Today, river corrections aiming at better flood protection must consider ecological aspects such as “naturalness” and biodiversity. Gaining acceptance among local residents for these projects is important, since they impact local infrastructure and alter the familiar landscape. The question addressed in this paper is whether there are differences between local residents regarding the question of whether they think a river restoration project at a section of the Swiss Thur River was reasonable. We also investigate whether there are differences regarding the reasons for this evaluation, such as improved flood protection, higher perceived naturalness, increased biodiversity, and aesthetics. Results show that for farmers flood protection and naturalness are more important factors than for others and that there are differences among the local villages.
1. From Flood Protection by River Corrections to River Restoration With Flood Protection
 Through a wider set of human activities (building of dams; interbasin transfers; locks, stream channelization, leveeing, human settlement), aquatic ecosystems have been altered on a global scale and often lost some of their function and related services for society [Döll et al., 2009; Vörösmarty et al., 2004]. For instance, channelization, lock, and dam systems that enable transportation and provide flood protection for settlements and farmland intensified in the early 20th century in the United States (US). The efforts during this period followed an official doctrine and a rational approach based on technical analysis [Committee on Restoration of Aquatic Ecosystems: Science Technology and Public Policy et al., 1992; chapter 2]. During the second half of the 20th century, this regime changed, giving way to higher valuation of aquatic ecosystems. The first endeavors to restore these ecosystems emerged toward the end of the century (cf. the example of the Kissimmee River, Florida: Koebel ). Increasingly, public support for restored and more “natural” rivers was sought and helped change the old water management regime. This development is evident globally although the context for river restoration may differ nationally and locally. For instance, in the US and Europe, restoration projects are mainly policy driven with top-down planning even if the current participatory approaches to restoration including social learning are state of the art [Mostert, 2006; Pahl-Wostl, 2006]. In contrast, in Japan, nongovernmental organizations (NGOs), in particular, trigger major restoration activities [Nakamura et al., 2006].
 In Switzerland, since the early 19th century, structural flood protection and river correction have been used to protect society and living spaces from the destructive forces of floods [Vischer, 2003]. These interventions increased flood protection (at least in the short term), but had significant impacts on ecosystems and the habitats of many animals and plants. For example, the ecologically very valuable wetlands largely disappeared [FOEN, 2012]. Meeting the need for arable land and protecting land and related land uses from floods has been the main driver for these kinds of “corrections.” In contrast, in the last few decades, more and more water bodies have been modified to restore the natural functions of the landscapes. At the same time, these projects should guarantee flood protection by giving more space to the rivers and thus enhancing their water discharge capacity. In other words, today flood protection should emphasize ecological aspects. However, communication regarding a specific project can set the frame by stressing either flood protection or other (e.g., ecological or aesthetic) aspects.
 Since these projects are often expensive, a monitoring process for measuring success is recommended, but seldom fully applied [Wohl et al., 2005; Woolsey et al., 2007]. Often, restoration projects are considered successful if post-hoc positive public opinion has emerged, an important dimension of success, but not the only one [Palmer et al., 2007]. General restoration strategies are seldom available—be it on the national level [Committee on Restoration of Aquatic Ecosystems: Science Technology and Public Policy et al., 1992, chapter 8], in a single US state [Kondolf et al., 2007], or a Swiss canton [Bratrich, 2004]—for steering and monitoring restoration projects, because standards for ecologically but also socially successful restorations are lacking [Palmer et al., 2005].
 Moreover, the alteration of the familiar landscape and the potential influence of accessibility and usability for residents often make river restoration projects a contested issue, and social acceptance of such projects in the planning phase is hard to gain [Wolsink, 2010]. Furthermore, residents generally want to be involved in river restoration projects [Tunstall et al., 1999].
 We are convinced that successfully planning participation processes requires a preferably detailed view on the current perception and evaluation of the project at hand. A participatory strategy for gaining public acceptance of the project might be beneficial if the strategy involves different subgroups or specific situations [Le Lay et al., 2013]. Of course, a retrospective study such as ours can help improve future projects just by informing whether different groups were equally included and are content with the result and by showing which aspects may play a role. That is, we assume the public is not a homogeneous group. Even if general social acceptance could be achieved after the project's end, this does not mean that everybody agrees on the reasonability of the restoration for the same reasons and that the same strategy would be successful in a new project in a different context [House and Fordham, 1997]. There might be differences within the population, for instance, regarding the question “to what” a river should be restored [Bain et al., 2011; Tapsell, 1995]. Expectations of user groups and thus judgments after restoration can differ. Therefore, social science investigations as a complement of engineering and natural sciences research are increasingly seen as important [Braden et al., 2009].
