The different values of nature: a comparison between university students’ perceptions of nature’s instrumental, intrinsic and relational values

Human–nature relationships are in a critical state, characterized by increasing environmental problems caused by humans. Thus, it is essential to know how to motivate people for environmental protection. Motivation can arise from intrinsic, instrumental or relational perceptions of why nature is valuable. The present study investigates empirically how university students differ in their view of nature’s value. For this purpose, students from 13 different majors across Germany as well as students engaged in the environmental protection organization Health-for-Future (HfF) were surveyed via an online survey (N = 1885). The majors were divided into four thematic groups (environmental, people-aligned, structures-exploring as well as economics and political science). While all groups showed high levels of agreement with intrinsic and instrumental values, the groups differed strongly in their relational perception of nature. Environmental disciplines were characterized by a stronger relational view, followed by people-aligned disciplines, while disciplines not directly associated with either people or nature agreed less with relational and also intrinsic values. HfF as a conservation organization showed the highest level of agreement in all three value components. Further we found that gender plays a role in value perception, with women holding stronger intrinsic and relational values than men. The study concludes that among students in Germany, regardless of major, there is a pluralistic understanding of values: ecosystem services (instrumental values) and the intrinsic value of nature are both appreciated, whereas relational values appear to be more group specific. Since students will be future decision-makers of society, relational values have special significance for the understanding of conservation.


Introduction
Humans and nature share an ancient and historically changing relationship (Costanza et al. 2007;Steffen et al. 2007). However, the relationship is strained these days: although humans depend on ecosystem services (Levine and Chan 2013;MEA 2005), they threaten nature by taking up more and more space, leaving ecological footprints so large that the "Anthropocene" has already been proposed as an epochal term (Crutzen 2002). Biodiversity is rapidly declining (IPBES 2019), land and marine ecosystems are changing and human livelihoods are affected as well: for instance, food security has already been impacted by warming, changed precipitation and extreme events (IPCC 2019). There is a crisis in the human-nature relationship, not least because society is distancing itself from nature (Amel et al. 2017;Jordan 2009;Zylstra et al. 2014). To find suitable solutions to environmental problems arising from this, it is important to look into what people in turn value about nature. Based on this, values describe the different ways in which nature, ecosystems or ecosystem services are important to individuals or social groups (Arias-Arévalo et al. 2017). Nature is a broad term that includes ecosystems but also e.g. biodiversity (for a useful working definition, see Díaz et al. 2015;Pascual et al. 2017). However, people differ in how they value nature (Riechers et al. 2021).
To convince people of the importance of nature conservation, it is necessary to identify the different motivators, i.e., what kind of value people assign to nature, and why they in turn would want to protect nature. In the following, the three common types of values are presented.
For a long time, two opposing values dominated research: instrumental and intrinsic values (Jax et al. 2013;Pascual et al. 2010;Tallis and Lubchenco 2014). Intrinsic values are independent of human valuation (objective value) or a specific benefit (non-instrumental value), taking a non-anthropocentric perspective (Díaz et al. 2015;O'Neill 1992;Pascual et al. 2017). Nature should be protected for the sake of nature itself (Soulé 1985). In contrast, there are instrumental values which depend on human valuation (Justus et al. 2009). Nature should be protected for the sake of humans (Díaz et al. 2015;Pascual et al. 2017) because humans benefit from nature. This further implies that environmental aspects are replaceable if the replacement can fulfill the same functions (Himes and Muraca 2018). Although instrumental values are often valued in monetary terms, the basic idea is dependence on a valuator and, hence, they are not merely defined by market value (Justus et al. 2009). They are associated with the concept of ecosystem services (Reyers et al. 2012), whose popular definition is found in the Millennium Ecosystem Assessment (MEA 2005). According to it, they are the benefits people derive from ecosystems and are categorized as providing, regulating, cultural and supporting services. This categorization serves as the basis for measuring instrumental values (see "Instruments and analysis").
