Trust in large carnivore science in Norway

Large carnivores are controversial species, and associated conflicts between stakeholders with opposing views on large carnivores are observed across the globe. Social trust, the public’s willingness to rely on those responsible for developing policies, has gained much attention regarding the acceptance of large carnivores and large carnivore management. However, trust in large carnivore science has not received as much consideration. In Norway, administrative management authorities are responsible to execute the political framework decided by the Norwegian Parliament while basing their decisions on recommendations from large carnivore science. As large carnivore science is the main knowledge provider for monitoring and measures implemented in management decisions to achieve viable carnivore populations, trust in science is crucial. Yet, scientific information is often challenged. As attitude studies show a tendency for the wider general public to be more positive towards large carnivores than people most adversely affected, we wanted to examine whether the trust in large carnivore science follows the same pattern. We used a geographically stratified sample of 2110 respondents, five respondents from each municipality in Norway, to model how trust varies across the sample. Our results indicate that elderly men, people with lower education, those who have experienced loss of livestock to carnivores associate with lower trust in large carnivore science. Lower trust was also found among big game hunters and people who fear large carnivores. This knowledge could help to guide targeted science communication and contribute to a more comprehensive understanding of cognitions important for management of conflicts involving large carnivores.


Introduction
Large carnivores are controversial species (Dickman 2008;Lewis et al. 2017), and associated conflicts between stakeholders with opposing views on large carnivores are observed across the globe (Chapron et al. 2014;Lozano et al. 2019;Treves and Karanth 2003). Emotional responses to large carnivores range from admiration to hate (Johansson et al. 2012;Sjölander-Lindqvist et al. 2015). As general attitudes towards an object are a relevant predictor of broad behavioral patterns (Fishbein and Ajzen 1977;Fiske and Taylor 1991), knowledge about the public's attitudes towards large carnivore has been used to predict the social foundation for future conservation (Vaske and Manfredo 2012). Negative attitudes at an individual level can, for instance, lead to poaching, a frequently observed problem for the conservation of several threatened carnivore species (Gangaas et al. 2013;Liberg et al. 2012;Woodroffe and Ginsberg 1998). At a societal level, negative attitudes can induce public resistance to conservation plans and policies (Bruskotter and Fulton 2012;Sandström et al. 2015). Social-psychological theories have contributed to an understanding of behavior based partly on concepts involving general cognitions, such as attitudes, beliefs, norms, and trust (Schwartz 1992;Siegrist et al. 2000;Stern et al. 1999). Such theories are used extensively to guide research in human dimensions of wildlife, and the importance of attitudes has, for example, been highlighted as the basis for people's assessment of environmental management (Bright and Manfredo 1996;Teel and Manfredo 2010;Vaske and Manfredo 2012). The cognitive hierarchy theory (Fulton et al. 1996;Homer and Kahle 1988;Vaske and Donnelly 1999) states that behavior can be predicted, to some extent, by lower levels of cognition, but more specific levels of cognition such as attitudes and norms have proved to be better predictors (Ajzen and Fishbein 2005;Whittaker et al. 2006). Attitudes are considered rather stable positive or negative evaluations of an object, and the more specific they are in terms of the object in question, the better the attitude becomes as a predictor (Heberlein 2012). Moreover, it has been suggested that people's attitudes towards objects such as policies are influenced by the trust they have in the management agency (Vaske et al. 2007).
Trust has been claimed to have a fundamental role in conservation conflicts (Stern and Coleman 2015;Young et al. 2010Young et al. , 2016. It has been suggested that a willingness to rely on those responsible for developing policies, called social trust (Cvetkovich and Winter 2003), is essential for establishing cooperation between an agency and the public (Beierle and Konisky 2000;Cvetkovich and Winter 2004). Ultimately, the public may not support management options because they lack trust in the management agencies (Borrie and Liljeblad 2006;Cvetkovich and Winter 2004;Nyaupane et al. 2009). It has been proposed that social trust mediates the relationship between shared values and attitudes towards environmental management such as wildland fire management strategies (Vaske et al. 2007), sharpshooting programs to reduce the spread of chronic wasting disease (CWD) (Harper et al. 2015), and large carnivore management (Sponarski et al. 2014).
In addition to management agencies, scientists have been argued to play an important role in conservation conflict management Sutherland et al. 2004), for example, by providing a better understanding of the root causes of conflicts, helping to discover and test mitigation techniques, and exploring trade-offs (Bennett et al. 2017a, b;Redpath et al. 2013). Yet, scientific inquiry is constantly at risk of being politicized in environmental controversies (Sarewitz 2004), challenging the trust people hold in scientific knowledge. Scientists can be perceived as biased if they advocate positions fitting with only some of the stakeholders (Sutherland et al. 2004) or frame questions and interpret results supporting one side rather than the other (Treves et al. 2006). Furthermore, stakeholders may focus solely on findings that support their own position, contributing to an impression of science being politicalized (Thirgood and Redpath 2005). The Norwegian large carnivore management is based on a national policy defined by the Norwegian Parliament, where the government prepares specific management goals for the large carnivore species. The administrative management authorities are responsible for ensuring that the political framework is followed at both local (Municipality) and regional (County) levels. This makes the large carnivore management a mixture of centralized and decentralized processes (de Boon et al. 2020). The county governor is responsible for implementing state directives together with local carnivore committees (rovviltnemder), and local politicians will work for local democratic influence that is not necessarily in line with national administrative goals. In the midst of this complexity between local and national governance and influence, large carnivore science works as a knowledge provider giving professional recommendations.
