Acceptance of wind energy – Theoretical concepts , empirical drivers and some open questions

The further development of wind energy is of major importance for the success of the energy system transformation in Germany and elsewhere. This transition process is not an easy task. For example, the yearly installed capacity of wind energy onshore in Germany is declining since 2017. Only relatively few new wind turbines were 10 constructed especially in 2019. Problems are for example minimum distance requirements (e. g. residential areas, air safety), the high complexity of planning processes and local protests. Social science research has now dealt with the topic of wind energy acceptance for quite some time. On the one hand, the specific kind of acceptance (e.g. local acceptance) has been subject to scientific discourse. On the other hand, different empirical drivers (e. g. perceived distributional or procedural fairness, trust in relevant actors of the transformation process, risk/benefit perceptions, participation) have been of special 15 interest. This review deals with central theoretical concepts as well as qualitative and quantitative empirical findings of social science research concerning the acceptance of wind energy in Germany and elsewhere. Although there has been already a lot of valuable scientific work done, there are still some open questions left.

local level, higher for onshore wind and highest for offshore wind (Jones and Eiser, 2010). This finding was successfully replicated by Sonnberger and Ruddat for Germany  as well as Hübner and Hahn for three 160 regions in Germany (non-representative-sample, n=704, Hübner and Hahn, 2013). Swofford and Slattery also find evidence in favour of the proximity hypothesis on the basis of a mail survey in the USA conducted in 2009. They use a random sample of 200 residents of a wind farm with 75 wind turbines in Texas. The distance of respondents' home to the wind farm was up to 20 km. They report " […] an inverse relationship between proximity and positive attitudes, whereby acceptance of wind energy decreases closer to the wind farm […]" (Swofford andSlattery, 2010: 2514). In the same way, Langer and 165 colleagues discover a relatively high relevance of distance to place of residence for the acceptance of wind energy projects whereas " […] respondents preferred larger distances between the wind turbines and their place of residence" (Langer et al., 2017: 68).
On the other hand, Hoen and colleagues report positive effects of proximity of wind turbines on acceptance using a randomly selected sample of residents near wind turbines in the USA (n=1.705, Hoen et al., 2019: 7). Wolsink interviewed 531 170 environmentalist (members of the Wadden Vereninging) and found no effect of distance on the attitude on the siting of wind turbines in the Wadden region (Wolsink, 2007a(Wolsink, : 1199. Hübner and Pohl summarise findings of four studies with residents of wind turbines in different regions of Germany and Switzerland (212 < n < 467). Measuring distances on a metric scale (for example from less than 600 m to more than 2000 m) they found no correlation between distance to the nearest wind turbine and acceptance of wind energy in general as well a locally (Hübner and Pohl, 2015: 11). 175 Differences in results are not very surprising given the fact that the studies vary with respect to acceptance subject, acceptance object and acceptance context. Because social science research (especially in an international context) has always to deal with some degree of cultural variation, this is just natural. On the other side, this highlights the great importance of longitude research and cross-national studies. For example, Wolsink reports a U-shaped curve of wind energy acceptance as a result of experimental studies conducted in the Netherlands (pre-post-test control group design, 333 < n < 180 680). He differentiates between three phases: before project planning, during the siting process and after the wind turbines started running. Although general attitudes towards wind power as well as local acceptance of wind farms are positive on average in all three phases, they are high in the first phase, relatively low in the second phase and high again in the third phase (Wolsink, 1988;1994;2007a;b). This positive effect of direct experience with the risk source has also been found in several other studies and countries (e.g. Ireland, Scotland and USA; Krohn and Damborg, 1999;Swofford and Slattery, 185 2010;Warren et al., 2005). It can probably be traced back to overexaggerated expectations about the negative environmental impacts of the wind turbines (van der Horst and Toke, 2010; Warren et al., 2005). But it is certainly not an automatism.
