Social media summary: New scales shed light on eco-evolutionary explanations for human morality and cooperation with other species.
Introduction
There is fundamental uncertainty as to why people care about wild species at all and why they feel strong moral obligations towards some species but not others (Chan et al., Reference Chan, Balvanera, Benessaiah, Chapman, Díaz, Gómez-Baggethun and Turner2016; Hare et al., Reference Hare, Blossey and Reeve2018; Lehnen et al., Reference Lehnen, Arbieu, Böhning-Gaese, Díaz, Glikman and Mueller2022; Soulé, Reference Soulé2013). Nevertheless, concerns about effects of human exploitation of other species are reflected in value systems (de Groot et al., Reference de Groot, Drenthen and de Groot2011; Gamborg & Jensen, Reference Gamborg and Jensen2016; Teel & Manfredo, Reference Teel and Manfredo2010; Teel et al., Reference Teel, Manfredo, Jensen, Buijs, Fischer, Riepe and Jacobs2010) and cultural norms worldwide (Artelle et al., Reference Artelle, Stephenson, Bragg, Housty, Housty, Kawharu and Turner2018; Berkes, Reference Berkes2017; Turner et al., Reference Turner, Ignace and Ignace2000, Reference Turner, Ari, Berkes, Davidson-Hunt, Ertug and Miller2009). Despite contradictory views about the role that human interests should play (compared with non-human interests) in conservation objectives (Hare et al., Reference Hare, Blossey and Reeve2018), the widespread concern about other species’ well-being in value systems and cultural norms across societies suggests that humans express moral responsibility towards other species. Global biodiversity declines and contentions about how conservation efforts should be allocated call for an interdisciplinary approach to understanding conservation ethics (italic terms are defined in the glossary in Box 1). Despite recent work showing an evolutionary cooperative foundation for morality among people, specific linkages between evolution, morality and conservation have not been explored. An eco-evolutionary framework for understanding conservation ethics could provide novel insights into the adaptive mechanism underlying cooperation with other species and propose an ultimate explanation for why conservation ethics vary (Hare et al., Reference Hare, Blossey and Reeve2018).
Conservation efforts are typically morally justified using frameworks of intrinsic (Lute et al., Reference Lute, Navarrete, Nelson and Gore2016; Vucetich et al., Reference Vucetich, Bruskotter and Nelson2015), instrumental (Justus et al., Reference Justus, Colyvan, Regan and Maguire2009) or relational value (Arias-Arévalo et al., Reference Arias-Arévalo, Martín-López and Gómez-Baggethun2017; Chan et al., Reference Chan, Balvanera, Benessaiah, Chapman, Díaz, Gómez-Baggethun and Turner2016), despite recognition that such value orientations cannot capture the full diversity of conservation ethics (Chan et al., Reference Chan, Balvanera, Benessaiah, Chapman, Díaz, Gómez-Baggethun and Turner2016; Sandbrook et al., Reference Sandbrook, Scales, Bhaskar and Adams2011). Although justifications for conservation can be scrutinised in terms of these ethical frameworks, evolutionary scientists have yet to explain why people care about other species at all (i.e. the ultimate adaptive reason for cooperation). The evolutionary theory of morality-as-cooperation (MAC) proposes that morality evolves to solve cooperation problems and can explain moral behaviour among humans (Curry, Reference Curry, Shakelford and Hansen2016; Curry et al., Reference Curry, Hare, Hepburn, Johnson, Buhrmester, Whitehouse and Macdonald2020; Curry et al., Reference Curry, Mullins and Whitehouse2019; Tomasello & Vaish, Reference Tomasello and Vaish2013). Empirical tests of MAC also demonstrate that people's moral psychology (i.e. proximate mechanisms) maps onto this evolutionary cooperative theory. In fact, cooperation between species could explain why people express moral concern for non-human organisms, why they express greater concern for some organisms than others, and why this concern for different organisms varies across diverse ecological and socio-cultural contexts (Hare et al., Reference Hare, Blossey and Reeve2018). However, whether MAC's fundamental logic applies to a conservation context has yet to be empirically tested.