 Former studies [e.g., Buijs, 2009] showed that residents living close to a river generally have a positive attitude regarding its restoration: The average perceived scenic beauty, for instance, is higher for a restored river. Further important categories for residents are attachment to the river, biodiversity, functionality, and risk [Jacobs and Buijs, 2011]. Connelly et al.  used the theory of planned behavior to better understand the relationships between environmental beliefs and support for ecosystem restoration in the Hudson River estuary, New York State, US. They showed that support for general restoration plans was higher than support for specific implementation actions. In particular, concern for the environment was positively and concern for human interests was negatively related to support for conservation actions.
Tunstall et al.  suggested that a restoration has a positive influence on perceived aesthetics. Likewise, Junker and Buchecker  found in their study that natural conditions are perceived as more aesthetic. They concluded that the aesthetic outcomes of minor restoration efforts could nevertheless be highly appreciated by the public. However, groups of residents or residents in spatially different proximities to the project area may differ regarding their relation to the river/riparian area and thus regarding the perception of different aspects of a restoration. For instance, affected farmers mainly focus on economic aspects, since their land/land use is directly threatened by restoration as much as it is by floods [Buijs, 2009]. Thus, farmers' perspective may be different from that of other groups.
 The basic idea of this paper is to assess the degree of acceptance of restoration projects. We examine attitudes toward nature in the current study regarding the relationship between environmental attitudes among respondents and the perception of riparian qualities and acceptance. Moreover, we address potential differences between (groups of) participants and between villages. Specifically, we assume farmers and nonfarmers differ regarding degree of acceptance and perception of flood protection. We also assumed that participants from those villages close to the river might report higher concern about floods and rate the importance of flood protection as higher than other villages.
2. Material and Methods
2.1. Two Corrections of the River Thur in Switzerland
 Over the past two centuries, various projects have been conducted to solve flooding issues in the region of Thurtal (literally Thur Valley). Between 1874 and 1893, the “1. Thurkorrektur” (first correction of the Thur) was implemented, in which the river was straightened and put into a different profile [Zaugg, 2002]. The levees were not high or strong enough; consequently, two extreme floods in 1977 and 1978 caused significant damage to the region. The affected canton of Thurgau decided to intervene to better protect the people and infrastructure surrounding the Thur. After considerable discussions, debates, investigations, and evaluations, a plan was presented addressing the shortcomings (“2. Thurkorrektur”). Instead of just building higher levees, the plan was to combine structural measures with restoration to enhance flood protection, restore ecological functions, and reduce erosion of the riverbed. The plan was realized from 1993 to 2002. Levees were built higher, and stronger and additional space was provided to the river. The second correction thus exemplifies a trend toward a “more integrated and ecosystem-based flood management philosophy” [Hill, 2013, p. 135].
 At the time of the second correction, it was not easy to gain acceptance for Swiss restoration projects, and conflicts among the public and the cantonal governments were apparent [Zaugg, 2002, 2006]. Strong opposition in the beginning, especially from landowners and land users, was observed [Junker et al., 2007]. Following these difficulties, a participatory approach was designed for the Thur restoration [Junker and Buchecker, 2006]. The literature provides reasons participation is important for the success of this project [Mostert, 2006]. For instance, conflicts among project leaders, stakeholders, and the public are more likely to be avoided, the public's ideas and requirements are respected, and long-term acceptance and identification with the project are more likely to be guaranteed [Junker and Buchecker, 2008; Junker et al., 2007]. In general, public engagement leads to better decisions [Petts, 2006].
 Restoration along the specific section of the Thur (the focus of this paper) took place from 1993 to 2003, and the large widening at the section was conducted between 2000 and 2003. During this later phase, a research project that investigated the social and participatory processes accompanied the restoration process [Junker and Buchecker, 2006]. Flood prevention responding to the devastating floods in the late 1970s was the project leaders' main “selling point” of the restoration plans for the second Thur restoration in the beginning, which neglected ecological aspects, however. Only later were ecological “restoration” aspects involved due to pressure from the Federal Agency for Water Management and the Swiss Agency for the Environment, Forests and Landscape. After the plans were revised, the main goals of the second Thur correction were protecting against flooding events, restoring ecological functions (increased “naturalness”), and reducing erosion on the river grounds.
Meier-Dallach and Walter  investigated the cultural importance of the Thur for the residents of the surrounding villages (two are included in the current study). The authors found that the most important aspects for the residents were a multipurpose and free river channel, high ecological diversity, followed by a broader space where the Thur can spread out, thus enhancing protection from floods. Ten years after the project was completed, no scientific study has dealt with the local residents' perception, while the effects of the restoration have become apparent. We investigated how local residents rate the conditions before and after the restoration at the “Schäffäuli,” as the river section is termed by local inhabitants (see Figure 2), whether differences between participants with different professions occur, and how the judgments in four villages near the river differ. The dependent variable was whether the restoration is generally accepted by individuals or, in other words, whether these individuals agreed that “the restoration was reasonable.” The research described in this paper was part of a transdisciplinary case study in which scientists, graduate students in the Department for Environmental Systems Science, and representatives of the villages collaborated closely [Ghisla et al., 2012; Rösch et al., 2013]. This transdisciplinary process facilitated the appropriate focus of the study.
