to accept or reject the aforementioned hypotheses, a survey was conducted via a paid consumer panel from October to November 2019. The target group included adult owners of a self-inhabited property with a PV system. An online questionnaire was distributed via a market research institute to 2828 panel participants, of whom 2266 started to respond to the survey. The questionnaire also offered the option of being sent by email to other owners of PV systems. In addition, the link to the questionnaire was published in the German online community “Photovoltaik-Forum,”the largest German online community for people interested in renewable forms of energy. A total of 844 community and snowball participants, according to Atkinson and Flint (2004), clicked on the link to the questionnaire. The aim of the study was not to map a representative crosssection of the population but to identify the difference between two consumer groups. Since both groups had to be owners of photovoltaic systems, both the participants from the online forum and the panel participants fulfilled this criterion for the study. A total of 2603 individuals edited the survey, and 605 of these completed the questionnaire. After deleting all cases in which the participant had not given serious answers, did not own a PV system, or had not reached at least the page with risk perceptions, 587 cases remained. In a further step, all cases with obviously incorrect or tooquick answers were deleted. Since only decision-makers were to be examined for the investigation, only those individuals who had made the purchase decision either alone or with a partner were included in the working sample. Therefore, the resulting working sample comprised 460 persons. 3.2 |Construction of variables Since a positive attitude toward a product or the intention to buy it does not necessarily result in a positive purchase decision—a phenomenon known as the energy-efficiency gap (Allcott & Greenstone, 2012, p. 19; Häckel et al., 2017)—we did not measure attitude toward batteries or purchase intention. Instead, the actual purchase decision was examined as the dependent variable. To obtain comparable results, we decided not to develop our own scales for the measurement of individual differences. However, this study relied on sufficiently well-known and widely tested scales from the fields of personality and values research. 3.3 |Higher-order values The basic human values were measured with a 57-item scale, the Portrait Values Questionnaire (PVQ)-57 on a six-point Likert-type scale where 1 denoted “Not like me at all”and 6 “Very much like me” (Schwartz et al., 2012). Each item consisted of one statement about a person, and each statement was formulated to take gender into consideration (“it is important to her…”;“it is important to him…”). For each of the four higher-order values, a raw average score and a centered score were calculated. The centered score is the raw score reduced by the participant's average score over all items and is thus adjusted for the participant's scale use preferences. In the Coding & Analysis Instructions (Schwartz, 2016b), Schwartz provided details for how the individual items should be assigned to the different values. Regarding the 10 original individual values, “face” was not considered to contribute to a certain value, while “humility” items were part of tradition. Within the revised 19 values, both form their own values. In the majority of about 100 samples, “humility”fits best with “self-transcendence”and “face”is best combined with “self-enhancement”regarding the higher-order values (Table A1 in the Appendix A). For t-tests, higher-order values were computed as the mean of all related items. FIGURE 3 Hypotheses in the structural equation model. Figure 3 presents the hypotheses (H_01–H_09) within the structural equation model with mediation of higher-order values through perceived risk (Risk) and environmental concern (EC) on battery purchase with CO, conservation, OC, openness to change, SE, self-enhancement, and ST, self-transcendence 914 POIER ET AL. Since many owners of PV systems indicated protection against electricity price increases or power failure as a reason for their purchase, safety values could also contribute to the decision-making process (Figgener et al., 2018), while for technology enthusiasts, “openness”values could play an important role. Environmental concern was measured in detail with the 15-item NEP scale, according to Dunlap, van Liere, Mertig, and Jones (2000, p. 433). 