Although cooperation apparently contradicts Darwin's influential idea of ‘nature red in tooth and claw’ (Bowles & Gintis, Reference Bowles and Gintis2011; Darwin, Reference Darwin1859), cooperation is a critical component of social behaviour in many species, from microbes to plants to people. Natural selection should favour behaviour that benefits one's own reproductive success or that of genetically related individuals (i.e. inclusive fitness; Hamilton, Reference Hamilton1964). Yet interspecific cooperation is widespread (Kiers et al., Reference Kiers, Duhamel, Beesetty, Mensah, Franken, Verbruggen and Bücking2011) and manifests across all levels of ecological organisation (Barker et al., Reference Barker, Bronstein, Friesen, Jones, Reeve, Zink and Frederickson2017; Harcombe, Reference Harcombe2010; Sachs et al., Reference Sachs, Mueller, Wilcox and Bull2004; West et al., Reference West, Griffin and Gardner2007). However, ultimate justifications for cooperation, including inclusive fitness (Hamilton, Reference Hamilton1964), reciprocal altruism (Axelrod & Hamilton, Reference Axelrod and Hamilton1981; Trivers, Reference Trivers1971) and stakeholder theory (Roberts, Reference Roberts2005), constitute a rich body of work on what motivates cooperation and determines who individuals should cooperate with. Hamilton demonstrated that natural selection would favour cooperating with kin over non-kin by virtue of shared genes (Hamilton, Reference Hamilton1964). Nevertheless, cooperation amongst unrelated individuals is widespread and prevalent in our everyday lives, from cooperating with peers to helping strangers even without expectation of reciprocation (Fehr & Fischbacher, Reference Fehr and Fischbacher2003; Roberts, Reference Roberts2005).
A powerful explanation for non-kin cooperation rooted in evolutionary logic is fitness interdependence: ‘the degree to which two or more organisms positively or negatively influence each other's success in replicating their genes’ (Aktipis et al., Reference Aktipis, Cronk, Alcock, Ayers, Baciu, Balliet and Winfrey2018: 429). Fitness interdependence among individuals, including individuals who are not genetically related, can arise from shared fates and interests established and enhanced by socio-cultural norms and institutions (Aktipis et al., Reference Aktipis, Cronk and de Aguiar2011, Reference Aktipis, Cronk, Alcock, Ayers, Baciu, Balliet and Winfrey2018). Fitness interdependence can be positive, as when interacting individuals increase each other's fitness (e.g. symbiosis or mutualism), and negative, as when interacting individuals compete for the same limited resources (e.g. predation or parasitism), or neutral (Ayers et al., Reference Ayers, Sznycer, Sullivan, Guevara Beltrán, van den Akker, Muñoz and Aktipis2023; Cronk et al., Reference Cronk, Steklis, Steklis, van den Akker and Aktipis2019) (Figure 1). However, it is impossible in any given situation for an individual to know definitively their degree of fitness interdependence with others (Ayers et al., Reference Ayers, Sznycer, Sullivan, Guevara Beltrán, van den Akker, Muñoz and Aktipis2023). The perceived fitness interdependence (PFI) scale captures underlying proximate aspects of interdependent relationships by assessing the degree to which an individual's emotional and fitness outcomes intertwine with potential outcomes of specific others (Ayers et al., Reference Ayers, Sznycer, Sullivan, Guevara Beltrán, van den Akker, Muñoz and Aktipis2023). Because proximal indicators of aligned fates, such as ‘closeness’ and ‘oneness’ are considered cues to infer interdependence and long-term fitness outcomes (Balliet et al., Reference Balliet, Tybur and Van Lange2017; Columbus et al., Reference Columbus, Righetti, Balliet, Machia, Agnew and Arriaga2020; Gerpott et al., Reference Gerpott, Balliet, Columbus, Molho and de Vries2018; Korchmaros & Kenny, Reference Korchmaros and Kenny2001), fitness interdependence can be approximated using measures of these variables.
