Introduction

In many parts of the world, wildlife poaching has reached crisis levels and is a high-priority policy issue [1]. Conservationists often define poaching as the illegal taking of wildlife in violation of laws or rules [2]. However, different sectors and disciplines acknowledge nuanced variations in the definition of poaching. Many of these differences revolve around the motivation or intent underlying the behavior. It is important to note that there is a difference between wildlife that is held in public trust, as is the case in the United States, and wild animals that are privately owned, for example by private game ranches in South Africa. Thus, poaching can and does often refer to both the illegal killing of otherwise protected animals and the unlawful exploitation of wildlife. Moreto and Lemieux [3] provide a useful description of the various activities that may be subsumed under this umbrella term, and propose a crime specific approach which recognizes the essential steps required for the completion of a criminal act, which in term can help identify entry points for intervention. Given this crime specific approach, we adhere to Eliason’s [4] definition, which has since been applied by Crow et al. [5], Filteau [6] and Moreto and Clarke [7]. Our definitional approach informed the format and content of our data collection instruments as well as our analysis.

Poaching threatens the viability of protected areas and species conservation in almost every biome [8]. Alongside other forms of biodiversity exploitation and environmental crime, such as illegal commercial logging and illegal unreported and unregulated (IUU) fishing, poaching can be politically destabilizing, subverts the rule of law, undermines sustainable development, and generates proceeds that fuel other organized crimes and conflict [9]. Furthermore, poaching can detrimentally affect the livelihoods of people living in proximity to wildlife 10]. Research on illegal resource use seeks to understand what species are being extracted, where illegal resource use transpires, who uses resources illegally, and why illegal use occurs [8]. This research has demonstrated that wildlife populations are not uniformly vulnerable to poaching [11]. Species biology, distribution, and interactions with human socio-economic systems can all impact susceptibility to poaching [12]. For example, the vulnerability of Bolivian parrots (Family Psittacidae) to poaching for the pet trade is significantly higher for species found within 80 km of a city and for animals from relatively more abundant populations [13]. Species vulnerability to poaching can also differ according to human motivations to poach. Animals may be poached because of negative attitudes about human-wildlife conflict [14], avoidance of future human-wildlife conflicts [15], or economics [12]. Although some scholars are recognizing limits to focusing solely on why environmental crimes occur as opposed to how [16], understanding the incentives and motivations of poachers remains key for discouraging illegal resource use on-the-ground [17].

Reducing the risks associated with poaching, such as biodiversity loss, defaunation, and detrimental livelihood impacts, necessitates understanding the sociocultural context that informs the risk perceptions of poachers and local communities. The way in which a species interacts with humans influences vulnerability to poaching [12], and human perceptions of poaching-related risks inform attitudes toward the causes and consequences of poaching. For example, risk perceptions affect attitudes toward human-wildlife conflict (a potential cause of poaching) and can be used to evaluate stakeholder support for different management options aimed at reducing poaching [18]. However, scholars have yet to apply a cultural theory of risk to attitudes toward poaching. Our paper addressed this gap in order to explore the potential of cultural risk theory for predicting poaching-related attitudes. Our analysis also responds to the emphasis in the green criminological literature on the social construction of crime [19]. We use a theoretical framework rooted in sociology in order to offer a nuanced examination of how various social institutions influence the social construction of poaching. We examined how cultural institutions (of various sizes and types) respond to poaching and how these responses influence individual risk perceptions of poaching. Madagascar is an ideal site to apply cultural risk theory to poaching since: (a) it is a center of biodiversity threatened by rampant poaching and corruption [20], (b) it has distinct cultural institutions such as the fokonolona that predate colonialism and that collaborate with post-colonial institutions in natural resource management, and (c) social norms are known to tangibly impact wildlife protection in Madagascar [21]. Our objectives were to: (a) explore beliefs about the poaching rules and responses promulgated by four social institutions (the family, village, fokonolona, and Madagascar National Parks or MNP), (b) examine the effects of these social institutions’ responses to poaching on motivations to and not to poach, and (c) analyze the effects of these social institutions’ responses to poaching on the perceived impacts of poaching on people and the environment.

Background

Social scientific work on poaching

Social scientists have been characterizing the types, motivations, and rationalizations of poachers for decades. Wildlife poaching has been increasingly criminalized (wildlife poaching is now one activity that can be criminalized under the United States Racketeer Influenced and Corrupt Organizations Act), but this activity has long been normatively considered wrong or immoral by various community-level institutions, such as the Malagasy fokonolona (described in more detail below). Social science has studied poaching from both a crime science perspective [3] and a normative perspective [22]. The latter perspective has focused on the lack of cultural acceptance of certain activities as crimes irrespective of what the rule of law states, particularly when research has transpired in a developing nation where there is a relatively weak environmental rule of law, such as Madagascar.