2.2. Questionnaire Study: Four Different Villages Showing Different Proximity to the River
 To investigate the opinions of local residents about the restoration project at the Thur and especially the question of whether the restoration was perceived as reasonable and accepted (representing the dependent variable), we conducted a questionnaire study in villages at varying proximities to the restored river section: Altikon, Niederneunforn, Oberneunforn, and Thalheim (see Figure 1). We sent the questionnaire as a postal survey to inhabitants in the four villages. The addresses were received directly by the population registries of the municipalities, as we had received consent from the respective mayors. The questionnaire comprised questions covering perception of ecological diversity, flood protection, and aesthetic aspects, as well as personal attitude toward nature (see the section below for more details). Participants had to indicate the degree of agreement on a Likert scale (ranging from 1 = “not at all” to 7 = “absolutely”). These topics were discussed with representatives of the villages before the questionnaire was finalized to make sure that all relevant aspects were addressed. It was also important that representatives from all villages were happy with the content, as they also initiated access to the municipalities' registry of residences. The cover letter of the questionnaire also carried the logos of the municipalities in the bottom line to stress the “official” acknowledgement of the study to improve the response rate among residents.
 In this paper, we compare the perception of the condition of the river area before and after the restoration. Moreover, we asked participants to rate the perceived importance of each issue. To illustrate the different conditions, two photographs were used, showing the condition before and after the restoration from the same viewpoint (see Figure 2). This also helped individuals who were unfamiliar with the “before” condition to imagine the change. The restoration took place downstream from the bridge, which is more visible in the picture on the right. One can also see the newly emerged gravel banks as well as the loss of land and forest on the left bank (right picture).
 From the addresses provided as random samples by the four municipalities from the residence registry, 1000 adults (around a third of the population) were randomly chosen to receive the questionnaire. Of these questionnaires, 408 (41%) were returned; 398 questionnaires were usable for the analysis. The response rates for the villages are shown in Table 1.
Table 1. Sent and Returned Questionnaires Including the Achieved Return Rate From the Different Villagesa
In brackets the number of farmers in each sample.
Return rate (%)
 The sample consisted of 197 (50%) women and 196 (50%) men plus five people who did not indicate their gender. The bottom rows in Table 1 show that the gender ratio is also balanced within each village. The age of the participants ranged from a minimum of 15 years to a maximum of 94 years with a mean of M = 51.8 (SD = 16.23) and was also balanced across the villages.
 As indicated by local village representatives and by comments in the open section of the questionnaire, individuals from outside the region also appreciate the river section. Some individuals perceive the nonresidents' presence as somewhat concerning. Especially in the daytime during summer, the parking area is full of cars, while dog walkers, joggers, and bikers flock to the site. As shown by the ratings of the item “I feel personally connected to the Thur,” respondents feel considerably connected to the river (M = 5.9, SD = 1.36). Importantly, there are no significant differences between the resident groups or villages regarding this connectedness.
 Roughly 30% of the participants noted down a personal comment. The comments varied from very positive to strongly negative opinions about the project. The most important category was naturalness: Many participants hope that through restorations a reapproach to nature and the rich flora and fauna can be accomplished. Restorations were thought to be very welcome and the wish for further restorations was expressed.
2.4. Dependent Variable
 The dependent variable measures acceptance (as acknowledgement) of the restoration project and is represented by the item “The restoration of the Schäffäuli was reasonable” (N = 385, M = 5.3, SD = 1.73).
2.5. Attitude Measures
 We developed items comprising a block of six statements that measured the general attitudes toward nature and economic development and the right of humans to alter nature (see Table 2). All items had to be rated on a seven-point scale (from 1 = “not at all” to 7 = “absolutely”). There is a striking difference between the mean values for the first four items (ranging from M = 4.3 to 5.4) and the two last items (M = 3.2 and 3.4). In contrast, the standard deviations are comparable. These six items were fed into a factor analysis (alpha factoring using Oblimin rotation), which revealed two factors. The first factor is denoted by nature conservation, because the items represent the value of nature in itself and the view that human influence has to be minimized. The second factor comprises one item on economic progress and one on the right to alter nature for humans' sake; this factor is denoted as anthropocentrism.
Table 2. Six Items Addressing the Attitudes Toward Nature and Economic Aspectsa
Statistics and results of the factor analysis are shown (factor loadings below 0.35 omitted, Costello and Osborne ). “°” indicates items used to calculate the scale nature conservation.