The PFI scale has hitherto only been applied to human–human cooperation. Yet its strong correlation with other broad measures of interdependence, such as welfare tradeoff ratios (Delton & Robertson, Reference Delton and Robertson2016; Sznycer et al., Reference Sznycer, Delton, Robertson, Cosmides and Tooby2019; Tooby et al., Reference Tooby, Cosmides, Sell, Lieberman, Sznycer and Elliot2008) and willingness to help others even in the absence of reciprocity (Roberts, Reference Roberts2005), makes it a plausible framework for assessing cooperation between humans and non-humans. Using eco-evolutionary models, Hare et al. (Reference Hare, Blossey and Reeve2018) extend the fundamental logic of fitness interdependence to include interdependence between humans and other species. Specifically, Hare et al.'s (Reference Hare2018) models illustrated the possibility that conservation ethics evolve because people's fitness covaries with the success of other species. For all individuals, the strength and the sign (positive or negative) of this covariance will vary with different species. Hare et al. (Reference Hare, Blossey and Reeve2018) predict that individuals will care most about species with the strongest positive fitness interdependence and express antipathy towards species with the strongest negative fitness interdependence. However, no study has yet empirically tested predictions from these models. Although social and environmental psychologists have explored many proximate mechanisms for conservation behaviours including attitudes, beliefs, values and norms, the evolutionary bases of such behaviours have not been extensively investigated (van Vugt et al., Reference van Vugt, Griskevicius and Schultz2014). Bridging ultimate and proximate explanations requires this important knowledge gap to be filled.
Our objectives are to develop new scales for (1) human–non-human PFI and (2) conservation ethics as the necessary first steps to testing these predictions. Our work expands on related but separate aspects of evolved moral psychology derived from published measures that index fitness interdependence among humans (Ayers et al., Reference Ayers, Sznycer, Sullivan, Guevara Beltrán, van den Akker, Muñoz and Aktipis2023) and moral standing regarding other humans and non-humans (Berndsen & van der Pligt, Reference Berndsen and van der Pligt2005; Crimston et al., Reference Crimston, Bain, Hornsey and Bastian2016; Graham et al., Reference Graham, Nosek, Haidt, Iyer, Koleva and Ditto2011; Piazza et al., Reference Piazza, Landy and Goodwin2014). To this end, we devised an online study using 15 wild plant and animal species as target organisms (Figure 2). We develop these scales independently to provide the necessary tools for future investigation on the theoretical possibility that the perceived relationship of other species to human fitness (i.e. PFI) might predict conservation ethics towards those species.
Methods
Item generation
We generated items for each scale focusing on indicators of (1) how people view their well-being as associated with the well-being of other species (i.e. PFI) and (2) their moral concern for those species (i.e. conservation ethics) (see Table 1 for all items that were considered). We either adapted or used unchanged nine of 25 PFI items from a recent scale assessing perceived and emotional shared fate among humans (Ayers et al., Reference Ayers, Sznycer, Sullivan, Guevara Beltrán, van den Akker, Muñoz and Aktipis2023). We adapted the items by modifying their wording to be used in the context of human interaction with other species. We adapted or used unchanged all conservation ethics items except for one (i.e. people should deny moral concern for [target]) from published measures of moral standing regarding other humans and non-humans as no single validated measure of conservation ethics existed. We derived one item from Berndsen and van der Pligt (Reference Berndsen and van der Pligt2005), one item from Crimston et al. (Reference Crimston, Bain, Hornsey and Bastian2016), two items from Graham et al. (Reference Graham, Nosek, Haidt, Iyer, Koleva and Ditto2011) and six items from Piazza et al. (Reference Piazza, Landy and Goodwin2014). All authors were involved in item creation and provided feedback on the appropriateness of these items in the context of human–non-human relationships. Colleagues not involved in the study (n = 8) pretested items for clarity and comprehensibility.
Participants and procedure
Following Ayers et al. (Reference Ayers, Sznycer, Sullivan, Guevara Beltrán, van den Akker, Muñoz and Aktipis2023), we conducted a power analysis with α = 0.05, 80% power and an effect size estimate of ƒ2 = 0.06 and determined that we needed a sample size of at least 514 respondents. We recruited 665 adults at least 18 years of age residing in the US via Qualtrics (Qualtrics, www.qualtrics.com), a respondent recruitment platform enabling researchers to access diverse samples. We minimised unreliable responses by automatically screening out respondents who failed at least one of two randomly inserted attention checks (i.e. if you are reading this, select harmful to me; if you are reading this, select agree) or completed the task in less than 100 s (half the median response time). We further excluded 10 respondents who took longer than 800 s (four times the median response time) and four respondents who provided identical responses to all survey items (i.e. straight-lining) – an indicator of low response quality (Zhang & Conrad, Reference Zhang and Conrad2014). Our final dataset contained a total of 651 completed sets of responses (334 males, 313 females, two non-binary, one transgender, one other; age, mean, M = 47.73, standard deviation, SD = 16.57), with an average of 43.4 responses per target organism (minimum n = 37, maximum n = 49). Cornell University's Institutional Review Board approved this study (IRB 2110010651), and all respondents provided informed consent prior to completing the survey.