By approaching the old dichotomy of intrinsic and instrumental values philosophically, Muraca (2011) presented relational values as a new category to focus on the relationships between the human and the non-human. Although the idea of values being relational was for instance already mentioned by Brown (1984), the definition differs from current ones (by focusing on the origin of values) as described in Chan et al. (2018). Relational values as a third category of values were then popularized by Chan et al. (2016), since human decisions are not only based on an intrinsic value or the satisfaction of one's own needs, but also on relationships. This means the multiple interconnections between the individual, the society and the natural world (Chan et al. 2016). Relational values in this sense do not reside in the things themselves, but arise out of these relationships (Chan et al. 2016). There are for example landscapes, which are especially important to someone because he or she grew up there (Neuteleers 2020). This is why relational values are connected with the own as well as with cultural identity (Chan et al. 2016;Neuteleers 2020;Riechers et al. 2021). Something that has relational value is not replaceable, which is an important distinction from instrumental values (Himes and Muraca 2018). These relationships can result in responsibilities toward nature (stewardship) and other people by considering what is appropriate for these relationships. Also, this is how relational values can also take the form of eudemonic values or can even include moral values (Chan et al. 2018). They comprise one's idea of a good life, which is meaningful and satisfying, elements such as social responsibility, as well as social norms like justice and considerations such as general welfare (Chan et al. 2016). The literature suggests that (especially) cultural ecosystem services might be relational rather than instrumental values (Arias-Arévalo et al., 2017;Chan et al. 2016) or at least have more than an instrumental dimension (Díaz et al. 2015). But relational values are also criticized, at least as a separate value category. Luque-Lora (2022) makes two major points: all value types were in their substance relational and the concept of "values" was defined differently for relational than for the other two values.
Although perceptions of nature's value influence engagement in conservation (Admiraal et al. 2017;van den Born et al. 2018), there is still much room to investigate how people differ in these values. Some authors already addressed different groups quantitatively (e.g., Klain et al. 2017;Kleespies and Dierkes 2020b;Saito et al. 2021;See et al. 2020;Winter 2007). They provided important clues as to why certain groups want to protect nature. Nevertheless, more quantitative research on relational values can help to discover more unifying elements (Schulz and Martin-Ortega 2018). Therefore, our study focuses on the value perceptions of university students using a quantitative method. Students are an interesting and important research group because of two reasons. First, they form thematic interest groupsdepending on their field of study. Although many aspects play into the choice of major, interest in the subject plays a significant role for women and men in Germany and Austria (Elster 2014). Personality traits with their ideas about life are also important (Porter and Umbach 2006). Second, university students will be involved in decision-making at many levels in the future, contributing their perspectives, personality and interests. Even if not all later work in the field they studied, for many this is the entry into a related career (Lemieux 2014), especially since many professions also require appropriate education. Both points imply that today's students are likely to integrate their perceptions of nature into their careers later in life. Thus, our research questions are: (1) Are there subject group-specific differences in perceptions of the values of nature and (2) if so, what do these differences look like? An approach with different fields of study offers the opportunity to better assess future decision-makers in different fields (e.g., business, politics, ecology). Therefore, students from 13 different majors, grouped by their broader field of study, and medical students engaged in an environmental protection organization (Health-for-Future, HfF) were surveyed about their perception of why nature is valuable.

Participants and sample
In the study, 2353 students participated, of which 1885 participants (31.88% male, 65.46% female, 0.48% non-binary; 2.18% n/a) were divided into subject groups. These groups were then compared (reduced data set). In the final assignment of returned questionnaires to a major, questionnaires were excluded if a second or minor subject was studied that was not named by the student or was one of the other majors chosen to be analyzed.
The majors were chosen to reflect a broad spectrum of disciplines: natural sciences, humanities and social sciences. A total of 13 majors were compared: biology, geoscience and geoecology categorized as Environmental; medicine, psychology, sociology, theology and teacher training as people-aligned; economics and political science as E + P; chemistry, computer science and mathematics as structuresexploring. The categorization was based on orientations and content of the majors. Hence, the People-aligned group is specified through its focus on human well-being, whereas the Structures-exploring group tends to focus on inanimate objects.
Theology refers to evangelic and catholic theology. Teacher training comprises various subject specializations and school systems and is therefore connected by the idea of pedagogy. The subject designations and contents may vary, e.g., "biology" includes mainly biosciences or biology, but also specialized majors such as marine biology. Such specialized forms make up only a small proportion of the respective groups. Additionally, medical students engaged in the movement Health-for-Future, an offshoot of Fridays-for-Future, were surveyed. This organization focuses on the connection between environment and health, giving them a specific goal in conservation linked to human beings. For that reason, they are an interesting group to study. The organization is open to all, both people from the health sector and everyone else who identifies with their aims (KLUG 2022b), but we limited our sample to medical students.
