Introduction

Wild meat remains a critical part of people’s diet (Nóbrega Alves & Van Vliet, 2018). In rural areas, aside from being a valuable source of food, hunting also provides income and medicine; it contributes to strengthening social bonds and identity and is practiced to reduce crop predation by animals considered pests (Coad et al., 2019; El Bizri et al., 2015; Fischer et al., 2013; Nóbrega Alves & Van Vliet, 2018; van Vliet & Nasi, 2008; Van Vliet et al., 2015). In urban areas, the consumption of wild meat has often evolved from a necessity to a delicacy (Wilkie et al., 2016) to keep cultural links with the “village” in a context where increased availability of domestic sources of protein and the decreased availability of wild meat reduced the demand for wild meat daily (Van Vliet et al., 2017).

As a response to the ecological impacts of over-hunting in tropical regions (Coad et al., 2019), social marketing has been increasingly recommended to address the unsustainable use of wildlife (Drury, 2011; Challender & Macmillan, 2014; Thomas‐Walters et al., 2020), and evidence of its effectiveness to reduce wild meat consumption has recently been shown by Chaves et al. (2019) in the Amazon region. To develop such behavioral change strategies based on social marketing, understanding value orientations is essential to avoid imposing western conservation interests and risk social conflict over wild meat management issues (Rickenbach et al., 2017; Van Vliet et al., 2018). Value orientations of wildlife are a set of values, beliefs, attitudes, and norms that help to describe how people think about and act towards wildlife (Manfredo et al., 2017). Incorporating value orientations towards wild meat in conservation strategies is essential to develop adequate wild meat policy and management responses and represent a shift away from the conventional vicious cycle of illegality, un-sustainability and criminalization (Van Vliet, 2018).

Several studies from Sub-Saharan Africa have already provided an understanding of hunting and consumption patterns in both rural and urban contexts in terms of quantities consumed and hunted per species and the drivers associated with wild meat consumption (Luiselli et al., 2018; Chausson et al., 2019; McNamara et al., 2019; Van Vliet et al., 2017; Wilkie et al., 2016; Nguyen et al., 2021). These studies show that availability, price, and cultural attachment are among the main drivers of wild meat consumption. Generally, wild meat is perceived as healthy food and relates to childhood and village life (Nguyen et al., 2021; van Vliet & Mbazza, 2011). However, within a given context, different segments of the population may relate differently to wild meat. This complexity is not captured in current studies, leaving a significant gap in understanding the value orientations that shape behaviors towards wildlife consumption and hunting in Sub-Saharan Africa.

In this study, we sought to elucidate the value orientations towards wild meat and wildlife in the Tshopo Province of the Democratic Republic of Congo in Central Africa, distinguishing between the provincial capital and rural areas. We investigated which socio-economic variables explain differences in the observed value orientations.

Methodology

Study Site

The study was carried out in the Tshopo Province, in the eastern Democratic Republic of Congo. Kisangani, the provincial capital of Tshopo, is the third-largest city in the country. The study covered the urban environment of Kisangani (its six communes, namely Makiso, Mangobo, Kabondo, Kisangani, Lubunga and Tshopo), and a rural environment of 16 villages distributed along the main access roads towards Yangambi, Bengamisa, Ubundu and Madula at about 100 km from Kisangani. Since 2003, at the end of a violent conflict, Kisangani started to slowly re-flourish as an influential trading center of agricultural products and imported merchandise between the DRC hinterland and neighboring Uganda (Takamura, 2015). During the last fifteen years, the urban population from Kisangani has exploded from 247 000 inhabitants in 2002 to 628 000 in 2009 and 1 600 000 in 2015 (Institut National de la Statistique, 2015).

The Tshopo province is marked by a long dry season from December to mid-March, followed by a short rainy season from April to May, a short dry season from June to July, and a long rainy season from August to November. The majority of the province is covered by dense semi-deciduous forests rich in commercial timber species like Pericopsis elata (Afrormosia) or Entandrophragma spp. (redwoods) interspersed with mono-dominant Gilbertiodendron dewevrei forest, associated with a mosaic of secondary forests and agriculture. Wildlife largely contributes to food security and dietary diversity in the region, is consumed daily in rural areas, and is also sold in the urban market (van Vliet et al., 2012; Kyamakya et al., 2018; Van Vliet et al., 2017). In Kisangani alone, about 296 tons were sold between July 2015 to June 2016.