We randomly assigned each respondent to a single organism selected from a set of 15 plants and animals (Figure 2). We first asked whether respondents knew the target organism. If respondents reported that they did not know the assigned target, we provided them with a randomly selected alternative. Once they affirmed that they knew the organism, we asked them to answer two sets of questions – PFI and conservation ethics – about that organism and complete a short demographic questionnaire. First, we measured perceptions of shared outcomes with the target as degree of agreement with the nine statements in the set using seven-point Likert-type scales (i.e. very harmful to me–very beneficial to me; strongly disagree–strongly agree) (Table 1). We then measured respondents’ moral concern for the target as degree of agreement with a set of 11 statements rated on the same seven-point scale (i.e. strongly disagree–strongly agree) (Table 1). To prevent priming effects, we randomised the order of items within each set. Lastly, we measured respondents’ demographic characteristics using four factors (i.e. sex, ethnicity, age, ZIP code). Wording for all survey items is presented in the Supplementary Materials (Supplementary Table S1).
Species selection
We selected target organisms based on results from a multidimensional study of how people in the US think about wild organisms based on 20 characteristics, such as beauty, charisma, harmfulness to humans, trophic type, familiarity and moral standing (Hare, Reference Hare2018). This study found that wild organisms cluster into three distinct groups: (1) ‘plants’, (2) ‘nasty animals’ and (3) ‘nice animals’. To ensure a range of responses with regards to PFI and conservation ethics, in the current study we selected the five most familiar organisms from each cluster identified in that initial study. ‘Plants’ and ‘nice animals’ were characterised as more charismatic, of higher moral standing and more ecologically, economically and culturally valuable whereas ‘nasty animals’ were perceived significantly more as threats or pests to humans (Hare, Reference Hare2018).
Analysis
We assessed the underlying factor structure for each set of items (PFI and conservation ethics) by performing exploratory factor analyses using oblique rotation with principal axis factoring (Supplementary Figures S1 and S2). We selected the oblique rotation method based on the assumption that the extracted factors are somewhat correlated (Osborne, Reference Osborne2015). To determine the suitability of the data for factor analysis, we used the Keyer–Meyer–Olkin (KMO) criterion (Kaiser & Rice, Reference Kaiser and Rice1974) and Bartlett's sphericity test (Bartlett, Reference Bartlett1950), and used screeplots (Cattell & Vogelmann, Reference Cattell and Vogelmann1977) and parallel analysis (O'Connor, Reference O'Connor2000) to determine the number of factors to extract for data across targets. For both sets of items, screeplots suggested that one factor should be extracted whereas parallel analysis suggested two factors. We therefore considered one-, two- and three-factor (one above what parallel analysis suggests) solutions for PFI and conservation ethics data across targets. We eliminated any inadequate items (i.e. items that are low-loading or cross-loading; Costello & Osborne, Reference Costello and Osborne2005) by conducting a stepwise series of item removals based on factor loadings and re-running the factor analysis after each removal. We retained items with a rotated factor loading of > 0.5 (for the PFI scale) and 0.6 (for the conservation ethics scale) and excluded items with loadings below the selected thresholds. Scores ≥ 0.4 are considered stable (Guadagnoli & Velicer, Reference Guadagnoli and Velicer1988), and we selected the said cut-offs to account for item communality, potential cross-loading and number of items retained.
Items removed for loading < 0.5 for the two PFI factors included (using deer as an example target): ‘What is beneficial to deer is beneficial to me, and what is harmful to deer is harmful to me’, ‘I feel detached from deer’ and ‘Deer and I have different fates’ followed by ‘A world without deer would be worse for me’. Items removed for loading < 0.6 for the two conservation ethics factors included: ‘I feel no personal sense of responsibility to help deer’, ‘It would be important to protect deer from extinction’ and ‘People should deny moral concern for deer’. We determined the internal reliability (i.e. whether scale items consistently measure the same concept) of the extracted factors after item reduction using Cronbach's α (Cronbach, Reference Cronbach1951) and McDonald's omega (McDonald, Reference McDonald1999). We conducted all analyses in R v4.2.1 (R Core Team, 2022) using the psych (Revelle, Reference Revelle2022) and GPArotation (Bernaards & Jennrich, Reference Bernaards and Jennrich2005) packages.