Norwegian studies report that large carnivore science is regularly being challenged ) and cherrypicked by stakeholders to benefit their own interests (Skogen and Thrane 2007). This may cause people to lose trust in carnivore science. We believe it is important to include trust in carnivore science when trying to understand the cognitions important for management of conflicts involving large carnivores. People holding more negative attitudes towards large carnivores often claim that local knowledge is ignored by those in power, e.g., politicians, managers, biologists, and conservationists (Skogen and Krange 2003). Judgments on trust relies on heuristics (Cvetkovich and Winter 2007), including similarity between oneself and the person or organization to be trusted (Balliet and Van Lange 2013), and positive affect towards that person/organization (Schoorman et al. 2007). People tend to trust those they perceive to hold the same values as themselves (Johansson et al. 2017;Stern and Coleman 2015). We consider it urgent to better understand which variables underlie the lower trust of some people in both carnivore scientists and the management agency. This would strengthening trust.
Direct experience with wolves (Canis lupus) have been observed to associate with negative attitudes towards the species (Eriksson et al. 2015;Williams et al. 2002). People living in large carnivore areas are potentially negatively affected by large carnivores (Krange et al. 2017b), for example, by experiencing losses of domestic animals (Røskaft et al. 2007), being a hunter and thus experiencing competition for big game species (Naughton-Treves et al. 2003;Treves and Martin 2011), and/or fear of meeting these animals in the wild (Krange et al. 2017a). In addition, as the influence from what friends, peers, and enemies think can strongly affect a person's attitude (Boninger et al. 1995;Petty et al. 1997), indirect experiences (e.g., relying on other people's experiences rather than personally being exposed to carnivores or seeing tracks or signs; Eriksson et al. 2015) can be important influencing attitudes (Karlsson and Sjöström 2007). People living in areas with big game hunting traditions and high sheep densities have earlier reported more negative attitudes towards large carnivores (Gangaas et al. 2013). Further, sociodemographic variables have also been included in attitude surveys. While the association between gender and attitudes toward large carnivores have been observed to vary, with both males and females being most negative (Kleiven et al. 2004;Røskaft et al. 2007;Williams et al. 2002), higher age Dressel et al. 2015;Røskaft et al. 2007) and lower education (Dressel et al. 2015;Williams et al. 2002) have been associated with more negative attitudes towards large carnivores.
Large carnivore presence varies geographically in Norway, with some regions having all four large carnivore species, brown bear (Ursus arctos), wolf, Eurasian lynx (Lynx lynx), and wolverine (Gulo gulo) present, while others have none. To meet the two-folded objective of large carnivore management in Norway, both ensuring viable populations of large carnivores and securing a sustainable grazing industry, the Norwegian parliament have established a zonal management system Strand et al. 2019). Carnivore management zones prioritized for large carnivore species are thus separated from areas prioritized for grazing livestock. The uneven distribution of large carnivores and consequently the variation in experiencing negative consequences associated with living close to these species are therefore expected to vary geographically. To capture skepticism and negative perceptions at smaller spatial scales, without losing those voices in proportional surveys, nonproportional sampling methods are useful (Ericsson et al. 2006). Considering the important influence of geographical variation in carnivore presence and any associated negative experiences, on people's attitudes towards these animals, we wonder whether this spatial variation also influences trust in large carnivore science.
Here we wanted to examine whether variables known to influence peoples' attitudes towards large carnivores associate with trust in large carnivore science, as this could contribute to improve future science communication. In order to build trust among the public, a better understanding of the characteristics that describe people with low trust in carnivore science is important, potentially helping communicators to identify needs leading to more efficient science communication. This study aims to contribute to that understanding by exploring the association between trust in large carnivore science and variables related to carnivore presence and consequently how people perceive them. We examine whether the level of trust in carnivore science can simply be described by sociodemographic patterns and local presence of carnivores, or whether it follows a more complex pattern based on multiple factors. Therefore, we hypothesize that respondents living in areas with large carnivores show lower trust in large carnivore science, as they are more directly affected compared to respondents living far away from large carnivore species. We hypothesize that variables within the variable-group of rural context associate with lower trust in large carnivore science. Last, we hypothesize that trust in carnivore science will differ with age, gender, and education level among our respondents, and that respondents with higher levels of self-reported fear will have lower trust in large carnivore science. Ultimately, we hope that by improving the knowledge of trust in carnivore science, we can facilitate a better understanding of the cognitions important for management of conflicts involving large carnivores.

