Hybrid choice model dataset of a representative Swiss online panel survey on peoples’ preferences related to mixed renewable energy scenarios in landscapes and the effect of landscape-technology fit

We present stated preference data based on a national representative Swiss online panel survey related to preference of mixed renewable energy infrastructure in landscapes. Data were collected between November 2018 and March 2019 via an online questionnaire and yielded 1026 responses. The online questionnaire consisted of two main parts – (1) questions covering meanings related to landscapes, nature and renewable energy infrastructure and questions regarding the “fit” of landscape/renewable energy infrastructure (REI) combinations and (2) a stated choice experiment. While in the first part of the questionnaire we asked respondents about their personal connection to certain landscapes, to nature and to specific REI, we also asked them to evaluate the fitting of seven different Swiss landscapes (near natural alpine areas, northern alps, touristic alpine areas, agricultural plateau, urban plateau, Jura ridges, urban alpine valley) with five different REI (wind, PV ground/agricultural, PV ground/other, PV roof, power lines) combinations. In the second part of the questionnaire, the stated choice experiment confronted respondents with 15 consecutive choice tasks, with each task involving a choice between two “energy system transformation” options and an opt-out option (none). Each choice option (beside the opt-out option) included four unlabeled attributes (landscape, wind energy infrastructure, photovoltaic energy infrastructure, high voltage overhead power line infrastructure) with varying levels. Due to data cleaning procedures (item nonresponse) the number of responses used within hybrid choice modeling and analysis was n = 844 (12,660 choice observations). An analysis of the hybrid choice model and further insights are presented in the article “How landscape-technology fit affects public evaluations of renewable energy infrastructure scenarios. A hybrid choice model.”

roof, power lines) combinations. In the second part of the questionnaire, the stated choice experiment confronted respondents with 15 consecutive choice tasks, with each task involving a choice between two "energy system transformation" options and an opt-out option (none). Each choice option (beside the opt-out option) included four unlabeled attributes (landscape, wind energy infrastructure, photovoltaic energy infrastructure, high voltage overhead power line infrastructure) with varying levels. Due to data cleaning procedures (item nonresponse) the number of responses used within hybrid choice modeling and analysis was n = 844 (12,660 choice observations). An analysis of the hybrid choice model and further insights are presented in the article "How landscape-technology fit affects public evaluations of renewable energy infrastructure scenarios. A hybrid choice model."

Value of the Data
• Presented data provide information on public preferences across different energy scenarios. They also provide a proof-of-concept for "landscape-technology fit" and contain information about predictors (landscape-and renewable energy meanings, exposure) of peoples' preferences related to landscape developments. Also, the dataset highlights the interconnectedness of landscape and energy aspects in terms of the perceived landscape quality and its potential relevance for decision making processes. • The consideration of meanings for decision making processes and policy making (not only visual aspects) could be brought into all policy areas and technical decision-making tools, even those that are not landscape-oriented. During communication and planning residents of potential energy sites could be (1) informed early on and (2) invited to participatory workshops in which the meaning of landscape and REI is addressed in addition to usual visual scenarios and (3) discussing siting alternatives. • The dataset can be used to operationalize landscape-technology fit (LTF) concept which derived from place-technology fit (PTF). In particular, this dataset may be used as a base line for future LTF model improvements in alpine regions. They contain explicit information on meanings ascribed to alpine landscapes and to specific renewable energy infrastructures.