3.4 |Socioeconomic variables Sociodemographic variables such as age, gender, income, and household size were included as control variables. In addition, economic variables such as feed-in tariff and average electricity price in the year of installation of the PV system were taken into account. Finally, the analysis was controlled for two variables that were investigated in the context of energy-efficient investments: perceived risk as an obstacle to investment and concerns about the environment as a possible altruistic or ecological driver of the investment. However, both constructs were considered not only as control variables but also as mediator variables. Age, gender, and level of education were observed variables as well as household income and educational level. Household income was a selection field with nine gradations from 0 to more than 5000 euros per month. The variable educational level was generated from school education and vocational education (university degree), based on the International Standard Classification of Education (ISCED). The average electricity price as well as the average feed-in tariff in the year of PV installation were calculated using data provided by the BDEW for that year (BDEW Bundesverband der Energieund Wasserwirtschaft e. V., 2020). 3.5 |Environmental concern A standard scale was also used to measure environmental worldview. In its current form, the New Ecological Paradigm (NEP) scale comprises 15 statements on environmental attitude (Cronbach's α=.812), each rated on a Likert-type scale from 1 (do not agree at all) to 5 (fully agree). The eight odd-numbered items represent a pro-ecological worldview, while the seven even-numbered items reflect a more anthropocentric worldview according to the dominant social paradigm (DSP). To calculate the five individual facets of the NEP, the evennumbered pro-DSP items had to be reverse-coded so that all items measured the same dimension. A German translation of the original scale was adapted from a study by Menzel and Bögeholz (2010). 3.6 |Perceived risk For the construction of the global variable for perceived risk (Cronbach's α=.846), four modified questions from the SocioEconomic Panel (SOEP), a representative nationwide German household study, were used (Goebel et al., 2019). These originally asked about the consumer's risk preference and were reformulated in this study to capture the strength of the perceived risk (“Did you have the feeling of risk about purchasing the solar power storage?”). They included the domains of general and financial risk, risk related to the health of the participant, and risk related to trust in the persons or companies involved. The answer options on the Likert-type scale ranged from 0 (no risk) to 10 (high risk). 4|ANALYSIS AND RESULTS Since decision-making plays the central role in this study, only those survey participants who made the decision for or against purchasing a battery either by themselves or with a partner were considered. In the case of non-owners, it was assumed that decision-makers regarding the PV system would also decide about the battery; therefore, these persons were considered the decision-makers regarding storage as well. Of these 460 participants involved in the decision-making process, 173 (37.6%) owned a battery storage system. Of the 173 storage owners, 108 had made the purchase decision alone, while 65 had agreed on the decision with their partner. Eight stated that the battery was their partner's idea, and 38 said the idea was a joint one; the remaining 127 had the idea themselves. Of the 287 non-owners of a storage battery, 191 stated that they had made the decision to buy the PV system on their own, and 96 had made the decision with their partner. All control variables except for gender and environmental concern showed significant differences between the two groups (see Tables A2 and A3 in the Appendix A). 