Respondents and data collection
The study included 2110 respondents (43% female, 57% male), aged between 15 and 92 years (mean age = 45.63, SD = 17.72 years), representing five respondents from each of Norway's 422 municipalities. The sampling was done by a data collection agency (www. norst at. no), and the data frame was based on existing registers that are publicly available, and respondents were sampled randomly within a municipality until five persons from that municipality had completed the survey. This geographically stratified sampling procedure allowed for the inclusion of people from all over the country, and with a mixture of different backgrounds, carnivore presence and experiences with large carnivores. The data collection agency is not required to seek permission for this kind of data collection from the Norwegian Social Science Data Service (NSD; www. nsd. no). NSD is the institution reviewing research proposals for data collection, but an ethics review and a permit are only required in cases where the researchers and/or the data collection agency retain a register of respondents for purposes such as reminders or follow up surveys. This was not the case for our study, and we have no register or any other kind of information that can be used for linking individuals to the data set.

Survey
The questionnaire used for the survey included items on trust, rural context, and sociodemographic variables. Trust in large carnivore science was measured based on how the respondent rated the statement "I trust large carnivore science produced in Norway," and this was used as the response variable in the statistical analysis. Respondents were also asked to rate statements about their trust in science in general, in climate science, and in medical science in the same format. Respondents were asked to respond to these statements using a five-point Likert scale from 1 = "strongly disagree" to 5 = "strongly agree" with 3 = "neutral".

Predictor variables
The predictor variables were collected both through the survey and from official available databases, and were divided into four groups with different themes before the analysis (Table 1). The first group included sociodemographic variables such as age, gender, and education. Age was included as a continuous variable to explore if there is a general trend of changes in trust with increasing age. The second included variables for whether carnivores in general or wolves in particular were present or not in the respondents' residential municipality. Data on large carnivore presence were obtained from Rovdata (https:// rovda ta. no/), while hunter ration and sheep density were obtained from Statistics Norway (https:// www. ssb. no/ en). Given that people may be more aware of carnivore management zones than carnivore distribution, a variable about whether the municipality of the respondent was covered by a management zone of at least one of the four carnivore species was included. In addition, a variable on whether the municipality is within the national wolf zone was also included. Lastly, this group included a question in which respondents were asked which of the four large carnivores they thought were present within their municipality, to examine whether their awareness of carnivore presence was more important than data on carnivore presence regarding their reported trust in carnivore science. The third group was named rural context and focused on respondent's experience of losses of livestock, density of free ranging sheep in the municipality, and big game hunting traditions, which according to previous research are important characteristics defining this context (Gangaas et al. 2013). In addition, we asked whether respondents were big game hunters themselves. In Norway, hunting traditions are deeply established in the country's history and culture, with big game hunting including the hunting of moose (Alces alces), roe deer (Capreolus capreolus), wild reindeer (Rangifer tarandus), and red deer (Cervus elaphus). Generally, local hunters have good access to hunting through informal connections with landowners or through organized hunting teams. The final group of factors consisted of a self-reported level of fear towards each of the four large carnivore species present in Norway (see full variable description in Table 1).