(Van Vliet et al., 2017). Nonetheless, over the last ten years, chicken and pork prices have decreased compared to wild meat, linked to the post-conflict increase in the availability of those domestic meat sources (either imported from Ituri, Uganda, or locally produced) (Van Vliet et al., 2017). This decreased per capita frequency of wild meat consumption, but considering urban population increase, that may not necessarily indicate an overall decrease in urban wild meat consumption.

Some large-sized and potentially vulnerable species such as okapi (Okapia johnstoni), orycterope (Orycteropus afer), giant pangolin (Manis gigantea), leopard (Panthera pardus), chimpanzee (Pan troglodytes), and elephant (Loxodonta Africana) are still present in the region, even if they may have been locally extirpated in some areas (Van Vliet et al., 2018). The region is experiencing an irreversible post-depletion wildlife population process, with hunters increasingly entering relatively undisturbed areas to find larger prey and vulnerable wildlife species becoming extinct over larger proportions of the landscape (Van Vliet et al., 2019).

Data Collection

Before the interviews, permits were acquired from mayors and the heads of the neighborhoods and village chiefs. Nine hundred ninety-seven persons have been interviewed: 517 persons in Kisangani (about 90 in each of the six communes) and 480 persons in villages randomly chosen among those located less than 100 km from Kisangani situated along the primary access routes (about 30 persons per village). All interviewees were informed about the purpose of this research and were requested to give their consent to participate orally. In Kisangani, one household per street was chosen for an interview to spread the sample over a larger area and ensure a better representation. In villages, households were visited in their homes and selected based on convenience sampling. To ensure an equitable representation of gender, we interviewed the male head of the family in one interview and then the female head of the household in the following interview. Audio of the interview was recorded if the interviewee granted permission. The interviews were carried out either in Lingala or Swahili by two bilingual researchers (coauthors of this paper), depending on the origin of the household.

We adapted Dayer’s methodology (2007) based on emotional prompts for personal stories to elucidate people’s values regarding wild meat. As in Dayer et al. (2007), our method uses four fundamental emotions that have been identified as present in all cultures (Kemper, 1987): happiness, sadness, anger, and fear. Dayer et al. (2007) state that emotions are considered universal and that people in many cultures categorize emotions in similar ways (Dayer et al., 2007). The crosscultural nature of emotions reduces the possibility of conceptual confusion because all the respondents should recognize specific emotions. However, because the expression of emotion is culturally determined, each emotion may prompt very different stories related to wildlife from people in various cultures, depending on how the culture regards wildlife. The wildlife-related experiences that interviewees share through their personal stories are therefore expected to be very diverse. The word “Nyama” was used to refer to wildlife as it has the same meaning in Lingala and Swahili (and several other Bantu languages). The word “Nyama” indifferently means wild meat or wildlife. The interviewee was asked to tell a personal story (recent or in the past) about “Nyama.” Before telling the personal story, the interviewee was asked to pick up a colored ball which indicated whether the story was to be characterized by anger/happiness/fear or sadness depending on the color of the ball. This methodology always resulted in very friendly interviews, therefore overcoming the potential shyness of the interviewee or the dissimulation of the truth based on fears of stigma or lack of trust.

In addition, we asked the interviewees if they would be willing to respond to a short questionnaire on their socio-economic background, explaining the background of the research, the institutions involved, the type of questions that would be asked, and the potential use of the results. When the response was negative, we did not ask why to respect the free and informed decision of the interviewee. In total, 555 respondents agreed to participate to the second part of the interview. The socio-economic background was described by place of residence (rural/urban), gender, age (if age < 30 = young, if 30 < age < 50 = adult, if age > 60 = elder), ethnicity (Lokele, Mukusu, Nande, Mulengola, Mumanga, other), religion (catholique, église du reveil, protestant, musulman, other), profession/activity (hunter/fisherman, farmer, teacher, business owner, government employee, other), education (primary, secondary, none), wealth and nature relatedness. Wealth was a score based on the sum of the scores for type of house (concrete = 2, bricks = 1 or earth = 0), access to electricity (yes = 1; no = 0); access to tap water (yes = 1; no = 0); private transportation (car or moto = 1; bicycle or none = 0)); and household size (< 5 = 3; between 5 and 10 = 2; > 10 = 1). Household size was included as a variable determining wealth category because household size shows that larger households are associated with poverty in African contexts (Meyer & Nishimwe-Niyimbanira, 2016). Wealth scores were categorized into three groups from 1 to 3 (if score < 3, wealth = 1; if 3 < score < 6, then wealth = 2; or score > 5 wealth = 3), with 3 representing the wealthiest group. Three variables described nature relatedness: Hunt (if the person hunted once a day or once a week = 1, or less frequently = 0); Consume (if the persons consumed wild meat once a day or once a week = 1, or less frequently = 0); Forest (If the person went to the forest once a day or once a week = 1, or less frequently = 0).