A total of 95 colleges and universities are represented in the reduced data set. A breakdown of the sample size by major and gender is shown in Table 1. The average age is 23.01 years and the average of semester studied is 4.2, whereby 9th semester or higher was the highest possible answer in the survey (Sup . Table S1).

Procedure
The survey took place over a period from August 2020 to February 2021. The questionnaire was shared via various social networks (WhatsApp, Discord, Instagram, Facebook, Telegram, Moodle), email distribution lists and websites, partly via intermediaries such as student councils or secretary offices. The questionnaire was completed online using the author's university survey system to ensure the highest possible level of data protection and anonymity. The students were informed that participation was voluntary and 1 3 anonymous. There was neither an advantage nor a disadvantage due to non-/participation. The survey was conducted throughout Germany in German language; eight students from Austrian universities in the sociology group, and one from a Swiss University in Chemistry took also part in the study. There was no time limit, and the questionnaire included a short introduction, which stated that the aim was to investigate the connections between diverse environmental concepts. All groups received identical questionnaires.

Instruments and analysis
The students were asked about their agreement with the statements on a five-point Likert scale (disagree to agree), which represent the three types of values. The designed statements for intrinsic values are based on their definition as non-instrumental, i.e., the benefits of nature to humans do not matter. Additionally, they are built on the notion that nature is independent of an evaluator. Item INTR4 (Table 2) was inspired by Winter (2007). Three intrinsic value items (INTR2, INTR4, INTR7) were negatively defined, so a sign change was necessary.
For measurement of instrumental values, four items capturing the four types of ecosystem services were designed. Finally, for relational values, the questions of Klain et al. (2017) were adopted in German according to Kleespies and Dierkes (2020a), which were tested for reliability and achieved good results, including the change of the personal pronoun "I" to "we" in RV_resp. The German version of all items is shown in supplementary Table S2.
For analysis, SPSS Statistics 28 by IBM was used. To form components that represent the three value types, an exploratory factor analysis in the form of a principal component analysis (PCA) was conducted with all received questionnaires (raw data set). A pairwise exclusion criterion and varimax rotation were applied. It was checked whether items had to be removed due to small loadings or moved to another component (this refers in particular to the opposition between instrumental and intrinsic values). This procedure is described in the results. To test for reliability, Cronbach's alpha was calculated (Tavakol and Dennick 2011). To verify the applicability of a factor analysis, the Kaiser-Meyer-Olkin (KMO) test for sampling adequacy and the Bartlett test of sphericity were performed. Correlation coefficients were determined using a two-sided Pearson correlation. Further evaluation was carried out with the reduced data set of 1885 participants. The initial focus was on comparison between the five subject groups, but differences between majors were also tested to control for significant differences between them. After establishing the components with PCA, the statistical analysis was performed with the calculated mean values of the items belonging to the respective component (equally weighted). The Kolmogorov-Smirnov test was applied to test for normal distribution. Because there were no normal distributions aside from individual groups and majors in the relational dimension, comparisons between groups and between majors were performed using Kruskal-Wallis tests. To compensate for type I error due to multiple groups, the Dunn-Bonferroni test was applied as a post hoc test. p values ≤ 0.05 are considered significant. Effect size was calculated with r = �z�/ √ n according to Fritz et al. (2012). According to Cohen (1988), effect sizes are classified as follows: r > 0.1, weak effect; r > 0.3, medium effect; r > 0.5, strong effect. In addition to the differences between disciplines that are the focus of this study, gender differences (male/female) were also tested to examine the influence of this variable. For this, the Mann-Whitney U test was used. Due to the small amount of data, for HfF, no difference between genders was tested.

Results
The first analysis showed the eigenvalues for each component. Of these, four components had eigenvalues over 1 [Kaiser-criterion (Kaiser 1960)], explaining 46.45% of the variance. The fourth component had an eigenvalue of 1.19. Due to content validity, the items were then fixed on three components, representing the three value categories. Together, these three components explain 47.75% of the variance with eigenvalues of 1.55, 1.89 and 4.67. RV_other (Humans have a responsibility to account for our own impacts to the environment because they can harm other people) was removed since it loaded on the instrumental instead of the relational component. Kleespies and Dierkes (2020a) also removed the corresponding item. INSTR4 (cultural ecosystem services) showed very similar loadings for the relational and the instrumental component. This ambiguity was to be expected, as cultural ecosystem services contain both instrumental and relational elements (Díaz et al. 2015). In our study, we define them as instrumental because of the chosen wording ("as it can serve"), indicating the benefit humans derive and moreover representing one of the four ecosystem services. Also, RV_resp showed a loading difference < 0.2, but was not removed as it is part of the construct used by Klain et al. (2017) and reflects the idea of responsibility toward the land, an important idea of relational values. All other items had loading differences > 0.2.