Data Description
We conducted a representative online panel survey in Switzerland between November 2018 and March 2019 to elicit the preferences of Swiss residents for landscape oriented renewable energy infrastructure developments. The questionnaire was developed by WSL and operated by panel provider BILENDI GmbH. The survey is representative in language, age, gender, education and landscape.
The questionnaire consisted of two major parts, where within the first part questions were related (1) to meanings ascribed to landscapes and renewable energy infrastructure, (2) to aspects of landscape-technology fit and (3) to exposure of people to landscapes and renewable energy infrastructures. Within the second part a stated choice model was presented. All respondents were designated to one of seven landscapes (near natural alpine areas, northern alps, touristic alpine areas, agricultural plateau, urban plateau, jura ridges, urban alpine valley) according to the ZIP code of their origin. The landscape visualizations used in this study are illustrated in Fig. 1 , whereas further details about its joint development can be found in Spielhofer et al. [1] . All survey items and scales are presented in Table 1 , whereas the questionnaire is added to the supplementary material of the present artice. Socio demographic items and respondent ID were provided by the panel provider (items 1 to 6). After starting the survey, respondents were first asked to select landscapes that most closely represent the landscape of their living, recreation and childhood environment (variables 160-162). In a next step, respondents were asked to evaluate (randomly presented) meanings ascribed to each of the seven landscapes presented. A generalized overview of the evaluation of landscape meaning items (variables 84 to 153) is provided in Table 2 . Consequently, respondents were asked about (randomly presented) mean-      ings they ascribe to each of three renewable energy infrastructures (wind, PV ground, PV roof). A descriptive overview is provided in Table 3 (variables 57 to 83). As a consequence, people were asked to evaluate their personal feeling of the "fit" of each landscape/renewable energy infrastructure combination (variables 22 to 56). Within this landscape-technology fit evaluation photovoltaic infrastructure was separated into open space ground mounted PV and agricultural PV infrastructure. In addition, high voltage overhead power lines were integrated. For the evaluation, the landscape/energy infrastructure combination for each landscape was randomized in appearance. An exemplary illustration of the operationalized landscape-technology fit concept can be found in Fig. 2 , while an overview of respondents evaluation can be found in Table 4 . Lastly, people were asked about how they would feel if they would be exposed to renewable energy infrastructure in their living (items 154 to 156) and their recreation environment (items 157 to 159). The second part of the online panel survey consisted of a discrete choice study in which respondents faced 15 consecutive choice tasks. Respondents were asked to choose among two landscape oriented renewable energy infrastructure alternatives and one opt-out option. Each of these alternatives (beside the opt-out option) had four attributes (landscape, wind energy infrastructure, PV infrastructure, power line infrastructure). Choice design, consecutive choice tasks and choice attributes are presented in Table 5 . An exemplary choice task is illustrated in Salak et al. [2] .
For reasons of confidentiality we anonymized the data by removing all fields that would enable personal identification. The complete questionnaire, the dataset and data description are available on the Environmental Data Platform EnviDat of the Swiss Federal Institute for Forest, Snow and Landscape Research WSL ( https://doi.org/10.16904/envidat.206 ).

Experimental Design, Materials and Methods
The representative online panel survey was open for response from November 2018 to March 2019. Within this time, two reminders were sent. The survey targeted active Swiss panel members of panel operator BILENDI. In five months of operation we received a total of 1026  responses. We administered the online questionnaire with the hosting service provided by Sawtooth, while respondents were provided by panel operator BILENDI GmbH. For the layout of the questionnaire we used Sawtooth's survey software Lighthouse Studio [3] . Data cleaning due to item-nonresponse led to a total number of 844 respondents (12,660 choice observations). The questionnaire consisted of two main parts. The first part consisted of item-based questions regarding landscape and renewable energy infrastructure related aspects. The second part contained a stated choice experiment with fifteen consecutive choice tasks.

The item-based part
The first part of the questionnaire included questions regarding meanings ascribed to landscapes and renewable energy infrastructure, questions related to aspects of landscape-technology fit and questions examining the exposure of people to landscapes and renewable energy infrastructures. All items are presented in Table 1 . Item description of items regarding landscape meanings, meanings ascribed to renewable energy infrastructure and landscape-technology fit are presented in Table 2 , Table 3 and Table 4 .

The choice experiment part
The choice experiment consisted of fifteen consecutive choice tasks. Ich each choice task respondents had to choose between three alternatives. Option 1 and 2 described mixed landscape related renewable energy scenarios (action), whereas option 3 described an opt-out (noaction). Relevant attributes and credible attribute levels were developed based literature research, project meetings and workshops with the project steering group from different disciplines We identified four relevant attributes and the respective levels. The choice design was generated with Ngene software [4] and was designed as d -efficient design that varies the attribute levels in Options 1 and 2. Attribute, attribute levels and the generated choice design are presented in Table 5 . A detailed description of the attribute levels and the choice experiment can be found in the accompanying publication [2] .

Ethics Statement
The participation in the survey was operated and organized by a panel provider. Respondent participation was voluntary and respondents were informed that the data will be analyzed anonymously. Data collection and handling were implemented in accordance with the social data gathering ethics regulations of the institution conducting this research.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships which have or could be perceived to have influenced the work reported in this article.