Analysis

The analysis of a subset of recordings (100 randomly chosen within the dataset) was used to adapt the list of value orientations (VOs) from Dayer et al. (2007) and create a new, better-adapted list of value orientations regarding wild meat. In particular, we added Concern for health, which relates to the concern of interacting with wildlife because of the possibility of contracting diseases (e.g., Ebola), and Concern for crop/livestock destruction, which relates to the fact that wild animals destroy crops, kill small livestock or destroy fishing nets. We also added repugnance, which relates to the non-consumption value of wild meat due to the lack of hygiene (primarily due to its bad preservation status: rotten, dirty, exposed to flies). A regulation value was added to accounts because people’s values about wild meat are dictated by what is legal and what is not. Last, we added control over wildlife, linked to the capacity of humans to dominate their fear and bravely take control over the wilderness. The following VOs described in Dayer et al. (2007) were not represented in our sample: “Scientific,” explained as the belief that humans can solve any environmental problems by using science and technology, and “Rational,” related to scientific explanations about the way the natural world works, and the way animals behave, as opposed to spiritual or religious explanations. Based on the newly created list of value orientations, the stories from the whole sample were categorized into VOs (Table 1). A single-story could relate to more than one VO. We categorized the VOs elucidated in our study area into two major categories: 1. anthropocentric, which considers humans as the central fact, or final aim of the universe emphasizing the utilitarian or instrumental value of nature (Vaske & Donnelly, 1999); 2. biocentric, which attributes an inherent worth and life to environmental objects other than humans, regardless of their usefulness to humans (McFarlane & Boxall, 2000).

To investigate significant associations between multiple and diverse qualitative variables, we conducted the statistical analyses with 547 interviews, having removed interviews with incomplete answers or with singular observations. All analyses were conducted in R 3.6.3 (R Core Team, 2020). We used Multiple correspondence analysis (MCA) to explore possible associations between all variables in the FactoMineR (Le et al., 2008) and Factoextra (Kassambara & Mundt, 2020) packages. Differences in consumption of wild meat based on socio-demographic variables were evaluated with a G-test in the R-package DescTools (Andri et al., 2021).

Because respondents were asked to relate the story to the randomly chosen emotion, we tested for an association between emotions and the VOs with a Multiple Response Categorical Variables Analysis applying Bonferroni corrections in the MRCV package (Koziol & Bilder, 2014). Due to the significant relationship (P < 0.0001), we next applied generalized linear mixed models (GLMM) to predict VOs based on socio-demographic variables. For the three most common VOs, generalized linear models with VO as the binary response variable, binomial distribution, and the emotion as a random effect were run in the package glmmML (Brostrom, 2019). We selected the final most parsimonious model from a complete model based on the lowest AIC value.

Results

Less than 3% (N = 551) of the respondents consumed wild meat daily. Others ate it weekly (38%), monthly (27%), infrequently (29%) or never ate it (4%). Determining the variables that explain wild meat consumption was beyond the scope of this study; however, we found that wild meat consumption differed according to place of residence (urban = 29%; rural = 59%; G = 43.642, df = 1, p < 0.0001), ethnic background (Nande = 13%; Lokele = 45%, Mokuso = 39%, Mulengola = 48%, Mumanga = 69% and other = 31%; G = 44.248, df = 5, p < 0.0001), and wealth category (1 = 52%; 2 = 39%; 3 = 25%; G = 22.177, df = 2, p < 0.0001), but did not differ substantially according to gender (M = 40%; F = 40%; G = 0.0104, df = 1, p = 0.9189), age (elder = 33%, adults = 45%, young = 37%; G = 4.3852, df = 2, p = 0.1116), religion (Muslim = 29%; non-Muslim = 41%; G = 8.7785, df = 4, p = 0.0669) or education (none = 40%; primary = 40% and secondary = 42%; G = 0.2170, df = 2, p = 0.8972). Percentages indicate the percentage of respondents stating that they consume wild meat frequently (once a day or once a week).