The loadings of the final item set are shown in Table 2. The KMO criterion had a value of 0.865 and the Bartlett test was significant (p < 0.001), confirming that the data set is suitable for factor analysis (Brosius 2013;Kaiser 1974 Alpha increases when this item is removed. However, since the instrumental values here cover ecosystem services, and each item represents one of these services, there is a valid reason to keep the item. All three value categories correlated significantly (p < 0.001). Relational and intrinsic values were moderately correlated with a coefficient of 0.414, whereas relational and instrumental values correlated only by 0.258. Between instrumental and intrinsic values, the correlation was moderate with a coefficient of 0.362.
An overview of mean values, standard deviations and medians for all groups by component is given in Table 3. While instrumental and intrinsic values were rated similarly high by the groups, the level of agreement with the relational items varied most (see Fig. 1).
All groups showed high levels of agreement with intrinsic values. The mean values ranged from 4.48 ± 0.656 (E + P) and 4.48 ± 0.577 (structures-exploring) to 4.88 ± 0.169 (HfF). Sorted by descending mean value, HfF is followed by environmental and people-aligned. Structures-exploring and E + P agreed significantly less with intrinsic values than the other three groups (p < 0.001). The effect sizes ranged from  (Table 4a).
For relational values, the mean values ranged from 3.48 ± 0.723 to 4.28 ± 0.380; HfF had the highest and Structures-exploring the lowest level of agreement. Eight significant differences emerged as shown in Table 4b. HfF valued this dimension significantly more than all other groups-followed by environmental, people-aligned, E + P and at last Structures-exploring as shown in Fig. 2. This order resembles the ranking in the intrinsic dimension, but is more distinct.
The strongest significant effect was between HfF and structures-exploring, with a moderate effect size of 0.34. The smallest effect was between Environmental and Peoplealigned, with an effect size of 0.10. In total, there were six small and two moderate effects (Table 4b).
The instrumental items were rated at a similar high level as those of the intrinsic dimension. The means ranged from 4.59 ± 0.560 (E + P) and 4.59 ± 0.512 (Structures-exploring) to 4.84 ± 0.289 (HfF). With the exception of HfF, which had significantly higher values than all others, there were no significant differences between the groups (Table 4c). The order of groups sorted by their mean values that showed in the other two components barely showed in instrumental values because of the very similar agreement with the instrumental items (see Fig. 1).
Next, the individual majors were compared without regard to the group. The associated descriptive statistics are shown in supplementary Table S3. In the relational dimension, there were many differences with a total of 27 significances, but within the chosen groups (e.g., between geology and biology), there were no significant differences (Sup . Table S4). This also holds true in the intrinsic dimension, with 15 significant differences (Sup. Table S5).
The only exception showed in the instrumental component, where one of the four significant differences was between psychology and medicine-both People-aligned (Sup . Table S6).
Finally, we tested for gender differences. Significant gender differences were found in all groups for both intrinsic and relational values (except HfF, which due to its small number of male participants could not be tested). For instrumental values, a significant gender difference was found only in People-aligned. In each significant case, women showed higher agreement with the values. Table 5 shows the associated p values and effect sizes. The mean values by gender and component can be found in the appendix (Sup . Table S7). If the groups are then sorted by mean, the resulting order differs between men and women for instrumental values. For relational and intrinsic values, there is only a slight difference between genders with Environmental and HfF still having the highest mean values. Note that this order of means is not the same as significant differences.

Discussion
The aim of our study was to find out if students of different fields differ in their perceptions of nature, and if, what these differences look like. To discuss our findings, we first take a look at the measurement instruments and then at the correlations between the values. Finally, we discuss the results in terms of the different study fields and other influencing factors to answer our research questions.