Rotation converged in six iterations. Extraction Method: Alpha Factoring. Rotation Method: Oblimin with Kaiser Normalization.
°Nature must be protected from human influences as completely as possible
°Nature is more beautiful than anything created by man
°The wilder nature is, the better I like it
For me, nature serves primarily for recreation
Nature must not obstruct economic development
Humans have the right to change nature for their own benefit
 For a more convenient further analysis, we computed one scale (labeled nature conservation) of the items clearly loading factor one (in Table 2 indicated by °). The scale comprises the mean of the individual item ratings (scale mean value = 4.9, SD = 1.33). Cronbach's α indicates a good consistency value of 0.73. The other items that could not clearly be assigned to a specific factor (note the structure matrix) were used as separate items.
 Furthermore, we introduced an item on perceived flood risk (“I feel endangered by floods,” M = 1.8, SD = 1.45) to control for differences in risk perception. As the low mean value indicates, participants currently do not feel much at risk from floods.
2.6. Riparian Qualities Before and After Restoration
 Another block of variables—we developed our own and adapted items from Buijs —represent statements on features of the river/the riparian area “before” and “after” the restoration: naturalness, ecological diversity, flood protection, and aesthetics, including the importance of each respective aspect (see Table 3). Respondents were asked to judge these statements. Table 3 shows that the conditions (e.g., naturalness, flood protection) after restoration are rated higher than those before. All aspects are rated as highly important (mean values between 5.2 and 5.9).
Table 3. Statistics of the Questionnaire Answers to the Respective Itemsa
The mean values and standard deviations are referring to a scale from 1 to 7, where 1 = “I do not agree at all” and 7 = “I totally agree.”
I perceive the condition “before” (Figure 2 left) as natural
I perceive the condition “after” (Figure 2 right) as natural
A natural condition is important to me
I perceive the animal and plant diversity in the condition “before” as high
I perceive the animal and plant diversity in the condition “after” as high
The animal and plant diversity is important to me
In the condition “before,” I perceive the flood protection as good
In the condition “after,” I perceive the flood protection as good
Flood protection is important to me
I perceive the state “before” as beautiful
I perceive the condition “after” as beautiful
The scenic beauty of the Thur section is important to me
 The section of the Thur we investigated is used mainly for recreational walking, picnicking, and sunbathing during the summer months. Therefore, accessibility and usability are important aspects, and respondents had to rate two items: “The restored river situation restricts my personal use (walking, jogging, fishing in the river, etc.)” and “I perceive the restored river section as inaccessible.” An additional item asked whether respondents use the Thur section more frequently than before the restoration took place. People rated the two items on accessibility and use rather low: “The restored river situation restricts my personal use” (N = 383; M = 2.1; SD = 1.60) and “I perceive the restored river section as inaccessible” (N = 379; M = 2.7; SD = 1.91). Most people therefore do not think that the restoration restricts their use pattern. Participants, however, did not indicate they use the area more frequently for recreation than before (N = 381; M = 3.3; SD = 1.99).
 We used different analyses to assess the sample. We used correlation analysis to assess the relation between attitudes and the ratings of the conditions after versus before. We used stepwise linear regression analyses to measure the respective contributions of the latter items explaining the dependent variable (how reasonable was the restoration of the Schäffäuli?). We applied the stepwise function of SPSS 19, which selects from all included variables those with the highest β contribution to end up with the best subset of variables to predict as well as possible with as few variables as possible.
 Thus, we conducted regression analyses for each of the four villages separately. The regression analyses revealed that in fact not all aspects are equally important for all villages; that is, the regression models differ. The same strategy was used to compare farmers and nonfarmers. We expected that different aspects contribute significantly to explaining the dependent variable for farmers versus nonfarmers as well as for the different villages. In addition to farmers, individuals in various other professions participated in the study (see Table 4). In the preliminary analyses of the case study, the students involved noted a different response pattern for farmers compared to that of participants in other professions (other profession groups differ only marginally). Particularly, 12 farmers (25%) reported that they had been damaged by previous floods, whereas only 4% of the remaining sample did.
Table 4. List of Participants' Professions (Besides Farmers)
3.1. Attitudes and Riparian Qualities
 In this section, we report on the correlation between attitudes and riparian qualities (values after minus before restoration, because mean values after are higher than those for conditions before; this procedure thus measures improved riparian conditions; see Table 3, for instance Mnaturalness after (5.4) − Mnaturalness before (2.5) = 2.9). As can be seen in Table 5, the nature conservation scale correlates positively with all items except with the one stating flood protection is important. The item on the recreational purpose shows the same pattern, but with lower r values.
Table 5. Shown Are the Results of the Correlation Analysis Between the Items and the Scale of the First Block With the Difference Between the Values for the Conditions Before and After and the Related Importance of Each Issuea
Nature Must Not Obstruct Economic Development (N = 374–387)
Humans Have the Right to Change Nature for Their Own Benefit (N = 374–382)
For Me, Nature Serves Primarily for Recreation (N = 372–384)
Nature Conservation (N = 375–390)
The last row shows the correlations with the dependent variable.