Value Orientations Towards Wildlife/Wild Meat

The most frequent VOs encountered in our sample were Concern for safety (337 answers), Nutrition/taste (333 answers), and Caring/respect (105 answers). Some VOs related to “Nyama” in general, but others specifically related to certain species. Rodents and ungulates were mostly mentioned in VOs about Nutrition and Taste, Income and Concern for crop/livestock destruction, all promoting the consumptive use of those species. Small monkeys were mostly associated with Concern for health and Mutualism, both promoting the non-consumption of those species, for divergent reasons. Chimpanzees, elephants, and leopards, all vulnerable and emblematic species, were mostly mentioned in VOs about Concern for safety. Chimpanzees were also present in VOs about mutualism. Elephants were present in concern for crop/livestock destruction too. Okapi was mostly mentioned in Attraction/Interest and Identity.

Association of VOs with emotions

The first two dimensions in the MCA with emotions and VOs explained 22.5% of the variation. Happiness, Fear, and Sadness, together with the VOs Nutrition and Taste, Concern for Safety, Caring/respect, Regulations and Environmentalism were sufficiently well-explained by the first two dimensions (Cos2 > 0.2). Only significant associations yielded by the MRCV (Table 2/) were considered for further interpretation. Significant negative associations between VOs and emotion mean that a given VO was not or rarely mentioned in stories prompted by that emotion. For example, Nutrition and taste were negatively associated with anger and fear, which means that this VO was not described in stories prompted by those two emotions but instead mentioned in stories prompted by other emotions.

Mutualism and Caring/Respect were positively associated with sad stories; Nutrition and taste were associated with happiness; Concern for crop/Livestock destruction, Concern for safety, and Concern for health were more associated with anger or fear. Detailed results are presented in Table 2.

Determinants of Wildlife/Wild Meat Value Orientations

We tested the influence of place of residence, gender, age, ethnicity, religion, profession, education, wealth, and nature-relatedness on the VOs expressed by the interviewees. The multiple correspondence analysis explained 14.9% of the difference with the first two axes (F1 = 9.8% and F2 = 5.1%). The variables that contributed more to F1 were place of residence (v- value: rural = 20.4, urban = -20.4), forest (v-value: Forest-0 = -17.7; Forest-1 = 17.7) and wealth (v-value wealth-1 = 14.1; wealth-3 = -10.8). The variables that contributed most to F2 were gender (v-value: M = 13.6; F = -13.6) and hunt (v-value: hunt-0 = -10.9; hunt-1 = 10.9). Rural people, who use the forest frequently and belong to the least wealthy category, were concerned with Concern for crop/livestock destruction and Concern for safety, and Nutrition/taste.

In contrast, urban people from the wealthiest category were associated with Mutualism, Identity, Attraction for wildlife, Concern for health, and Repugnance for wild meat. The associations on F2 show that men were more likely to be associated with Income, Control over wildlife, Attraction/interest, Environmentalism, and Regulations. In contrast, women, who generally do not hunt, were associated with Nutrition and taste and with Taboos over wild meat. All other associations were non-significant on the two first axes.

Generalized linear mixed models, ran for the three most mentioned VOs, resulted in the following predictions (Table 3): Nutrition and taste can be predicted for non-Nande older women, no matter the place of residence (rural/urban). Safety concerns can be predicted for rural dwellers who are likely to consume wild meat frequently. Care/respect for wildlife can be predicted for young men.

Discussion

In this study, we identified the main value orientations (VOs) found in our target audience and the socio-economic profile of the people that were more likely to express the most frequent VOs. Our results from DRC, together with those from Rickenbach et al. (2017) and Chausson et al. (2019) in Congo, are the only available studies that elucidate value orientations over wild meat/wildlife in Central Africa, based on a human dimensions’ perspective. Our typology differentiated 14 VOs (as compared to 6 in Rickenbach et al., 2017). The most frequent VOs encountered in our sample were Concern for safety, Nutrition/taste, and Caring/respect. Place of residence (rural/urban), gender, age, wealth, and frequency of forest use and wildlife consumption best explained differences in VOs.