4.1 |Statistical analysis To test the hypotheses, the latent variables were included in a structural model that depicts the dependencies of higher-order values, perceived risk, environmental concern and battery purchase. SEM was chosen as the analysis method to estimate the strength and significance of the effects between the interrelated constructs and to account for measurement error. The Analysis was conducted using Mplus, a statistical software to conduct structural equation modeling with latent variables and factor analyses with a binary outcome as it was needed in the present study (Muthén & Muthén, 1998–2020; Muthén et al., 2016). Mplus also provides the possibility to report total indirect, direct and total effects, even with a binary outcome (for a detailed explanation see Muthén et al., 2016, pp. 307–330). In addition, age, gender, household size, feed-in tariff, level of education, average electricity price, and household income were included as control variables that were regressed on battery ownership in the first model and also on environmental concern and perceived risk in the second analysis. To avoid misspecifications resulting from those assumptions that do not reflect reality, we used exploratory structural equation modeling (ESEM) in the final analyses (Asparouhov & Muthén, 2009; Marsh et al., 2020). The difference with confirmatory POIER ET AL.915 factor analyses (CFA) is that cross-loadings to non-intended items are not constrained to zero but allowed as in exploratory factor analysis (EFA). The following values represent common guidelines for an acceptable model fit: comparative fit index (CFI) > 0.90, Tucker-Lewis index (TLI) > 0.90, root mean square error of approximation (RMSEA) < 0.08, and standardized root mean square residual (SRMR) < 0.08 (Hu & Hu & Bentler, 1999; Marsh et al., 2010). A good model fit can be assumed with the following values: CFI > 0.95, TLI > 0.95, RMSEA < 0.05, and SRMR < 0.05. However, it is important to note that scales for personality research with real-world data and many items and factors rarely come close to a mediocre fit (Marsh et al., 2010, p. 477). As an example, a fit of CFI =0.744, RMSEA =0.062 was considered acceptable during the development of the regular German Big Five Inventory 2 (BFI-2) measurement instrument (Danner et al., 2016). If the fit is sufficient, it can be assumed that the data collected have satisfactory validity. To achieve an acceptable,ifnotgood,modelfit,itis almost always necessary in social science practice to adjust the data without distorting the underlying theories and models. In the present study, this was accomplished through a series of EFAs and CFAs. First, an EFA was performed, revealing which items loaded on the intended factors (i.e., were good predictors) and which items had overly strong significant cross-loadings (i.e., were weak predictors). Following Cieciuch and Davidov (2012, p. 40) and Purc and Laguna (2019, p. 8), a CFA was conducted for all four higher-order values separately to evaluate model fit. With the original item structure, the results suggested that the models of higher-order values were not well-represented by the data. As a consequence, two methods were used to improve the model fit until the CFI reached a value of at least .90: 1. Items with non-significant loadings were excluded from the CFA; 2. For questions with very similar wording, it was assumed that the residuals also correlate. Following Cieciuch and Davidov (2012), this is the justification to allow those residuals to be correlated. Table 1shows the results for higher-order values' model fits. In the case of the NEP scale, the first step was to follow the approach of Xiao, Dunlap, and Hong (2019, p. 63) in order to increase the model fit. For this purpose, the item with the lowest factor load was deleted from each of the five domains. The result was a coherent measurement tool for the NEP, which still covered all five facets with an improved—but not yet good—model fit. Subsequently, as with the human values, those residuals with a very similar meaning or wording were allowed to be correlated. For perceived risk with four items (trust, health, financial, general), an acceptable model fit could be determined; therefore, this construct was not further optimized (Table 1). In order to avoid misspecification of the models due to low case numbers, multiple imputation had to