Correlation is significant at the 0.05 level (1-tailed).
 The items “Humans have the right to change nature for their own benefit” and “Nature must not obstruct economic development” show the opposite correlation pattern. This suggests that these notions are more in line with the condition before restoration and negatively related to the resulting condition. “Flood protection is important” is significantly correlated with the item on economic development. The correlation of the items with the dependent variable is shown in the last row. It is highly positive for the nature conservation scale and the recreation item but negative for the other items. Moreover, age is correlated significantly positively with economic development (r = 0.21, p < 0.01) but negatively with nature conservation (r = −0.10, p < 0.05).
3.2. Comparison of Farmers and Nonfarmers
 As indicated in section 2.6, we focused on farmers (N = 49) and compared this user group with nonfarmers (N = 349). Regarding acceptance of the restoration, the analysis of variance (ANOVA) shows lower mean values for farmers (nonfarmers: M = 5.4, SD = 1.58; farmers: M = 4.2, SD = 2.25; ANOVA: F(1, 383) = 23.9; p < 0.001). Farmers are less convinced (and in a less homogeneous way; note the higher standard deviation of farmers' attitude to the dependent variable) than nonfarmers that the restoration of the Thur section was reasonable. Noteworthy, farmers show a bimodal U-shaped distribution regarding their acceptance ratings (skewness = −0.197), meaning that some question the reasonability of the restoration. Nonfarmers, in contrast, show a clearly negatively skewed distribution of ratings, thus demonstrating clear support (skewness = −1.18). Farmers also differ significantly from nonfarmers in their attitude regarding the perceived threat of flooding (nonfarmers: M = 1.7, SD = 1.29; farmers: M = 2.4, SD = 2.17; ANOVA: F(1, 388) = 9.5; p = 0.002).
 We also conducted two stepwise linear regression analyses to compare the groups' ratings regarding the change in conditions, that is, the difference in values before and after the restoration (see also Table 5). The items on the general importance of each issue were also included.
 The regression model (see Table 6) for farmers comprised five steps, revealing three significantly contributing items, namely, improved flood protection, the change in the natural condition, and (with a smaller beta value) “scenic beauty is important.” The model for nonfarmers also comprised five steps but yielded five significant items with improved scenic beauty explaining most of the variance (β = 0.33). The total amount of the explained variance is higher for farmers (70%) compared to nonfarmers (50%), indicating that for farmers major points that make them accept the restoration have been included in the regression model.
Table 6. Two Stepwise Linear Regression Analyses for Farmers and Nonfarmers in the Samplea
Farmers (N = 40) R2 = 0.703
Nonfarmers (N = 309) R2 = 0.487
Shown are results of the final models. We included all items in the table, even those that turned out not to contribute significantly.
Condition is natural (after-before)
Natural condition is important
Animal and plant diversity is high (after-before)
Animal and plant diversity is important
Flood protection is good (after-before)
Flood protection is important
Scenic beauty is high (after-before)
Scenic beauty is important
 The items “Flood protection is important” and “Animal and plant diversity is high (after-before)” did not make a significant contribution to the model.
 There is the question whether farmers are concerned more with flood protection and if that may be because of a higher risk perception (they are more personally affected). Indeed, farmers perceive higher flood risks, answering the item “I feel endangered by floods” significantly (F(1, 388) = 9.57, p < 0.05) more pronounced (M = 2.4, SD = 2.17) than nonfarmers (M = 1.7, SD = 1.29).
3.3. Comparison of the Four Villages
 Regarding the dependent variable, all villages agree on average that the restoration of Schäffäuli was reasonable (M = 5.1–5.4). The differences are minimal and not significant (ANOVA: F(1, 3) = 0.84; p = 0.475). However, even if there are no significant differences between the villages regarding acceptance of this restoration project, there might be differences regarding how specific aspects of the perceived riparian conditions relate to acceptance.
 We conducted a stepwise linear regression analysis using the stepwise function in SPSS for each village. The number of steps applied by this function differed: three for Altikon and Niederneunforn but two for Oberneunforn and four for Thalheim. Table 7 shows the contributions of the change (represented by items calculating the condition “after” minus the condition “before,” respectively). The one item on aesthetics is influential for all villages (scenic beauty is high). This item also is the most important for explaining the variance in the perceived reasonability of the restoration, followed by the importance of a natural condition. The importance item (denoting “naturalness is important” in general) is significant only for the villages Niederneunforn and Oberneunforn, whereas the actual perceived change in naturalness is significant for Altikon and Thalheim.