As in Rickenbach et al. (2017) we found that rural people were more likely to express concern for crop/livestock destruction or materialism VOs. This is probably explained by the fact that rural families depend on wild meat for food security and are the most exposed to damage to crop or livestock destruction caused by wildlife (Rickenbach et al., 2017). Therefore, our findings support Inglehart’s thesis that people who are more dependent on natural resources for fulfilling their needs have a stronger materialistic value orientation (Inglehart, 1997). This conclusion, however, should be taken with nuance because the dichotomy between biocentric and anthropocentric values is overly simplistic to understand the complexity of values that shape people’s attitudes and behaviors. Indeed, being rural and consuming wild meat does not necessarily associate negatively with caring/respect, suggesting that people who view wildlife through materialistic lenses may also be sensitive to biocentric values. Happiness was equally conveyed in stories related to Nutrition and taste and Attraction/interest. From a behavior change perspective, our results suggest that messages that acknowledge the fact that wildlife users also care about wildlife (e.g., “eating wildlife but conserving it too”) may be more efficient than those who criminalize them. In fact, findings from Bonwitt et al. (2018) within communities from Sierra Leone showed that the criminalization of wild meat consumption fueled fears and rumors, entrenching distrust towards public policies, with negative effects on compliance with the wild meat consumption recommendations.

Biocentric VOs were well explained by age and gender in our study area, with young men more likely to express care/respect. As opposed to results from Brazzaville (Chausson et al., 2019; Rickenbach et al., 2017), education level was not found to be a good predictor for biocentric VOs, and the frequency of wild meat consumption did not differ between men and women.

Whereas in Western contexts, there is evidence of the increase of mutualistic values in urban areas (Manfredo et al., 2017), mutualism does not dominate in Kisangani. If expressed, it is exclusively related to small monkeys or chimpanzees, species for which respondents also express anger due to concern for safety and concern for crop/livestock destruction. Management decisions and messaging in behavioral change campaigns should consider that concern for safety and concern for crop destruction are among the values that represent significant barriers to change behaviors towards the killing of wildlife, and those VOs are associated with anger and fear. Humanizing animals, a strategy often used to generate compassion or affinity, may generate a rejection response in our study site if it ignores why people fear wildlife. We would therefore caution about using compassion conservation or anthropomorphic attributions to wildlife in the context of our study region.

Although this was not the focus of our study, we also found that different wild meat species associated differently with different VOs. Our methods based on personal stories from the interviewees allowed relating specific wildlife species with different VOs and emotions. From a behavior change perspective, this may imply that different VOs need to be considered for species-specific conservation. Different species might be used in behavior change messages depending on the desired change.

As used in this study, we argue that the human dimensions approach provides crucial information to understand differences in value orientations across groups, identify barriers to change, and tailor behavior change campaigns to the local context. In particular, the type of associations shown in this study allows to clarify the target for behavior change campaigns and elucidate positive images, messages, and emotions that can be used to create empathy in social marketing campaigns. We do, however, acknowledge that demand reduction strategies cannot achieve tangible impacts on their own. Systems that encourage the participation of civil society, equal access to conservation benefits, rights to land and resources and access to funding and alternatives to diversify from the use of natural resources are required to develop efficient conservation strategies that account for the multiple motivators that fuel the wild meat sector and reduce the barriers to change for more sustainable practices (Thomas‐Walters et al., 2020).

Table 1 Anthropocentric Value Orientations and associations between VOs and specific wildlife species. Biocentric Value Orientations and associations between VOs and specific wildlife species
Full size table
Table 2 Results of the MRCV analysis showing positive (+) and negative (-) associations between the emotions of the stories and the VOs expressed, with statistical significance at the P<0.001 (***), P<0.01 (**) and P<0.05 (*) level.
Full size table
Table 3 Result of GLMM for the most common VOs with socio-economic variables of gender, age, urban versus rural sites, ethnicity, religion, wealth, forest visit and wildmeat consumption. Only the model with the lowest AIC is presented here with the coefficient (Coef), standard error (SE) and P-value (P). For categorical variables, the referent (ref) is indicated. Sample size (n) is provided for all categories.
Full size table