Wolsink states that " […] it is by no means a guarantee for improvement of attitudes after a facility has been constructed. The effect can only be seen if the existing environmental impact is adequately dealt with, in the eyes of the local population" (Wolsink, 2007a(Wolsink, : 1199 viii . These research results help to explain at least in part the differing results with respect to the 190 proximity hypothesis: wind turbines near to residential areas can have a negative effect on acceptance in case of proposed sites but a positive effect in case of existing sites. This differentiation between proposed and existing sites is also emphasized in the literature (van der Horst, 2007;Hoen et al., 2019;Swofford and Slattery, 2010;Warren et al. 2005).
The proximity hypothesis can also be linked to the famous but meanwhile outdated NIMBY ("Not In My Backyard") phenomenon (e.g. Aitken, 2010b;Breukers and Wolsink, 2007;Devine-Wright, 2007;van der Horst, 2007;Jones and Eiser, 195 2010;Sauter and Watson, 2007). It means "[…] that people have positive attitudes towards something (wind power) until they are actually confronted with it, and that they then oppose it for selfish reasons" (Wolsink, 2007a(Wolsink, : 1199. NIMBY is problematic for at least three reasons. First, it is certainly not the only explanation for resistance (alternatives are for example place attachment or a lack procedural fairness, Jones and Eiser, 2010;Wolsink 2007a). Second, it is a very simplistic form of explanation (i.e. there are certainly more reasons for human behaviour than just selfishness, Bell et al., 2005;Devine-Wright, 200 2007;Wüstenhagen et al., 2007). Third, it is a one-sided negative label for respondents ("[…] it is never a compliment to call someone a NIMBY […]" Haggett, 2011: 504). Although the negative consequences of the NIMBY concept are clearly acknowledged here, there is one convincing argumentation by Bell and colleagues who connect NIMBY to Rational Choice Theory in order to explain the social gap: "The Nimby explanation of the social gap is the only explanation that depends upon an individual gap between attitudes to 205 wind power in general (unqualified positive) and attitudes to a particular development (negative) […] On the Nimby account, the individual gap is the gap between collective rationality (or concern for the public good) which people will express in opinion surveys when it costs them nothing and individual rationality (or self-interest) which will motivate their behaviour" (Bell et al., 2005: 465) This means "collective rationality" refers to the general support of wind energy (i.e. socio-political acceptance) while 210 "individual rationality" refers to local acceptance. This is in line with research findings referring to several distance measures of wind energy projects (local, onshore, offshore) instead of one overall measure of wind energy acceptance.

Trust
The role of trust for risk perception, risk management, (risk) acceptance and facility siting has been well researched in the last decades (e.g. Butler et al., 2011;Earle and Cvetkovich, 1995;Johnson, 1999;Renn and Levine, 1991;Slovic, 1993;215 Wüstenhagen et al., 2007). For example, the moderating effects of trust on risk and benefit perceptions are well known. If trust in relevant actors (e.g. official agencies, scientists, environmentalists, industry) is high, benefit also tends to be rated high and risks low and vice versa. This in turn has effects on risk acceptance (Siegrist, 2000;2001).
Trust can be defined as "[…] a feeling or belief that someone (or some institution) will act in your best interest" (Bellaby, 2010: 2615). But why should someone or some institution do that for me? Earle and Cvetkovic argue that trust (or social 220 trust as they call it because it is socially constructed) is based on value similarity (Earle and Cvetkovich, 1995 ix ). People sharing common values can more easily trust each other. In the modern, complex world there are many new technologies and associated risks that no one and no institution can handle alone (Renn, 2008). This is one reason for the importance of trust. Additionally, a lot of people don't know much about these technologies. In such a situation of high complexity and https://doi.org/10.5194/wes-2021-118 Preprint. Discussion started: 27 January 2022 c Author(s) 2022. CC BY 4.0 License. little knowledge, trust becomes even more important. It reduces complexity to a certain degree and creates possibilities for 225 joint action. Of course, trusting someone or some institution is a risk by itself because expectations always can be disappointed. In this case, trust is lost. In fact, it is lost very easily and regaining it is (very) difficult (Huijts et al., 2007;Kasperson et al., 2003;Luhmann, 2014;Siegrist, 2001;Slovic, 1993).