Results
For the reduced PFI and conservation ethics scales, two factors explained the data across targets with five and eight items, respectively. Indeed, fit statistics for both sets of items prior to item reduction indicated that the two-factor solution was the most appropriate fit for the data (Supplementary Table S2). The standardised root mean square residual (SRMR) (PFI 0.05; conservation ethics 0.03), root mean square of the residuals (RMSR) (PFI 0.01; conservation ethics 0.02), root mean square error of approximation (RMSEA) (PFI 0.082; conservation ethics 0.056) and Tucker–Lewis index (TLI) (PFI 0.952; conservation ethics 0.979) of the reduced scales also indicated a good fit within the recommended criteria, namely SRMR < 0.08 (Hu & Bentler, Reference Hu and Bentler1998), RMSR < 0.05 (Hu & Bentler, Reference Hu and Bentler1999), RMSEA < 0.08 (Browne & Cudeck, Reference Browne and Cudeck1992), and TLI > 0.95 (Hu & Bentler, Reference Hu and Bentler1998) (Table 2).
Abbreviations for indices: χ 2, chi-square statistic for goodness-of-fit test; d.f., degrees of freedom; p, significance level; SRMR, standardized root mean square residual; TLI, Tucker–Lewis index; RMSEA, root mean square error of approximation; RMSR, root mean square of the residuals; 90% CI, 90% confidence interval for the RMSEA.
The data for PFI and conservation ethics were suitable for factor analysis, and their scales demonstrated good reliability. Specifically, two-factor solutions for the reduced PFI and conservation ethics scales showed acceptable sampling adequacy (PFI 0.78–0.83; conservation ethics 0.87–0.93) (KMO; Kaiser and Rice, Reference Kaiser and Rice1974) and non-identity correlation matrices (PFI, χ 210 = 929.04, p < 0.001; conservation ethics, χ 228 = 2715.74, p = 0.000) (Bartlett's sphericity test; Bartlett, Reference Bartlett1950), suggesting a good fit of the data for factor analysis (Table 3). The results also indicated acceptable internal reliability across targets (Cronbach's α (PFI 0.79; conservation ethics 0.89); McDonald's omega (PFI 0.83; conservation ethics 0.92)) based on a cut-off value of Cronbach's α = 0.60–0.70 (Hair et al., Reference Hair, Black, Babin and Anderson2010; Nunnally, Reference Nunnally1978) and McDonald's omega = 0.70 (Hermsen et al., Reference Hermsen, Leone, Smalbrugge, Knol, van der Horst and Dekker2013) (Table 3). Since Cronbach's α increases with the number of items (Hair et al., Reference Hair, Black, Babin and Anderson2010), we expected relatively low α-values given that the reduced PFI and conservation ethics scales consist of only five and eight items, respectively.
Abbreviations for indices: χ 2, chi-square statistic for goodness-of-fit test; d.f., degrees of freedom; p, significance level.
For both the PFI and conservation ethics scales, the two-factor structure also emerged as the most interpretable solution (i.e. solution producing the cleanest factor structure, with item loadings > 0.3 and no or few cross-loadings; Costello & Osborne, Reference Costello and Osborne2005). Responses to PFI items resulted in a factor structure with one factor (PA1) identified by three items and the other factor (PA2) by two items (Figure 3, Supplementary Table S3). The factors were correlated, r = 0.75, and accounted for 47.8% of the total variance (explained variance per factor = 19.9–27.9%; Figure 3, Supplementary Table S3). Similarly, responses to conservation ethics items resulted in a factor structure with one factor (PA1) identified by five items and the other factor (PA2) by three items (Figure 4, Supplementary Table S4). The factors were correlated, r = 0.66, and explained 56.9% of the total variance in participants’ responses (explained variance per factor = 19.8–37.1%; Figure 4, Supplementary Table S4). All items loaded highest on their respective factor, and cross-loadings were all smaller than 0.24 (Supplementary Tables S3 and S4), which is < 0.32 and negligible (Tabachnick & Fidell, Reference Tabachnick and Fidell2001).