Table 7. Results From the Stepwise Linear Regression, Dependent Variable “The Restoration at the Schäffäuli Was Reasonable”a
Altikon (N = 95) R2 = 0.456
Nieder neunforn (N = 76) R2 = 0.646
Ober neunforn (N = 61) R2 = 0.622
Thalheim (N = 114) R2 = 0.538
We included all items in the table, even those that turned out not to contribute significantly. For the whole sample (N = 352), R2 = 0.553.
Condition is natural (after-before)
Natural condition is important
Animal and plant diversity is high (after-before)
Animal and plant diversity is important
Flood protection is good (after-before)
Flood protection is important
Scenic beauty is high (after-before)
Scenic beauty is important
Farmer or not
 We included a binary item denoting if the respondent was a farmer or not to control for specific effects. In no village did the contribution of this item reach significance. Thus, this item was excluded from the regression model by the stepwise function.
 In this study, we investigated the degree of and the reasons for acceptance of a restoration project at the River Thur in Switzerland. Residents of four villages in the respective area responded to a questionnaire that included items on general attitudes toward nature conservation and anthropocentric interests as well as items assessing participants' perception of the area's riparian qualities before and after the project.
 Results of the correlation analysis, focusing on the relation of participants' attitudes and responses to the before and after conditions, show a clear pattern. A nature conservation orientation is positively connected to the restoration outcomes whereas human-related concerns are positively related to the status quo ante. This pattern is in line with results of studies by de Groot and de Groot  and Connelly et al. . The items “Humans have the right to change nature for their own benefit” and “Nature must not obstruct economic development” are in turn negatively related to conditions after restoration. This suggests that individuals rating high on these human-concerned attitudes accept a canalized river more readily, whereas increased naturalness due to restoration is less appealing. Moreover, there is a hint at a potential trade-off between nature conservation and flood protection measures that supposedly impact naturalness. Obviously, the changed condition is more in line with the idea of conservation of natural environments whereas the importance of flood protection and ecological conditions “before” are more related to anthropocentric (economic) aspects.
 The older the participants are, the more positive they are about the idea that economic aspects are more important than environmental aspects and vice versa. There is a negative connection to a nature conservation attitude but a noncorrelation with acceptance. This hints at changes in the perception of riparian qualities and thus the purpose of river management. We consider this an interesting research area: Can we observe parallel development from a technocratic to a more ecosystem-oriented river management among specialist experts and the broader public? Few studies have investigated these questions. For instance, Schaich  found more support among older residents for grazing practices in Luxembourg floodplains and differences between “lifestyle groups”: A modern lifestyle group showed higher acceptance of restoration efforts compared to the mixed and the premodern lifestyle groups. However, Buijs , for instance, found no significant effect of age regarding adherence to different frames that provide meaning to a restoration of floodplains in the Netherlands. In addition, other studies, dealing with the economic valuation of restorations, found only insignificant relationships with age and other demographic variables [Loomis, 1996; Loomis et al., 2000]. In our study, age and acceptance of restoration are likewise not significantly correlated.
 Explaining the dependent variable acceptance of the restoration (i.e., the statement that the restoration at the Schäffäuli was reasonable) was the main task of the second part of this questionnaire. The majority of participants evaluated the conditions after restoration as superior to the conditions before. Thus, improved riparian qualities by the restoration can explain the acceptance by the participants to a considerable extent. However, participants did not indicate that they use the river section more often after restoration. This could be because they used it to a high degree before. Even in the less natural and aesthetically valuable environment of the channelized river, local and nonlocal inhabitants used the area for recreation. Given the river's proximity to their homes, this is of course not surprising. Whether residents from non-neighboring villages spend more time after the restoration along the river, however, cannot be answered by our survey.
 As assumed, there are differences between user groups (specifically farmers versus nonfarmers) and between the four villages. Farmers have a specific relationship to “nature” because they directly deal with soil and plants and their financial well-being depends on the yield of their land. Thus, it is not surprising that farmers differ from nonfarmers in their judgment about acceptance (i.e., they agree to a lower degree that the restoration was reasonable). Farmers show significantly higher values for the perception of threat from flooding. Thus, as hypothesized, farmers actually focus more on flood protection and are more concerned about the flood threats. We assume that not only are flood risks more related to the work of farmers (land use and potential expected land losses) but also that in turn normal residents perceive the nearby countryside and the proximity of the river are decisive advantages for living quality that partly outweigh environmental risks [House and Fordham, 1997]. In fact, nonfarmers acknowledge flood protection and the importance of animal and plant diversity after restoration, but overall, they focus much more on aesthetic aspects.