Statistical analyses
Trust in carnivore science was a 5-level categorical response variable. We analyzed how it varied in response to different combinations of the explanatory variables (Table 1) using multiple ordinal linear regression models. We also wanted to understand which of the variables in the different group best described the variation in the level of trust in large carnivore science. For example, within the group "carnivore presence, was it the presence of carnivores in general, the presence of wolves in particular or whether, or not a respondent lived within a defined carnivore zone that best described the variation in trust? The candidate models constructed and the model selection procedure reflect these considerations. Although it might be expected that some of the variables would interact, we only considered additive effects and no interaction effects due to data restrictions. The models were constructed using the MASS-library (Venables and Ripley 2013) in the statistical environment (R Core Team 2019), and we used an information theoretic approach based on AIC-values (Burnham and Anderson 2002) to objectively select the most supported model for variation in trust.

Trust in science in general and specific scientific disciplines
We found lower levels of trust in large carnivore science and climate science compared to medical science (Table 2). While approximately four out of five either agreed or highly disagreed with the statements on having trust in medical science, only 58% did so for carnivore science. Almost one out of five respondents disagreed or highly disagreed in having trust in large carnivore science.

Models of trust in carnivore science
The most supported model describing the variation in trust in large carnivore science included variables from all groups and were gender, age, education, wolf presence, big game hunter, experienced loss, and fear of wolves (Table 3). The specific effects within each variable group are described in more detail below.

Sociodemographic variables
The probability of agreeing with the statement concerning trust in large carnivore science decreased as a function of age (Fig. 1a). Gender differences were less obvious, but females showed a slightly higher probability of highly agreeing to trust in carnivore science compared to males (Fig. 1b). People who completed high school as their highest education level agreed less with the statement concerning trust in large carnivore science compared to other types of education levels. Besides that, the probability of highly agreeing with the statement concerning trust in carnivore science generally increased with education level (Fig. 1c).

Carnivore presence
All variables included in the section of carnivore presence improved the model compared to a model setup containing only sociodemographic variables (Table 3). However, the most supported model included wolf presence, where respondents living in municipalities with wolves showed lower agreement with the statement concerning trust in large carnivore science, on average (Fig. 2).

Rural context
Including some of the variables describing a rural context improved the model fit ( Table 3). Respondents that reported to be a big game hunter themselves were more likely to report lower levels of agreement in the statement concerning trust in carnivore science (Fig. 3a). The likelihood of responding agree or highly agree were higher for non-hunters than for hunters. Relatively few had experienced loss (n = 194), but they agreed less with the statement concerning  1 The predicted probability of each of the five levels of agreement to the statement of trust in large carnivore science as a function of age (a), gender (b), and education level (c). The vertical lines depict the SD for each category Fig. 2 The predicted probability of each of the five levels of agreement to the statement concerning trust in large carnivore science as a function of whether the respondent lives in a municipality where wolves were present. The vertical lines depict the SD for each category trust in large carnivore science than people who had not experienced any loss (Fig. 3b).

Fear
The results suggest that trust in large carnivore science is negatively associated with a fear of wolves, with the probability of responding agree or highly agree to the statement concerning trust in large carnivore science declining with every increase in level of fear, from not scared, little scared, pretty scared, and very scared of wolves (Fig. 4). The probability of responding disagree and highly disagree increased the most with increased fear of wolves. Fear towards the other species (e.g., wolverine, lynx, and brown bear) were also associated with mistrust in carnivore science, but to a lesser extent.