Discussion
By combining insights from evolutionary anthropology, evolutionary psychology and human behavioural ecology, we provide empirical tools to test theory about a potential cooperative basis for conservation ethics. Specifically, the PFI scale derived from a validated measure of PFI among humans (Ayers et al., Reference Ayers, Sznycer, Sullivan, Guevara Beltrán, van den Akker, Muñoz and Aktipis2023) contributes a novel and reliable way to index PFI with diverse wild organisms. Moreover, we show that existing individual items measuring the moral standing of humans and non-humans (Berndsen & van der Pligt, Reference Berndsen and van der Pligt2005; Crimston et al., Reference Crimston, Bain, Hornsey and Bastian2016; Graham et al., Reference Graham, Nosek, Haidt, Iyer, Koleva and Ditto2011; Piazza et al., Reference Piazza, Landy and Goodwin2014) can be combined to produce a single reliable scale that measures conservation ethics.
Our finding of a two-factor solution for the PFI scale has potential implications for future investigations. Although all items contained in PA1 of the PFI scale measure perceptions of shared fate with a target, one of two items contained in PA2 measures emotional shared fate (Gervais & Fessler, Reference Gervais and Fessler2017; Sznycer & Lukaszewski, Reference Sznycer and Lukaszewski2019), and only one item explicitly measures perceived fitness responses to harming the target. Indeed, we intended to develop an internally consistent scale to measure perceptions of shared fate in general – regardless of its underlying subscales – rather than test for specific dimensions of PFI. Conversely, Ayers et al. (Reference Ayers, Sznycer, Sullivan, Guevara Beltrán, van den Akker, Muñoz and Aktipis2023) tested and found clear support for a priori hypotheses regarding perceived and emotional shared fate as different but related aspects of fitness interdependence among humans. Further, the two factors of our PFI scale are highly correlated, and it may be that information provided by one factor strongly mediates the effect of the other. Future investigations of particular dimensions of interdependence, including perceived and emotional shared fate, would therefore provide valuable further insights into the nature and origin of how people perceive fitness interdependence between humans and other species.
Our novel approach also produced nuanced insights into how items measuring conservation ethics can capture subtle differences in how people think about moral concern regarding specific organisms. For example, items based on fairness, and on care and compassion loaded onto different factors and may reflect distinct domains of morality consistent with MAC (Curry, Reference Curry, Shakelford and Hansen2016; Curry et al., Reference Curry, Mullins and Whitehouse2019), such as impulses promoting more equitable treatment of species based on ‘fairness’ (Curry et al., Reference Curry, Hare, Hepburn, Johnson, Buhrmester, Whitehouse and Macdonald2020), or care for companion or culturally significant species based on ‘kin values’ (Morris & Qirko, Reference Morris and Qirko2020; Qirko, Reference Qirko2017). For example, in several societies people use kin terms to describe relations with companion animals (Charles, Reference Charles2014; Erikson, Reference Erikson, Podberscek, Paul and Serpell2000; Rose, Reference Rose and Harvey2013; Wilson et al., Reference Wilson, Netting, Turner and Olsen2013).
Items measuring moral attitudes towards harming a target were equally distributed across both factors, but we cannot assess the relationships between these items with data from this study. Future investigations of harm-related vs. other aspects of moral psychology could offer insight into the nature of human moral attitudes towards non-human organisms, such as whether emotions like ‘sympathy’ or those invoked from the ‘killing’ or ‘harming’ of species resonate with different moral rationales for conservation of those species. Although items in the current scale at face value appear to measure the same construct, our finding of a two-factor solution for conservation ethics derived from multiple existing scales indicates subtle discrepancies in how respondents think about moral concern regarding specific species. Thus, this scale provides a more comprehensive tool for measuring conservation ethics and could be adopted by future researchers interested in investigating evolutionary foundations of conservation ethics, or even to measure the moral standing of non-human species beyond evolutionary psychology.