A PCA was used to establish the three components. This was important since most items had not been used before. Nevertheless, a few items need to be discussed in more detail.   Table 4b The Winter-inspired item (Winter 2007) INTR4 (ugly parts of nature have no meaning) loaded as expected negatively on the intrinsic factor, but aesthetics can also be seen as a relational value (Arias-Arévalo et al. 2017; Lliso et al. 2022). Ugliness would then be equivalent to a relational or instrumental disvalue (Lliso et al. 2022). However, if nature is valued even if there is no benefit of beauty, a non-instrumental (Brady and Prior 2020) and thus intrinsic value is acknowledged. Since INTR4 loaded well on the intrinsic component, and there are theoretical considerations emphasizing this decision, the item was kept as intrinsic.
Quantitative analysis can help to compare values, but as is evident in the case of cultural ecosystem services, which are sometimes considered relational instead of instrumental, it has the disadvantage of forcing certain ideas into a fixed value type. For example, Pascual et al. (2017) argue that Nature's Contributions to People, a value-integrative concept, cannot always be covered by one type of value system. Given the controversy surrounding the aesthetic value of nature and the difficult classification of cultural ecosystem services (instrumental or relational), we recommend that their classification be considered anew in each new study.
To investigate the relationship between values, we then calculated correlations. As expected, perceptions of nature's values were positively correlated, as also found in other studies (See et al. 2020;Winter 2007). The positive correlations suggest that high agreement in one component often brings high agreement in another (and vice versa). Higher correlations between instrumental and intrinsic as well as between instrumental and relational values were measured by See et al. (2020), while they found the correlation between relational and intrinsic values to be the weakest in their study despite a strong correlation. In contrast, it was the strongest correlation in the present study. Likely reasons include different cultural backgrounds and measurement instruments.
After this initial analysis, we turned our attention to answering the research questions. First we asked, if there were subject-specific differences in the valuation of nature. The results reveal differences in value perceptions between the disciplines, but especially in relational values. Instrumental values were rated very high by all groups and thus showed little difference. Since the associated items reflect ecosystem services, this means that students of all disciplines clearly acknowledge nature's benefits. However, the underlying scale is based on agreement, and in the case of ecosystem services, it seems difficult to deny their benefits. This may have caused the generally high agreement, the close ranking, and the few significant differences. But recognizing these benefits is not the same as understanding them, as Guo et al. (2010) states.
The likewise high levels of agreement with intrinsic values indicate that no group values nature only for selfinterest. This is comparable to Butler and Acott's (2007) conclusion that intrinsic values are no longer the preserve of environmental ethicists, but appear to be universally accepted. In conservation, intrinsic and instrumental values are often contrasted (Maguire and Justus 2008;McCauley 2006). Additionally, intrinsic values are also used synonymously with non-instrumental values and are defined as such (Justus et al. 2009;O'Neill 1992). Despite gradual differences, however, both categories, instrumental as well as non-instrumental, were seen as important by all groups and appear to be fundamental components of students' perception of nature.
Next, we asked what these differences between groups and majors were, and based on our results one can say that a rough pattern showed in intrinsic and relational values, but not instrumental values.
In both the intrinsic and relational dimensions, majors working closely with and on nature had the highest mean values. In the middle range, those majors that have or will have increased contact and work closely with people were ranked. The lowest mean values were for majors that study basic structures and are not directly related to people or nature. While this pattern was distinct in the relational dimension, it was weaker in the intrinsic component because of fewer significant differences and a mixed order of the individual majors belonging to Structures-exploring and E + P (as shown in Table S5). Now, there are many possible reasons why the groups have different levels of agreement. Relational values are not just individual, but also shaped culturally and socially (Ishihara 2018;Muradian and Pascual 2018)-therefore, the social environment may be crucial. Furthermore, relational values can already be formed in childhood through contact with nature (van den Born et al. 2018), whereby the significance in individual childhood memories and not the frequency alone is decisive (See et al. 2020). Thus, students may have chosen a major based on their experiences and personal interests, which is then reflected in their relational values. Additionally, biology, geoecology and geoscience courses address our environment specifically. Sometimes, practical applications are integrated, as well as human-environment relationships, while e.g., chemistry (although it is a natural science) focuses mostly on different aspects and more specific structures of nature. In a study by Kleespies and Dierkes (2020b), relational values between high school students (many with biology as an advanced course), firstyear biology students, and advanced students showed no significant differences. At the same time, concepts overlapping with relational values such as nature connectedness exist (Kleespies and Dierkes 2020a). By reviewing literature about such similar concepts, Britto dos Santos and Gould (2018) concluded that relational values are dynamic and changeable and, accordingly, positively influenced by environmental education. Together, this provides clues as to why environmentally related majors have higher relational values. Also noteworthy is the medium to high level of agreement of individual People-aligned majors to relational values.