be performed to reduce missing TABLE 1 Model fit for higher-order values and mediator variables Higher-order value CFI TLI RMSEA SRMR Openness to change 0.927 0.903 0.092 (0.079/0.102) 0.049 Self-enhancement 0.913 0.873 0.108 (0.095/0.122) 0.069 Conservation 0.940 0.918 0.075 (0.065/0.085) 0.055 Self-transcendence 0.923 0.900 0.075 (0.067/0.082) 0.054 Mediator variable CFI TLI RMSEA SRMR NEP 0.957 0.928 0.061 (0.045/0.078) 0.043 Perceived risk 0.942 0.826 0.232 (0.180/0.289) 0.039 Note: Table 1presents model fit indices of four higher-order values and the NEP as a measure of environmental concern and for perceived risk using CFA. Abbreviations: CFA, confirmatory factor analysis; CFI, comparative fit index; NEP, new ecological paradigm; RMSEA, root mean square error of approximation with 90% confidence interval in brackets; SRMR, standardized root mean square residual; TLI, Tucker-Lewis index. FIGURE 4 Significant regressions of battery purchase on higherorder values. Figure 4presents significant regression paths of battery ownership on higher-order values; estimator =WLSMV, multiple imputation generated 50 datasets (n=383 on average). Solid lines stand for positive (+) effects; dotted lines indicate negative () effects. CO, conservation; EC, environmental concern; Education, no degree; OC, openness to change; risk =perceived risk; SE, selfenhancement; ST, self-transcendence 916 POIER ET AL. Muthén, B. O., Muthén, L. K., & Asparouhov, T. (2016). Regression and mediation analysis using Mplus. Muthén & Muthén. Muthén, L. K., & Muthén, B. (1998–2020). Mplus User's Guide: Version 8 (8th ed.). Muthén & Muthén. O'Cass, A. (2001). An exploratory study of the influence of personal values on Purchase Decision & Advertising Involvement. Asia Pacific Advances in Consumer Research,4,67–72. Parks, L., & Guay, R. P. (2009). Personality, values, and motivation. Personality and Individual Differences,47(7), 675–684. https://doi.org/10. 1016/j.paid.2009.06.002 Parks-Leduc, L., Feldman, G., & Bardi, A. (2015). Personality traits and personal values: A meta-analysis. Personality and Social Psychology Review, 19(1), 3–29. https://doi.org/10.1177/1088868314538548 Pinto, D. C., Nique, W. M., Añaña, E. D. S., & Herter, M. M. (2011). Green consumer values: How do personal values influence environmentally responsible water consumption? International Journal of Consumer Studies,35(2), 122–131. https://doi.org/10.1111/j.1470-6431.2010. 00962.x Poier, S. (2021). Towards a psychology of solar energy: Analyzing the effects of the big five personality traits on household solar energy adoption in Germany. Energy Research & Social Science,77, 102087. https://doi.org/10.1016/j.erss.2021.102087 Primc, K., Ogorevc, M., Slabe-Erker, R., Bartolj, T., & Murovec, N. (2021). How does Schwartz's theory of human values affect the proenvironmental behavior model? Baltic Journal of Management, 16(2), 276–297. https://doi.org/10.1108/BJM-08-2020-0276 Purc, E., & Laguna, M. (2019). Personal values and innovative behavior of employees. Frontiers in Psychology,10,1–16. https://doi.org/10.3389/ fpsyg.2019.00865 Rai, V., Reeves, D. C., & Margolis, R. (2016). Overcoming barriers and uncertainties in the adoption of residential solar PV. Renewable Energy, 89, 498–505. https://doi.org/10.1016/j.renene.2015.11.080 Ramos, A., Labandeira, X., & Löschel, A. (2016). Pro-environmental households and energy efficiency in Spain. Environmental and Resource Economics,63(2), 367–393. https://doi.org/10.1007/s10640-015-9899-8 Roccas, S., Lilach, S., Schwartz, S. H., Ariel, K., Sagiv, L., & Knafo, A. (2002). The big five personality factors and personal values. Personality and Social Psychology Bulletin,28(6), 789–801. https://doi.org/10.1177/ 0146167202289008 Rockstuhl, S., Wenninger, S., Wiethe, C., & Häckel, B. (2021). Understanding the risk perception of energy efficiency investments: Investment perspective vs. energy bill perspective. Energy Policy,159, 112616. https://doi.org/10.1016/j.enpol.2021.112616 Rokeach, M. (1973). The nature of human values. The Free Press. Rundmo, T., Nordfjærn, T., Iversen, H. H., Oltedal, S., & Jørgensen, S. H. (2011). The role of risk perception and other risk-related judgements in transportation mode use. Safety Science,49(2), 