Potential extensions of our approach would assess whether our findings from a broad sample of the US public generalise to other populations. Because different populations live in different socio-cultural and ecological environments, any extensions would have to account for relevant socio-ecologies, such as cultural beliefs and traditions as well as species composition. We conducted our study with a diverse sample of the US population using 15 common wild plants and animals (Hare, Reference Hare2018). The current scales therefore promise a powerful starting point for research, particularly for investigating diverse methodological and theoretical aims in future studies (Ayers et al., Reference Ayers, Sznycer, Sullivan, Guevara Beltrán, van den Akker, Muñoz and Aktipis2023) such as target-specific or cross-cultural comparisons. Indeed, traditional ecological knowledge is often rooted in sustained relationships between people and ecosystems (Berkes et al., Reference Berkes, Colding and Folke2000) and comprises adaptations to local ecologies such as correct ways to relate to locally important species (Artelle et al., Reference Artelle, Stephenson, Bragg, Housty, Housty, Kawharu and Turner2018; Jones et al., Reference Jones, Andriamarovololona and Hockley2008). Recognising that traditional ecological knowledge carries important information about living sustainably in different socio-ecologies is increasingly relevant in a globalised system of biodiversity conservation (Rudd et al., Reference Rudd, Allred, Bright Ross, Hare, Nkomo, Shanker and Dávalos2021), as the role of Indigenous and local peoples in conserving biodiversity brings issues of equity and injustice into sharp relief (Berkes, Reference Berkes2017; Fletcher et al., Reference Fletcher, Hamilton, Dressler and Palmer2021; Kashwan et al., Reference Kashwan, Duffy, Massé, Asiyanbi and Marijnen2021). We recognise that the current scales were developed among an English-speaking Western, educated, industrialised, rich, democratic (WEIRD) sample from a single country (Henrich et al., Reference Henrich, Heine and Norenzayan2010) and are therefore relevant to that particular context. We expect PFI and conservation ethics to be sensitive to local ecological and socio-cultural conditions (Hare et al., Reference Hare, Blossey and Reeve2018) and would encourage researchers interested in applying our scales to a different context to replicate the study in its entirety. Developing relational frameworks that are easily adaptable for use with diverse populations would promote inclusive conservation research that embraces and celebrates cultural and biological diversity, rather than expecting all people to think alike or assuming that moral psychology in the US reflects moral psychology everywhere. Replications of the entire study across multiple socio-cultural and ecological contexts could reveal the similarities and differences in relationships between PFI and conservation ethics across societies. This in turn may provide insight into how ‘rigid’ or ‘flexible’ these relationships are, and whether they vary along with characteristics of different social systems, such as population size, kinship systems, predominant livelihood types and degree of market integration (Díaz et al., Reference Díaz, Pascual, Stenseke, Martín-López, Watson, Molnár and Shirayama2018; Mattison et al., Reference Mattison, Quinlan and Hare2019, Reference Mattison, MacLaren, Sum, Mattison, Liu, Shenk and Wander2023).
Piloting enables the testing of validity and reliability of the research instruments, and feasibility of the study design, prior to data collection of the main study (Alharbi et al., Reference Alharbi, Tolchard and Isouard2019). In addition, following exploratory with confirmatory factor analysis on a new sample allows for testing of whether the hypothesised factor structure of a scale is consistent across different samples and thus whether it reflects its intended construct (Knekta et al., Reference Knekta, Runyon and Eddy2019; Morgado et al., Reference Morgado, Meireles, Neves, Amaral and Ferreira2017). The construct validity of this scale can then be assessed with convergent, discriminant, predictive and concurrent validity by incorporating other validated measures of the same construct (Morgado et al., Reference Morgado, Meireles, Neves, Amaral and Ferreira2017). Therefore, we acknowledge that more detailed factor structure assessment and validation of the scales would assess how robust our findings are.
We do not intend the work we present here to replace, but instead build on, long-standing proximate theories for why humans care about members of other species. Specifically, measures of PFI and conservation ethics can be combined to empirically test predictions at the ultimate-adaptive level. For example, although we did not assess in this study whether different perceptions of organisms relate to their PFI or conservation ethics, we would expect people's PFI to positively covary with moral concern towards those organisms. According to Hare et al. (Reference Hare, Blossey and Reeve2018), people are more likely to express moral responsibility towards organisms they perceive as beneficial to humans and ecosystems (i.e. plants and ‘nice animals’). Therefore, we would expect people to care more about plants and ‘nice animals’ with strong positive fitness interdependence and express more antipathy towards ‘nasty animals’ with strong negative fitness interdependence.