Since relational values were conceptualized by Chan et al. (2016) as a network of relationships, which also includes relationships between human, there are aspects not associated with environmental considerations alone. Klain et al. (2017) argue that by linking nature to human well-being, family and more, these values have the potential to be perceived as a part of the individual's realm of care-they do not have to be environmental. This could be one reason why people-oriented majors agreed relatively strong with relational values. Their view on nature might be based on their focus on fellow people. For example, psychology students could feel addressed, in that they have an interest in nature or its protection out of a responsibility and a connection to other people as well as nature is a source of psychological well-being (Pritchard et al. 2020). This may be one reason why they stand out from majors with a less pronounced people or nature focus. As the relational items used do not differentiate which aspects exactly lead to agreement (whether environmental or people-related, and which elements of these exactly), other factors may also play a role. The subdimensions of relational values (community, connection, care; Kleespies and Dierkes 2020a) can be a starting point for further research.
Finally, HfF stands out due to its overall high values. The idea of a meaningful life, i.e., a basic relational idea, is an important motivation for environmental commitment (Admiraal et al. 2017;Molinario et al. 2020;van den Born et al. 2018), as is the intrinsic idea of nature (Admiraal et al. 2017). Both are reflected in the results of HfF by having the highest mean values in all value scales. With its motto health needs climate protection (KLUG 2022a), HfF combines two key elements of relational values: nature (climate) and society (health) (Chan et al. 2016;Kleespies and Dierkes 2020a;Saito et al. 2021;Schröter et al. 2020). The only significant difference between HfF-engaged and regular medical students occurred in the relational dimension, indicating the importance of this category. It may be decisive for engagement; or at least support a positive opinion about conservation (Riechers et al. 2021). HfF even differs significantly from the environmental group in relational and instrumental values, while no significant difference is seen in intrinsic values. One possible explanation is HfF's specific focus on people. This anthropocentric perspective is integrated in just these two values. Possibly, the human focus is not as present in the environmental group as it is in HfF, even if Environmental showed high means in general. Also, the step between an environmentally related study and the engagement in an environmental group might cause the difference. It must be noted though that HfF was the smallest test group.
Up to this point, the possible influence of the various disciplines has been discussed, but other factors should also be considered. In particular, the difference in gender distribution is striking. The proportion of women is higher in all groups except Structures-exploring. Due to that, the female perspective has a stronger influence on our results. For Germany, gender differences are known for relational values among biology students; women show a stronger relational understanding of nature (Kleespies and Dierkes 2020b). Our results support this finding. Here, women tend to have higher agreement with intrinsic as well as relational values than men. Concerning instrumental values, there was no clear gender difference (except for People-aligned). This might be caused by the ceiling effect, but fewer differences between gender or even an inverse trend in this component cannot be ruled out.
Indeed, gender differences are often (but not always) noticeable in studies in the environmental field. For example, Dornhoff et al. (2019) found that female high school students in Germany had higher environmental concern, which was not the case in Ecuador. McCright (2010) discusses the observation of higher climate concern among US females with the difference in socialization between men and women. Kleespies and Dierkes (2020b) also refer to this theory as a possibility for the differences in relational values, while Duong and van den Born (2019) observed no differences in the human-nature relationship in Vietnam. Gender differences and their causes are a controversial field. In this context, the important point is that differences in values could also be influenced by gender distribution in the majors and groups.
Other possible influences are age, semester and cultural background. The mean age is within a similar range.
Semesters of subject studying differed more. Even so, Kleespies and Dierkes (2020b) found no difference between semesters in relational values for biology students. Cultural differences are conceivable, as it is unknown whether cultural diversity is similar across groups. Finally, it is also unknown whether there are differences in socioeconomic statuses.
Overall, the results of this study indicate that there are correlations between values and field of study, although the results cannot prove whether these similarities and differences already existed before studying or whether the students developed, adopted or changed their values during the course of their studies. The combination of environmental education, a correspondingly supportive social environment, gender, interest in nature (under the assumption that the students chose majors of their interest) and experiences in nature offers at least possible explanations for the high mean values of environmental oriented majors.

Implications
Since students are likely to work in a field related to their choice of study, partly in high positions, two important approaches can now be derived from the findings.