226–235. https:// doi.org/10.1016/j.ssci.2010.08.003 Sagiv, L., & Roccas, S. (2021). How do values affect behavior? Let me count the ways. Personality and Social Psychology Review,25,1–22. https://doi.org/10.1177/10888683211015975 Schultz, P. W. (2001). The structure of environmental concern: Concern for self, other people, and the biosphere. Journal of Environmental Psychology,21(4), 327–339. https://doi.org/10.1006/jevp. 2001.0227 Schultz, P. W., & Zelezny, L. C. (1999). Values as predictors of environmental attitudes: Evidence for consistency across 14 countries. Journal of Environmental Psychology,19(3), 255–265. https://doi.org/10.1006/ jevp.1999.0129 Schwartz, S. H. (1994). Are there universal aspects in the structure and contents of human values? Journal of Social Issues,50(4), 19–45. https://doi.org/10.1111/j.1540-4560.1994.tb01196.x Schwartz, S. H. (1999). A theory of cultural values and some implications for work. Applied Psychology,48(1), 23–47. https://doi.org/10.1111/j. 1464-0597.1999.tb00047.x Schwartz,S.H.(2005).Basichumanvalues:Anoverview.InBasic human values: Theory, methods, and applications. The Hebrew University of Jerusalem. Schwartz, S. H. (2012). An overview of the Schwartz theory of basic values. Online Readings in Psychology and Culture,2(1), 1–20. https:// doi.org/10.9707/2307-0919.1116 Schwartz, S. H. (2016a). Basic individual values: Sources and consequences. In T. Brosch, D. Sander, F. Clement, J. Deonna, E. Fehr, & P. Vuilleumier (Eds.), Handbook of value: Perspectives from economics, neuroscience, philosophy, psychology and sociology (pp. 63–84). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780198716600.003.0004 Schwartz, S. H. (2016b). Coding and analyzing PVQ-RR data (instructions for the revised portrait values questionnaire) (Unpublished). https://doi.org/ 10.13140/RG.2.2.35393.56165 Schwartz, S. H. (2017). The refined theory of basic values. In S. Roccas & L. Sagiv (Eds.), Values and behavior: Taking a cross cultural perspective (Vol. 29, pp. 51–72). Springer International Publishing. https://doi.org/ 10.1007/978-3-319-56352-7_3 Schwartz, S. H., & Bilsky, W. (1987). Toward a universal psychological structure of human values. Journal of Personality & Social Psychology, 53(3), 550–562. https://doi.org/10.1037/0022-3514.53.3.550 Schwartz, S. H., & Bilsky, W. (1990). Toward a theory of the universal content and structure of values: Extensions and cross-cultural replications. Journal of Personality & Social Psychology,58(5), 878–891. https://doi. org/10.1037/0022-3514.58.5.878 Schwartz, S. H., & Boehnke, K. (2004). Evaluating the structure of human values with confirmatory factor analysis. Journal of Research in Personality,38(3), 230–255. https://doi.org/10.1016/S0092-6566(03)00069-2 Schwartz, S. H., Cieciuch, J., Vecchione, M., Davidov, E., Fischer, R., Beierlein, C., Ramos, A., Verkasalo, M., Lönnqvist, J.-E., Demirutku, K., Dirilen-Gumus, O., & Konty, M. (2012). Refining the theory of basic individual values. Journal of Personality and Social Psychology,103(4), 663–688. https://doi.org/10.1037/a0029393 Schwartz, S. H., & Howard, J. A. (1981). A normative decision-making model of altruism. In J. P. Rushton & R. M. Sorrentino (Eds.), Altruism and helping behavior: Social, personality, and developmental perspectives (pp. 189–211). Erlbaum. Schwartz, S. H., Sagiv, L., & Boehnke, K. (2000). Worries and values. Journal of Personality,68(2), 309–346. https://doi.org/10.1111/14676494.00099 Silberer, G. (1995). Wertedynamik und Wertemarketing. In B. Tietz, R. Köhler, & J. Zentes (Eds.), Enzyklopädie der Betriebswirtschaftslehre: Bd. 4. Handwörterbuch des Marketing (2nd ed., pp. 2703–2708). Schäffer-Poeschel. Simon, H. A. (1955). A behavioral model of rational choice. The Quarterly Journal of Economics,69(1), 99. https://doi.org/10.2307/1884852 Simon, H. A. (1993). Altruism and economics. The American Economic Review,83(2), 156–161. Slovic, P. (2011). The perception of risk (reprinted.): Risk, society and policy series. Earthscan Publ. Steg, L., & de Groot, J. I. M. (2012). Environmental values. In S. D. Clayton (Ed.), Oxford library of psychology. The Oxford handbook of environmental and conservation psychology (pp. 81–92). Oxford Univ. Press. https://doi.org/10.1093/oxfordhb/9780199733026.013.0005 