 Participants from all villages on average overwhelmingly agree that the restoration was reasonable. They partly differ regarding the underlying rationales. Compared to Meier-Dallach and Walter's  results, for example, our study confirms the importance of ecological diversity and adds the importance of aesthetic aspects and naturalness (shown, for instance, by Junker and Buchecker ). We also show differences between the surrounding villages regarding the factors perceived as most important for evaluating the restoration as reasonable. For all villages, the improved aesthetics of the Thur section is an important predictor of acceptance. This indicates that an anthropocentric perspective generally dominates. In contrast, animal and plant diversity is a significant predictor only for Altikon and Niederneunforn. The natural condition is perceived as improved on average by all villages. Improved flood protection has a significant impact only for the village Thalheim. For Thalheim, this result is reasonable because the village is situated directly on the river side (see Figure 1) and has experienced floods. For Altikon, also near the river and flood experienced, no flood item revealed a significant explanation, however.
 General importance (“naturalness is important”) has an interesting pattern. This item is a significant predictor for the Neunforn villages (Oberneunforn and Niederneunforn), but the actual change in naturalness is significant for Altikon and Thalheim. Without overemphasizing this finding, we assume distinguishing between the general importance (measured by items reading “XY is important”) of a natural river and the improved naturalness of the concrete river section through restoration (measured by the difference in the condition before and after) is critical. The same holds for the diversity of animals and plants. Here, too (this time for Altikon and Niederneunforn only), the pattern is that for Niederneunforn, improved diversity is significant, but for Altikon, the general statement about diversity is important. We assume that this has to do with local patterns of how the inhabitants of these villages experience the river. The aspects of the changed/improved riparian conditions the inhabitants experience most are significant predictors of acceptance [Robertson et al., 2000]. In some cases, only the general importance is significant.
 Interestingly, “importance of flood protection” does not show up as significant predictor although it is rated as high as the other importance items. We assume the consensus is that flood protection is generally important, but for accepting a specific restoration it is not a criterion.
 For this study as well as for two companion studies (unpublished reports) conducted essentially by students, it was difficult to access data about the financing, actual goals, and success criteria of the restoration project. This experience is in line with other literature [Kondolf et al., 2007; Palmer et al., 2005], and future work is necessary to investigate if and how these aspects may also play a role regarding acceptance [Rispoli and Hambler, 1999].
 After the first attempt to gain acceptance for the Thur restoration at the Schäffäuli failed, a new (this time more participatory) strategy was chosen, which yielded considerable acceptance, as also shown in our study (for details on the history of the process cf. Zaugg ). The project process became “‘more public’ and thus more complex and ‘more political’ context in which the project planning and implementation of the 1993 and 1997 construction projects occurred, and at the same time, to remain able to take action and pass decisions” [Zaugg, 2002] (p. 283). If we examine how restoration projects actually happen, it is clear that they do not follow a straight stages model. Much about the final design and potential outcome depends on “different distributions of power between the various public and private actors involved at the different stages of restoration policy making” [Baker and Eckerberg, 2013]. Obviously, there might be trade-offs between necessary landscape alterations to ensure a high degree of flood protection, a natural river with high ecological quality, and pleasant aesthetics that are liked by local and residents and recreationists [Pahl-Wostl, 2006; Palmer et al., 2005].
 A limitation of the study is that no appropriate comparison of the conditions “before” versus “after” was possible. The ratings had to be done by the participants in hindsight. If there is visible aesthetic improvement after the finished project, studies relying on hindsight evaluation might inevitably measure high acceptance. We can learn from other studies that captured this effect [Buijs, 2009]: After the restoration was finalized, only a minority still opposed river restoration. However, participants feel considerably connected with the Thur and seem to have enough experience with the river that we can assume reliable responses. Additionally, the pictures of the situation before and after the restoration help individuals recall and compare the respective conditions. A monitoring process regarding public acceptance starting before the building measures and continued and compared over time would have been helpful [Jähnig et al., 2011], though. It is a clear lack of the design of the Thur restoration project that no such evaluation measures were planned and thus no substantial contribution to an overall control of success or monitoring process can be made [Carr et al., 2012].
 In this study, we confirm findings in earlier studies. However, we also extend the knowledge about local inhabitants' perception of river restoration. We assume that it is important to further investigate the differences between user groups and communities we can identify regarding the reasons for (non)acceptance of a restoration project. Only then can one better understand the rationales of local inhabitants. The interesting thing about our study is that the straightforward answers to the hypothesized differences only partly hold. For instance, although the villages we investigated were different distances from the restored river section, the differences in acceptance are extremely low. However, the differences in the independent variables show that actually the reasons differ (why participants accept restoration) depending on their location. Contentwise, the emerging pattern cannot be explained by “proximity.” Thus, we assume some differences in the mindset and specific local culture play a role and should be investigated in future studies.