Scholars regularly cite two central elements of trust: competence and care. Competence entails the technical knowledge and capabilities to rationally manage risks. Empirical indicators can be for example education, qualification or perceived 230 performance in risk management. Care refers to the perceived responsibility to manage risks in the right way which means acting on the basis of shared cultural values. Empirical indicators can be respect of the common good or honesty (Johnson, 1999;Huijts et al., 2007;Renn and Levine, 1991;Zwick and Renn, 2002).
The concept of trust has been also used in the context of renewable energies in general and specifically wind energy. For example, Aitken found in a Scottish case study some hints for the effects of distrust on the perception of unfair processes in 235 the siting of wind energy projects. He notes that "[…] initial suspicions that the developers x would not act in the community's best interests led individuals to view decision-making processes concerning the development to be unfair.
From the earliest stages the community benefits package was perceived as representing a bribe […]" (Aitken, 2010b: 6074).
Sonnberger and Ruddat deliver mixed evidence for the role of trust. On the one hand, a multiple regression analyses revealed only two significant correlations (out of twelve possible ones). Trust in big energy companies and the acceptance of offshore 240 wind farms correlates negatively and trust in big energy companies and the local acceptance of wind farms correlates positively . On the other hand, a categorical principal component analysis with the same data showed the relevance of trust for risk perception of renewables. The analyses revealed risk-benefit/acceptance and trust/fairness as the two main latent dimensions underlying citizens' perception of the German energy system transition (Sonnberger and Ruddat, 2018) xi . Jones and Eiser report effects of trust in the target town group of their study in Sheffield: 245 "[…] the more target respondents trusted Sheffield City Council to act with due fairness and transparency when furthering their plans for wind development, the more likely they were to hold favourable attitudes towards development, and vice versa" (Jones and Eiser, 2009: 4609). Hall and colleagues conducted a qualitative case study on wind energy in Australia

Risk and benefit perceptions 255
The application of technologies always implicates benefits as well as risks (Fischhoff et al., 1981, Perlaviciute andSteg, 2014). There are no universally ideal options for the satisfaction of human needs like transportation, food, housing or energy production. For example, nuclear power provides on the one hand seemingly endless energy without the carbon emissions of fossil fuels. On the other hand, society has to deal with the catastrophic potential of a worst-case scenario (accident in a nuclear power plant) and the still unsolved disposal problem of heat generating radioactive waste. As a consequence, public 260 risk perception can be characterized by ambivalence with a strong tendency to rejection of the technology (European Commission, 2010;Gamson and Modigliani, 1989;Slovic, 1993;Zwick and Renn, 2002). Some authors speak at least of a "reluctant acceptance" of nuclear energy as a means to fight climate change (Butler et al., 2011;Corner et al., 2011).
The list of possible risks and benefits of wind energy is long. Hilary S. Boudet gives a rather comprehensive overview.
Economic development, tax revenue, landowner and/or community compensation, reduced air pollution and carbon savings 265 are examples for commonly cited benefits for utility-scale wind. Examples for commonly cited risks for utility-scale wind are ecosystem impacts, visual impacts, sound annoyance and health effects as well as impacts to property values, electricity rates, tourism and so on (Boudet, 2019). According to Bell and colleagues, the perception of these positive and negative aspects of wind energy can be related to what they call "qualified support" meaning people tend to accept wind energy not per se and unconditionally but instead only if certain conditions (i.e. an acceptable risk-benefit-ratio) are met. This is another 270 explanation for the social gap (Bell et al. 2005). Perlaviciute and Steg address collective as well as individual costs and benefits of energy applications: "People tend to ascribe high collective costs and low collective benefits to fossil fuels, including oil, coal, and gas, and to nuclear energy, whereas they tend to associate renewable energy sources with high collective benefits and low collective costs" (Perlaviciute and Steg, 2014: 363). This positive view of renewables (including wind energy) is not present on the individual level though. Irrespective of that, the relationship between costs and benefits on 275 the one side and acceptance on the other side is clear for both levels: Higher perceived costs correlate with lower acceptance and higher perceived benefits with higher acceptance (Perlaviciute and Steg, 2014: 363).