Stern, P. C. (2000). New environmental theories: Toward a coherent theory of environmentally significant behavior. Journal of Social Issues, 56(3), 407–424. https://doi.org/10.1111/0022-4537.00175 Sutalaksana, I. Z., Zakiyah, S. Z. Z., & Widyanti, A. (2019). Linking basic human values, risk perception, risk behavior and accident rates: The road to occupational safety. International Journal of Technology,10(5), 918. https://doi.org/10.14716/ijtech.v10i5.2165 Torelli, C. J., Özsomer, A., Carvalho, S. W., Keh, H. T., & Maehle, N. (2012). Brand concepts as representations of human values: Do cultural congruity and compatibility between values matter? Journal of Marketing, 76(4), 92–108. https://doi.org/10.1509/jm.10.0400 POIER ET AL.923 Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science,185(4157), 1124–1131. https://doi.org/10. 1126/science.185.4157.1124 Vega-Zamora, M., Parras-Rosa, M., & Torres-Ruiz, F. J. (2020). You are what you eat: The relationship between values and organic food consumption. Sustainability,12(9), 3900. https://doi.org/10.3390/ su12093900 Vernon, P. E., & Allport, G. W. (1931). A test for personal values. The Journal of Abnormal and Social Psychology,26(3), 231–248. https://doi.org/ 10.1037/h0073233 Voorn, R. J. J., van der Veen, G., van Rompay, T. J. L., Hegner, S. M., & Pruyn, A. T. H. (2021). Human values as added value(s) in consumer brand congruence: A comparison with traits and functional requirements. Journal of Brand Management,28(1), 48–59. https://doi.org/10. 1057/s41262-020-00210-w Voorn, R. J. J., van der Veen, G., van Rompay, T. J. L., & Pruyn, A. T. H. (2018). It takes time to tango: The relative importance of values versus traits in consumer brand relationships. Journal of Consumer Behaviour, 17(6), 532–541. https://doi.org/10.1002/cb.1737 Willman-Iivarinen, H. (2017). The future of consumer decision making. European Journal of Futures Research,5(1), 14. https://doi.org/10. 1007/s40309-017-0125-5 Wilson, C., & Dowlatabadi, H. (2007). Models of decision making and residential energy use. Annual Review of Environment and Resources, 32(1), 169–203. https://doi.org/10.1146/annurev.energy.32. 053006.141137 Wolff, K., Larsen, S., & Øgaard, T. (2019). How to define and measure risk perceptions. Annals of Tourism Research,79, 102759. https://doi.org/ 10.1016/j.annals.2019.102759 Wolske, K. S., Stern, P. C., & Dietz, T. (2017). Explaining interest in adopting residential solar photovoltaic systems in the United States: Toward an integration of behavioral theories. Energy Research & Social Science,25, 134–151. https://doi.org/10.1016/j.erss.2016.12.023 Xiao, C., Dunlap, R. E., & Hong, D. (2019). Ecological worldview as the central component of environmental concern: Clarifying the role of the NEP. Society & Natural Resources,32(1), 53–72. https://doi.org/10. 1080/08941920.2018.1501529 Zhang, H., van der Bij, H., & Song, M. (2020). Can cognitive biases be good for entrepreneurs? International Journal of Entrepreneurial Behavior & Research,26(4), 793–813. https://doi.org/10.1108/IJEBR-03-2019-0173 AUTHOR BIOGRAPHIES Stefan Poier is a doctoral student at the University of Gdansk. His research interest is consumer behavior with a focus on the personal dispositions of the consumer like personality traits and human values and their influence on the purchase decision in the context of energy efficiency investments of households. In addition to research, he has already gained 20 years of practical experience in advertising and marketing. In his dissertation, he focuses on the deeply rooted motivations to adopt electricity storage. Anna Maria Nikodemska-Wołowik is an associate professor at the University of Gdansk, holds a doctorate in economics, a degree of habilitated doctor and several certificates from foreign universities: Harvard Business School, IESE Business School, University of Strathclyde Business School. Professor NikodemskaWolowik focuses on the following research areas: consumer behavior, qualitative marketing research, intellectual property protection, family enterprises, corporate identity. She participates in many international academic programs and numerous research projects (including co-financed by the EU and the Polish National Science Centre). Professor Nikodemska-Wolowik