First, to motivate students to take conservation action, it may be useful to argue with values that they can identify with and relate to. Based on our results, this means connecting the protection of nature with its intrinsic value as well as linking protection with nature's services for humans since agreement was very high in both value categories. Yet, it is important to keep in mind that people can perceive nature differently (e.g., Riechers et al. 2021;Winter 2007) and the groups in our study did. So those whose (future) profession is already related to environmental topics may be more receptive than others to the idea of relational values, while everyone may have a basic understanding on how (instrumental) dependent we are on nature. There is no certain answer to the question of which values most motivate conservation action. Van den Born's (2018) results suggest a particular relational factor, and the conservationist group HfF had significant higher agreement with relational values than "normal" medical students or even environmental majors (except geoecology), supporting the importance of relational values. The acknowledgement of different values aims to encourage everyone to contribute to nature conservation and sustainability, by allowing everyone to find value in nature and thus in their actions.
This brings us to the second point: integrating the multiple values into the curriculum at universities can create an expanded awareness of how multifaceted the relationships between people, nature and society actually are. This awareness can prove helpful in situations where naturerelated decisions must be made. The goal is not to consider the values separately, but to use the range of values that nature has to offer and to use them consciously in the argumentation for nature conservation.

Limitations
Although this study was conducted with great care, some limitations must be considered. These limitations refer to the sample, the analysis as well as to special circumstances and will now be explained in more detail.
Due to the form of data collection, the samples are nonrandomized. The study refers to the whole of Germany, while the federal states are represented differently. The data for geoecology is based solely on one university, while the other data is from a varying number of universities. The number of participants varies between majors, so that the majors are represented to different degrees in the subject groups. HfF represents students engaged in environmental protection, but we cannot verify if some individual students of the other groups are engaged in similar organizations. Further, in the discussion, it has already been shown that other influencing factors, especially gender, are also conceivable. The gender distribution of our sample is not representative.
The exploratory factor analysis showed that some items are to be evaluated critically. This concerns which value dimensions cultural ecosystem services and the aesthetic value of nature are assigned to. Here, as above, the recommendation is to re-examine these items in each new study using data analysis and theoretical considerations. Moreover, the items used for measuring relational values, as mentioned earlier, do not clearly refer to only one idea of relational values, but combine different ideas in one item, limiting their explanatory power. James (2022) criticizes the different possible interpretations of Klain's items, which would make it difficult to distinguish them from the other value types.
Finally, some influencing factors need to be mentioned. The data were collected during the pandemic of COVID-19, and with that the students could be affected by, e.g., lockdowns or social distancing. A study by the German government showed that the appreciation of nature has changed for young people: for 52% nature has become more important, 45% noted no difference (BMU 2021). Beyond that, the surveys were completed online, so the decision to participate may have been an unintentional selection based on interest. Social desirability may play a role, but the survey was anonymous, and anonymity may mitigate this effect (Joinson 1999). Together, these effects could have led to higher levels of agreement. Assuming all participants had the same prerequisites, a comparison, however, which was the objective of this study, should be nonetheless possible.

Conclusion
In our results, all groups showed a pluralistic understanding of nature in the form of a general medium to high agreement with intrinsic, instrumental and relational values. The environmental organization HfF showed the highest level of agreement in all three value categories, indicating a link between values and engagement.
Although students have fundamental intrinsic and instrumental value concepts, relational values might be important motivators for conservation action (van den Born et al. 2018). Yet, it was these values that showed the most differences between subject groups. Based on our results, we come to three conclusions. (1) Since it is not fully known which values are most decisive for conservation action, none of them should be disregarded. (2) The spectrum of values should be used instead, as people differ in their valuation of nature. Instrumental values were generally agreed on by the students, while relational values do not seem to be that close to many students. To gain the support of different groups, their personal opinion of why nature is valuable should be considered. (3) The awareness of such multifaceted values of nature should be promoted in university education, since current students will soon be shaping the environment as professionals in decision-making positions. These decisions will also affect other people, which might value nature for other reasons.
Funding Open Access funding enabled and organized by Projekt DEAL. This study was partly supported by the Opel-Zoo Foundation Professorship in Zoo Biology from the "von Opel Hessische Zoostiftung".

Data availability
The raw data supporting the conclusions of this article will be made available by the authors to any qualified researcher.

Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Ethical statement
The studies involving human participants were reviewed and approved by the Ethics Committee of the Science Didactic Institutes and Departments (FB 13,14,15) of the Goethe University Frankfurt am Main.
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/.