There is plenty of empirical evidence for the connection between the perception of risk/benefit and wind energy acceptance (e.g. Jones and Eiser, 2009;Walter and Gutscher, 2013;Sonnberger and Ruddat, 2017). Jones and Eiser use amongst others benefits like general economic benefit, opportunity to invest and cheaper electricity. Risks are for example the spoiling of the 280 landscape, the lowering of house prices and a general unwanted change. Results show that these benefit and risk perceptions correlate significantly with specific attitudes towards wind energy turbines (i.e. local acceptance, Jones and Eiser, 2009).
Walter and Gutscher used an experimental setting in a rural community in Bavaria (Germany) to analyse the effects of different wind energy projects on the perception an evaluation of 350 respondents to a postal survey. Projects varied amongst others with respect to the implementation/result of a citizens' vote and the existence of local benefits. They found a 285 significant effect of regional benefit (e.g. a community fund) on the support of specific wind projects (i.e. local acceptance, Walter and Gutscher, 2013). The study of Sonnberger and Ruddat revealed significant correlations of perceived risks of wind energy (index containing spoiling of the landscape, noise and danger for birds) as well as perceived benefit (creation of new jobs) and the acceptance of wind energy (offshore, onshore and local, Sonnberger and Ruddat, 2017).

Fairness and participation 290
Social scientists have repeatedly emphasized a demand for participation in case of siting decisions (Allen, 1998;Rademacher et al., 2020;Renn 2004). Residents perceive possible negative impacts of infrastructure planning (e.g. roads, power plants, waste facilities) for human health and the environment in their neighbourhood. Because the risks are solely taken by the local population while the whole society benefits from the infrastructure, questions of distributive fairness arise (Bell et al., 2005;Hall et al., 2013). A similar argument is contained in the "Green on Green conflict" which means risking the local 295 environment for the sake of the global environment (e.g. Devine-Wright, 2007;Wolsink, 2007a;Swafford and Slattery, 2010) xii . Haggett cites several studies documenting the gap between local risk and global benefit of offshore wind parks.
Although all people on earth will benefit from successfully fighting climate change, the risks (e.g. environmental damage, negative effects on birds, fishes, fishing industry and tourism) are taken by the residents of the sites (Haggett, 2011).
Another part of the puzzle is process fairness meaning the appropriate participation of residents and other stakeholders in the 300 decision-making process (Aitken, 2010b;Devine-Wright, 2007;Hall et al., 2013)  process fairness on the one hand and the acceptance of wind energy onshore and wind farms in a distance of 500m from the 305 respondent's home on the other hand (Sonnberger and Ruddat, 2017: 61). The relevance of distributive as well as process fairness with respect to the acceptance of wind turbines also shows up in the qualitative study of Hall and colleagues (Hall et al., 2013: 205).