 Our results of course do not allow for a definitive answer why the relationship patterns between importance and perception emerge. Thus, we can only speculate but can partly substantiate with our observations and information from many contacts we had in the region. Looking at the descriptive statistics, only in the case of scenic beauty, assessment of the present situation outranks importance. There might be two rationales for this: (i) only here, the public is actually able to make an assessment independently from technical expertise and (ii) effects in this regard are probably better visible per se. We actually argue that the importance question hints at a rather general assessment while the perception stands for a more specific situational assessment. Both together but also independently can impact the decision to find the project overall reasonable. We do not believe in a full rationale decision model here (i.e., perception multiplied by importance and aggregated) but that these are two distinct dimensions of the overall assessment. We can actually support this with a factor analysis (not shown but available upon request from the authors), in which the perception and importance items clearly load on different factors. That means one can find the project reasonable even if the effects are not (yet) visible or finally established by expertise but rather one finds river restoration for the different reasons offered already important. Our own discussions with people from the region are in line with this. When talking about the project, most people just referred to the beauty of the restored river and for all the other aspects referred to a research project under way studying this in detail. That means, for the items for which individuals cannot make really an independent judgment, they have to rely on experts and have to trust their expertise [Luhmann, 1993; Siegrist et al., 2005]. Individuals, for instance, often refer to the “Flussregenpfeifer” (English: Little Ringed Plover; Charadrius dubius) found again in the riverbed; however, one can assume that mainly bird experts have observed increased biodiversity and communicated about certain species.
 Looking at the detailed results from the regression analyses, we can further speculate that the clear pattern visible for the farmers (domination of perception of improved naturalness und flood protection) exemplifies that, given a clear stake, one is making a serious assessment of the core ideas of the project, and is then basing one's decision about its reasonability on this concrete assessment of the actual situation. In contrast, the nonfarmer group is probably too diverse to allow for a concrete understanding of the underlying decision patterns. The only exception is the dominant importance of scenic beauty perception. Thus, this points out again that the only assessment the public can make on its own and rationally derive a decision about the project's reasonability is scenic beauty [Junker and Buchecker, 2008]. This can also be illustrated by citing from a recent PhD thesis on river management: “If flood protection, water quality and recreational uses of water are ensured, restored river segments are generally perceived to be fundamentally positive. It seems it does not matter whether there was an actual ecological improvement at the affected river sections or not. As the detailed analysis of the interview texts revealed, it is apparently enough that a revitalized water body ‘looks beautiful’ and has become more ‘accessible’ or ‘user friendly’ after implementation” (Bratrich , translation by the authors).
 Moving on to the differences between the villages, we must emphasize that these aggregates are probably too big and actually hide individual group differences. Nevertheless, some observations let us speculate about the decision mechanisms. Overall, we again see that the only consistently significant item is the perception of scenic beauty. Aesthetics thus seem to play an essential role in the success of river restoration projects. However, this is of course not enough. In two villages, the importance of natural condition plays an equally important role. These villages are in the same canton while the other two are in another canton. As the cantons play a decisive role in the communication of the project (there even were two harmonized but independent restoration projects in the two cantons), this result might reflect this difference. Our own observations are in line with this idea. In the canton of Thurgau, naturalness (in addition to flood protection) has always been stressed. The river restoration project was communicated as a measurement for restoring nature. In contrast, villagers from the canton of Zurich (Altikon, Thalheim) became aware of this probably only after the project's end.
 Overall, these results point to a delicate interplay between individuals' feelings, observations, and trust in experts when making a decision about the reasonability of river restoration. This can in fact be found in other fields, for instance, in risk perception. Project developers and communicators thus have to pay considerable attention by designing and managing a restoration process that cautiously balances input by technical experts and public deliberations (e.g., for the case of opinions on nuclear waste: Krütli et al. , Seidl et al. ).
 A cautionary remark about these speculations is of course necessary. We base our interpretations of individual decision processes on the aggregate statistics only. Further research is necessary that applies qualitative methods by using in-depth interviews to establish a more detailed account of the individual decision rationales, for example. We argue that this would in fact be very worthwhile, since this interplay between general importance and specific perception seems essential from a theoretical and practical perspective.
 Stressing improved aesthetics over and above greater naturalness (and flood protection as guaranteed) after restoration has been a good strategy for increasing acceptance. Since attitudes toward the restoration project before or after were not measured in the sense of process monitoring, a real comparison is not possible. Thus, it remains debatable if a huge effort for participation, stressing a more natural river section, is inevitably necessary in the case of river restoration, or if stressing the issues of guaranteed flood protection but improved aesthetics in communication to the public would have been sufficient to gain acceptance among the different user groups and villages.
 The study was conducted as part of the transdisciplinary case study “Sustainable river restoration: The case of the Thur in Niederneunforn” in 2012 in cooperation with the projects Record & Record Catchment supported by Competence Centre for Environment and Sustainability at the Swiss Federal Institute of Technology. The authors thank the involved students and Sandro Bösch for help with the figures.