Discussion
Lower trust in large carnivore science associated with the variables wolf presence, being a big game hunter, having experienced loss of livestock to carnivores, older age, less education, being a male, and self-reported higher levels of fear of wolves. Essentially, the results show that there is no single variable alone that predicts a lower trust in carnivore Fig. 3 The predicted probability of each of the five levels of agreement to the statement concerning trust in large carnivore science as a function of being a big game hunter or not (a) and having experi-enced loss to large carnivores or not (b). The vertical lines depict the SD for each category Fig. 4 The predicted probability of each of the five levels of agreement with the statement concerning trust in large carnivore science as a function respondents' self-reported level of fear of wolves. The vertical lines depict the SD for each fear level science, but rather a combination of variables, many of which vary spatially. Compared to the other categories of science, carnivore science and climate science experience lower levels of trust than medical science. The observed lower trust in climate science, and perhaps carnivore science, may be partly explained by the different nature of the science categories included in the survey. Previous studies have suggested a trend towards lower levels of trust in what may be called impact science, e.g., science that identifies environmental and public health impacts of economic production, than in production science, e.g., science that provides new inventions for economic production (McCright et al. 2013). Denying evidence that does not conform to one's ideological worldview, values or ideology, is not uncommon, exemplified by the denial of climate change by many political conservatives in the USA (Dunlap and McCright 2015;McCright and Dunlap 2011). The presence of carnivores varies greatly between different parts of Norway, and consequently so do the potential negative consequences of living with these animals. We observed that respondents who lived in municipalities with large carnivores generally had a lower trust in carnivore science than respondents living elsewhere, with the strongest effect being found for the presence of wolves. This result is consistent with previous studies showing that people who reported personal experience of signs of wolves in the area where they lived had relatively lower social trust in the managing authorities (Johansson et al., 2012). Wolf presence has been associated with negative attitudes towards carnivores in several previous studies (Karlsson and Sjöström 2007;Krange et al. 2017a;Skogen and Krange 2003). Direct experience with wolves has been observed to correlate with negative attitudes towards both the animals and wolf policy goals (Ericsson et al. 2006;Eriksson et al. 2015). As zoning has become an important management strategy in Norway, uneven costs of coexistence will continue to be the reality. Based on our results, this will exacerbate regional variation in levels of trust. A person's experience of large carnivores may be perceived differently depending on their attitudes towards the animal. For example, some people living in carnivore areas report their experience of wolves as being unwanted, agonizing, assertive, or unnatural ). If scientific knowledge presented by large carnivore scientists is not in line with people's personal experience, for example communicating that wolves tend to be shy, people may lose trust in large carnivore scientists.
Both wolf presence and fear of wolves are included in the top model, indicating the importance of wolves in influencing people's level of trust in large carnivore science. Wolves are perceived as a flagship species, and may thus be more frequently associated with social conflicts than other large carnivore species (Douglas and Veríssimo 2013).
Researchers have sought to explain the underlying sociocultural causes of the observed controversies accompanying wolf presence Skogen and Thrane 2007). Among other observations, wolves can symbolize urban dominance over rural areas, contributing to a divide between urban and rural inhabitants (Enck and Brown 2002;Skogen and Krange 2003). Attitudes towards wolves seem in general to be more negative than those towards other large carnivores species in Norway (Dressel et al. 2015;Krange et al. 2017b), and seem to have a special position in people's awareness (Figari 2008;Figari and Skogen 2011). Krange and Skogen (2018) observed an association between lower trust in environmental institutions and negative attitudes towards wolves in Norway. Consequently, it may not be surprising that we observed wolf presence and fear of wolves to be included in the top model rather than variables relating to other species.
Negative experience, such as loss of livestock, was associated with a mistrust of carnivore science and has previously been shown to associate with negative attitudes towards carnivores (Kleiven et al. 2004;Røskaft et al. 2007). Social scientists have argued that personal importance of an issue is a key feature in studies of attitudebehavior relationships (Ajzen and Fishbein 2005;Bright and Manfredo 1996). Farmers owning free ranging sheep and big game hunters wanting to maximize big game abundance will have an interest in reducing the population sizes of large carnivores. Previous research has suggested that vested interests in a topic may motivate people to reject science communicated to them that is not compatible with their interests (Hornsey and Fielding 2017;Kunda 1987). Stakeholders that perceive carnivore science as being unbeneficial may thus be more likely to repudiate information communicated to them. Also, new relevant information that stand against currently held experiences and views may result in cognitive dissonance, making people rejecting it (Festinger 1957). Further, discomfort about the notion of the potential influence of policy preferences of scientists has been discussed in previous research (Lackey 2007). It should be noted that people tend to trust agencies that they perceive to hold similar values as themselves (Siegrist et al. 2000). On the other hand, distrust happens if the management agency is thought to have other values, or if it operates in a perceived inconsistent way with shared values for non-legitimate reasons (Cvetkovich and Winter 2003). Lack of power, e.g., if the management agency is not being perceived to control the situation, can lead to lower trust, as observed in management of threatened and endangered species (Cvetkovich and Winter 2003). While carnivore scientists are not directly responsible for management decisions, people may still have expectations to scientific knowledge being implemented into the management principles. Distinguishing between the distribution of roles and responsibilities between decision makers, managers, and scientists in relation to the complexity of the large carnivore issues may be difficult for people in general. If people fail to recognize the contributions of large carnivore scientists, they might lose trust in them and their role as knowledge providers.