Participation of residents and other stakeholders in siting decisions is seen as one possible way to come to commonly agreed solutions (Aitken, 2010b;Jones and Eiser, 2009;Klinke and Renn, 2002;Ruddat and Renn, 2012;Wolsink, 2007a). 310 Although there is certainly no guarantee for success, "good participation" raises the chances to avoid or minimize conflict (Alcántara et al., 2016;Webler, 1995;Renn, 2004;Ruddat and Mayer, 2020;Schweizer-Ries et al., 2010). Like Breukers and Wolsink put it: "Participatory decision-making is unlikely to turn people who fundamentally oppose wind power into supporters. However, conditional supporters […] may accept a wind project when they have been given an opportunity to influence the design" (Breukers and Wolsink, 2007Wolsink, : 2738. 315 People can participate directly in the planning process or financially. Krohn and Damborg report empirical evidence for the positive effect of financial participation on acceptance (Krohn and Damborg, 1999). Pasqualetti also emphasizes the benefit for land owners in the USA through wind turbines on their property (Pasqualetti, 2001). In a mixed-method design using surveys, qualitative interviews, focus groups and workshops, Schweizer-Ries and colleagues examined the perception and evaluation of wind, solar and biomass energy in different German case studies and came to the conclusion that financial as 320 well as planning participation can have positive effects on (local) acceptance (Schweizer-Ries et al., 2010). Hübner and colleagues arrive at the same conclusion (Hübner et al., 2020).
With respect to planning participation, Firestone et al. report a positive correlation between the perceived possibility of the community to influence the outcome of the siting process and the attitude towards the respective wind project (Firestone et al., 2018: 377). In the study of Langer and colleagues, participation was under the three attributes with the highest average 325 relative importance values with respect to the acceptance of local wind energy projects (Langer et al., 2017: 68). Based on empirical evidence from several studies, Haggett asserts that "[…] opposition can be both because people perceive that they have no voice, or no power" and concludes that "the planning process for offshore projects should therefore ideally allow local people to have some say or even influence in the project […]" (Haggett, 2011: 507).
tween proposed and existing wind farms. But this would just be a starting point on the road to a more comprehensive theoretical concept or framework. Examples for such integrative theoretical frameworks are the Elaboration Likelihood 390 Modell (ELM, Petty and Cacioppo, 1986;Petty and Wegener, 1999) or the Social Amplification of Risk Framework (SARF, Kasperson et al., 1988;Kasperson et al., 2003;Renn et al., 1992). Additionally, it could be asked how proximity can be operationalized in the right way since the relationships between the distance of wind farms and the local acceptance of residents seem to vary with scales. If distances are measured on a metric scale, there is no relationship. If distances are measured using ordinal scales (e. g. 500 m, 5 km, onshore, offshore), relationships show up as expected. It is probably not 395 just the physical distance alone that constitutes opposition or support but the meaning of the different distances for the residents (i.e. social construction of distance, Devine-Wright 2005). What do they perceive to be their neighbourhood? How big is it really? This may vary between social groups as well as between different cultures. Taken together, this review has shown that despite a lot of valuable scientific work done until now, there are still some open questions left.

Sonnberger and Ruddat 2016:
All in all, how acceptable would you consider a solar farm / a wind farm / a high-tension power line at ca. 500m distance from your home?
vii "The visual impact of a wind energy landscape is indeed important, but this impact will fluctuate greatly across unique locations and societies. Levels of environmental concern will surely differ by location and will depend greatly on local context and place attachment" (Swofford andSlattery, 2010: 2514).
viii This can be linked to the role of visual effects and place attachment (see 3.1).
ix "Throughout its development, social trust was based on similarity of cultural values, and this was communicated within cultural groups by narratives constructed by community leaders. Social trust was socially based" (Earle and Cvetkovich, 1995: 19).
x According to Aitken, "[…] the developers are one of the largest energy companies in the UK" (Aitken, 2010b: 6070).
xi These different results can at least partly be ascribed to the different methods of analyses.
xii Here again the argument of "qualified support" by Bell et al. plays a crucial role. If residents perceive a high distributional fairness, acceptance of nearby wind farms is more likely. The same is true if the "price" for the local environment is not too high.
xiii Some authors (e.g. Firestone et al., 2018, Hall et al., 2013 use the terms "distributional / distributive justice" and "procedural justice". Although there may be some differences between these formulations, they are used here interchangeable. xiv I am totally aware that this may be a very critical topic for some researchers. Unfortunately, this cannot be discussed here in further detail.