has received many scientific and national awards. She authored and co-authored more than 100 publications in Poland and abroad. MichałSuchanek is an associate professor at the Faculty of Economics of the University of Gda nsk. He is the author of over 50 scientific articles including 7 JCR papers and author and editor of 9 books. In addition, he is coordinator and researcher in 10 scientific projects, including 3 Horizon 2020 projects. Suchanek is author of a research analysis for the European Commission. In addition to research, he is the owner of a consulting company with 10 years of business experience. How to cite this article: Poier, S., Nikodemska-Wołowik, A. M., & Suchanek, M. (2022). How higher-order personal values affect the purchase of electricity storage—Evidence from the German photovoltaic market. Journal of Consumer Behaviour, 21(4), 909–926. https://doi.org/10.1002/cb.2048 924 POIER ET AL. APPENDIX A TABLE A1 Assignment of the items to the different values classifications Higher order values 10 original values α19 revised values αItems Openness to change (α=.824) Self-direction .870 Self-direction thought .754 1, 23, 39 Self-direction action .810 16, 30, 56 Stimulation .510 Stimulation .510 10, 28,43 Hedonism .773 Hedonism .773 3, 36, 46 Self-enhancement* (α=.851) Achievement .741 Achievement .741 17, 32, 48 Power .863 Power dominance .764 6, 29, 41 Power resources .830 12, 20, 44 Face* .692 9, 24, 49 Conservation (α=.879) Security .821 Security-personal .668 13, 26, 53 Security-societal .777 2, 35, 50 Conformity .783 Conformity-rules .836 15, 31, 42 Conformity-interpersonal .627 4, 22, 51 Tradition** .718 Tradition .834 18, 33, 40 Humility** .584 7, 38, 54 Self-transcendence** (α=.902) Humility** .584 7, 38, 54 Universalism .870 Universalism-nature .807 8, 21, 45 Universalism-concern .778 5, 37, 52 Universalism-tolerance .758 14, 34, 57 Benevolence .879 Benevolence-care .784 11, 25, 47 Benevolence-dependability .816 19, 27, 55 Note: Table presents the assignment of PVQ-57 items to the 19 revised human values, 10 original human values and four higher order values according to the coding & analysis instructions by Schwartz (2016a,2016b). Face, as an own revised value, is part of self-enhancement but not part of one of the original values. Humility, as an own revised value, is part of the original tradition value to get closer to the original score. However, it is not part of the higher order conservation but part of the higher order self-transcendence; n=521. Abbreviation: α=Cronbach's alpha for uncentered values. POIER ET AL.925 TABLE A2 ttest for significant differences of continuous control variables Levene's test ttest for equality of means Fpt df pΔMΔSE 95% CI Lower Upper Age EV .002 .960 3.387 390 .001 4.426 1.307 1.857 6.996 UV 3.367 296.395 .001 4.426 1.315 1.839 7.014 Household size EV 4.343 .038 2.117 457 .035 .243 .115 .468 .017 UV 2.027 314.525 .043 .243 .120 .478 .007 Feed-in tariff EV 3.617 .058 2.794 381 .005 4.704 1.683 1.394 8.013 UV 2.758 283.807 .006 4.704 1.706 1.347 8.061 Average electricity price EV 2.571 .110 3.261 446 .001 1.266 .388 2.029 .503 UV 3.258 350.857 .001 1.266 .388 2.030 .502 Environmental concern (FS) EV .296 .586 1.476 458 .141 .142 .096 .047 .331 UV 1.468 356.101 .143 .142 .097 .048 .332 Perceived risk (FS) EV 4.430 .036 2.410 458 .016 .230 .095 .042 .418 UV 2.360 339.037 .019 .230 .097 .038 .422 Note: Presented are scores for mean differences of the control variables; n=460. Abbreviations: ΔSE =standard error of difference; 95% CI =95% confidence interval of the difference; EV, equal variances assumed; FS, factor score; M, mean; p, 2-tailed pvalue; SD, standard deviation; UV, equal variances not assumed. TABLE A3 Mann–Whitney test for differences of categorical control variables Ranks Test statistics nMean Sum Mann–Whitney UWilcoxon W Zp(2-tailed) Gender Non-adopters 286 222.80 63,721.00 22,680.00 63,721.0 1.862 .063 Adopters 173 241.90 41,849.00 Household income Non-adopters 265 198.68 52,649.50 17,404.50 52,649.5 3.070 .002 Adopters 159 235.54 37,450.50 Education level Non-adopters 285 217.96 62,119.00 21,364.00 62,119.0 2.550 .011 Adopters 173 248.51 42,992.00 Note: presented are scores for differences of categorical control variables; n=460. 926 POIER ET AL.