Social trust plays an important part in achieving efficient nature management (Cvetkovich and Lofstedt 2013;Sponarski et al. 2014). The existence of trust allows agencies to work without the need to continuously ensure that stakeholders will act acceptably. Zajac et al. (2012) argues that raising an individual's social trust in the management agency can indirectly lead to increased stakeholders' acceptance of large carnivores. Trusted organizations do not need to continuously argue for their decisions and defend their policies and actions. We found that trust in large carnivore science is associated with very much the same variables as attitudes to large carnivores. Consequently, provision of science-based information to the public, while simultaneously building trust in information, is an approach that may reduce extreme attitudes and increase acceptance of large carnivores (Arbieu et al. 2019). This may further influence support in management decisions (Heberlein 2012;Sponarski et al. 2014).
Altogether, our results show that distrust in large carnivore science is associated with a distinct suite of characteristics. The variables describing these are available to and could be used by managers and science communicators. Our approach, while not being proportionally representative of the population, show patterns that will be useful for detecting distrust. This can in turn help guide science communication and the use of resources spent on interventions aimed at providing prerequisites for local people to lower their level of fear or increasing their trust in large carnivore science. Large carnivore scientists need to better understand how different factors, processes, and dynamics play in building mistrust in general, and we also think that both large carnivore scientists and management authorities need to clarify roles they fulfill. Closer cooperation between scientists and managers has been suggested to improve management success, but this may make it more challenging for the public to distinguish between them, and may make people doubt scientists' motivation in management recommendations. If this is true, building trust in carnivore science will rely on simultaneously building trust in carnivore management agencies. Wildlife professionals could help in reducing controversy around large carnivores by clearly delineating policy decisions from the scientific contributions used to reach to those decisions (Bruskotter 2013). Science communicators should be aware of their role, and specific local experts could be appointed to engage with media and verify information accuracy of large carnivore science. Further, ensuring the communication of science-based information on both costs and benefits of large carnivores is crucial to gain trust in large carnivore science over time.
To reduce the risk of local communities feeling neglected, citizen science, e.g., involving citizens in data collection such as monitoring, and encourage participation in management strategy development, may help to develop a trustful relationship with science (Anhalt-Depies et al. 2019;Ostermann-Miyashita et al. 2021). In groups with low levels of trust in carnivore science, e.g., hunters, citizen science could aid to make results more likely accepted as unbiased and reliable as basis for management decisions. Increased effort in including these groups in the research projects, or being invited in different discussions, may help bridging the perceived gap between scientists and hunters, as well as other stakeholder groups. Acknowledgment of different types of knowledge and knowledge sources, including local knowledge, such as personal experiences, tradition, and cultural, in addition to science, can help to improve legitimacy in management decisions . Co-creation of knowledge including both scientists and stakeholders may improve perceived legitimacy of the knowledge used to inform management decisions and consequently increase support for large carnivore management policies.
Funding Open access funding provided by Inland Norway University Of Applied Sciences. This study is funded by Inland Norway University of Applied Sciences and the Competence, University and Research Fund in Hedmark (KUF-fondet).

Availability of data and material
We will make all data available upon request. The data set will be archived for at least 10 years.
Code availability Open access.

Declarations
Ethics approval and consent to participate The survey was conducted by a data collection agency (NORSTAT; www. norst at. no) between April and June 2019. The interviews follow a strict protocol as dictated by standard research ethics. Neither the research agency nor the data collection agency are required to seek permission for this kind of data collection from the Norwegian Social Science Data Service (Ross et al. 2016). NSD is the institution reviewing research proposals for data collection, but an ethics review and a permit are only required in cases where the researchers and/or the data collection agency retain a register of respondents for purposes such as reminders or follow up surveys. This was not the case for our study, and we have no register or any other kind of information that can be used for linking individuals to the data set.
Consent for publication All authors have been contributing to this work, and all authors agree to the publication of this research.

Conflict of interest
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.