Explanations at different levels of analysis do not compete, and both proximate and ultimate explanations are necessary to fully evaluate the costs, benefits and constraints that shape a given behaviour (Kenrick et al., Reference Kenrick, Griskevicius, Neuberg and Schaller2010; Nesse, Reference Nesse2019; Tinbergen, Reference Tinbergen1968; van Vugt et al., Reference van Vugt, Griskevicius and Schultz2014). Therefore, although PFI offers a particular perspective as to why humans cooperate with members of other species at the ultimate level, it in no way represents the one and only true explanation. In fact, domination (Ingold, Reference Ingold, Manning and Serpell1994; Schwartz, Reference Schwartz2006) and mutualism (Wildavsky, Reference Wildavsky1991) orientation scales (Teel & Manfredo, Reference Teel and Manfredo2010) are widely used to assess people's basic beliefs about wildlife (i.e. wildlife value orientations; Manfredo et al., Reference Manfredo, Teel and Henry2009). Consistent with our conservation ethics index, measures of domination assess people's beliefs regarding harming or killing wildlife whereas measures of mutualism assess beliefs about caring for other species. Conversely, these measures typically target wildlife generally rather than particular species. Moreover, they commonly assess people's agreement with beliefs specifically regarding hunting and managing wildlife and the social affiliation of wildlife with humans. Empirical studies of wildlife value orientations vary widely in the types of hypotheses tested, including emergent patterns across cultures (Jacobs et al., Reference Jacobs, Dubois, Hosaka, Ladanović, Muslim, Miller and Abidin2022), modernisation indicators (e.g. urbanisation, income, and education) (Dietsch et al., Reference Dietsch, Teel and Manfredo2016; Manfredo et al., Reference Mattison, Quinlan and Hare2019, Reference Manfredo, Teel and Dietsch2016; Teel et al., Reference Teel, Manfredo, Jensen, Buijs, Fischer, Riepe and Jacobs2010), socio-demographic characteristics (Bruskotter et al., Reference Bruskotter, Vucetich, Dietsch, Slagle, Brooks and Nelson2019; Gamborg & Jensen, Reference Gamborg and Jensen2016; Teel & Manfredo, Reference Teel and Manfredo2010) and attitudes towards wildlife-related issues and management actions (Dietsch et al., Reference Dietsch, Teel and Manfredo2016; Jacobs et al., Reference Jacobs, Vaske and Sijtsma2014; Teel & Manfredo, Reference Teel and Manfredo2010; Teel et al., Reference Teel, Manfredo, Jensen, Buijs, Fischer, Riepe and Jacobs2010). Further, studies of people's worldviews regarding conservation, such as anthropocentrism, zoocentrism, biocentrism and ecocentrism (Vucetich et al., Reference Vucetich, Bruskotter and Nelson2015), assess their relationship with moral inclusivity (Batavia et al., Reference Batavia, Bruskotter, Jones and Nelson2020), and the acceptability of management of specific wild species (Lute et al., Reference Lute, Navarrete, Nelson and Gore2016). However, no existing measures to our knowledge have tested hypotheses at the ultimate level for why humans cooperate with different species.
Conclusion
Our scales provide empirical, internally consistent tools for studying eco-evolutionary foundations of conservation ethics. Our study also underscores the importance of investigating both proximate and ultimate justifications for cooperation between humans and non-human organisms. Indeed, interspecific cooperation is widespread, and humans frequently cooperate with members of other species (Hare et al., Reference Hare, Blossey and Reeve2018). Despite having only applied these scales to wildlife species prevalent in the US and to a sample of a population of a single country, the high face validity of items allows for a wide range of targets and adaptation for use across diverse cultures. We suggest our approach as a model for producing insights in other ecological and socio-cultural systems.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/ehs.2024.10
Acknowledgements
We thank Jessica Ayers, one anonymous reviewer and members of the Blossey Lab for their helpful feedback on this study.
Author contributions
KL conducted data gathering, performed statistical analyses and wrote the first draft of the manuscript. DH and BB contributed substantially to revisions. All authors contributed to the early conceptualisation and design of the study and were actively engaged in the project from conceptualisation to manuscript submission.
Financial support
This work was supported by kindness.org and the Andrew W. Mellon Foundation.
Conflicts of interest
The authors declare none.
Research transparency and reproducibility
The data and R code that support the findings of this study are openly available in Figshare at https://figshare.com/s/06c21d8ed8855d17c0fa