The extant knowledge base draws from the fields of psychology, economics, political science, criminology, and sociology to tell us that poachers engage in illegal activity for a variety of reasons including lack of rule awareness, carelessness, recreation, trophy acquisition, tradition, monetary profit, food, protection of self or property, and disagreement with regulations or law enforcement [2, 23–25]. Poachers can broadly rationalize their behavior through denying injury (i.e., framing poaching as a victimless crime), defending it as a necessity, or condemning the condemners (i.e., evoking negative views of law enforcement). These rationalization and neutralization techniques may be used before, during, and after poaching [26]. Importantly, there is a growing body of applied research on wildlife poaching, where researchers are engaging directly with non-compliance behaviors or the specific behaviors that poachers commit [18, 27]. This body of literature compliments practical law enforcement experience and allows for an applied understanding of poaching within a broader environmental threat matrix [28].

Despite a substantial knowledge base, those tasked with wildlife protection still face challenges in reducing the risks associated with poaching. Whereas in the past poaching was viewed myopically as a key driver of biodiversity loss and therefore appropriately addressed by conservation biology alone, today it is understood as a cause and a consequence of environmental insecurity and thus under the purview of a broad array of public, private, non-governmental, civil society, military, and academic sectors. The extant poaching literature has been criticized as an overly static way of profiling criminals and predicting crime, since the categorization of environmental crime is contingent on the social reaction to the crime as much as the motives of the offender. Furthermore, poachers adapt their behaviors in response to the norms and responses of their culture [29].

Cultural perspectives on poaching risks

Cultural factors have the potential to affect both poaching practices and societal responses to poaching. A poacher’s relationship to a community and community norms impacts whether the act of poaching is seen as a “crime” and whether the community responds with tolerance, sanctions, or some combination thereof [30–32]. Data on tree theft in the Pacific Northwest [30], illegal fishing in New Jersey [31] and Lithuania [33], and wildlife poaching in Louisiana [34] and Greece [32] demonstrate that when a poacher is part of a community and is using poaching to express traditional land rights or adherence to the local culture, poaching may be tolerated and may even help sustain community identity and cohesion. In contrast, acts of poaching that violate community norms or are committed by community “outsiders” are often labeled as crimes [23, 30, 31]. Poachers are typically “recruited” into the practice by people from within their peer group and then become integrated into a culture or sub-culture where the violation of wildlife laws is normalized [35]. During the process of being socialized into a culture of illegal behavior, poachers often learn neutralization techniques that allow them to cognitively justify their socially deviant behavior [36].

Cultural theory of risk and group-grid theory

Cultural risk theory offers one mechanism for measuring and interpreting the cultural dimensions of poaching; it proposes that risk perceptions, such as perceptions of the risks posed to wildlife and people from poaching, are shaped by social context and culture [33, 37, 38]. Cultural theory relies heavily on the group-grid concept developed by Mary Douglas [39]. Group-grid summarizes a type of culture based on two dimensions: the level of affiliation between individuals (i.e., group) and the prescriptiveness of norms (i.e., grid) [40]. Theoretically, in cultures where individuals are more incorporated into bonded units or groups, groups have a stronger influence on individual risk perceptions [41] and thus behaviors. Within the wildlife poaching context, this could be exemplified as community members complying with rules against poaching when the community is considered “high group.” “Grid” designates the network of rules and sanctions that affect individuals in the course of their interactions [42]. Rules may be sharply delineated and enforced or, at the other extreme, may be absent. The intersection of group and grid creates four distinct cultural types: hierarchical (i.e., high group, high grid), equalitarianism (i.e., high group, low grid), fatalism (i.e., low group, high grid), and individualism (i.e., low group, low grid) [43]. Accordingly, these four cultural types are thought to hold specific views of nature [41, 42] and be concerned about different forms of risk [37] (Fig. 1).

Fig. 1
figure 1

Cultural types and corresponding myths of nature and risk perceptions

Full size image

These cultural types may also view regulatory responses to risk differently, which can fuel the discrepancy between regulators and the local community. Regulatory responses may take the form of risk management interventions, deterrence activities, or outreach and education. Specifically, perceptions about risk regulation among local people may not mirror those of regulators, and vice versa, when regulators are educated and socialized to risk-related regulatory strategies in foreign countries. Understanding how local communities perceive risk, and how their risk perceptions differ from expert risk perceptions, can improve the design and implementation of regulation and enforcement strategies in biodiversity hotspots threatened by environmental crime [44]. However, we recognize the considerable criminological literature that casts doubt on a linear relationship between deterrence and compliance, particularly within the wildlife-poaching domain. Wellsmith [45] discusses the lack of deterrent effect as a significant problem hampering effective enforcement in the wildlife conservation arena. The requirements for deterrent effects do not seem to be readily met for wildlife crimes. Thus, enforcement is best viewed as one part of the broader risk management solution [45]. Our research considers the extent to which multiple and different types of risk managers and/or authorities compare in helping to address the risks of wildlife poaching.

The cultural theory of risk is important to consider within the broader thesis of Ulrich Beck’s Risk Society [46], where risk is the dominant paradigm of postmodern society. The extent to which wildlife poaching fits within the framework of the risk society is not wholly clear. Many of the countries richest in biodiversity are developing and arguably may not be accurately categorized as fully postmodern. This research aimed to capture local people’s perceptions of risk in order to explore the extent to which a postmodern paradigm could be explored in a post-colonial society that retains pre-colonial institutions (such as the fokonolona).

Cultural risk theory has been applied to environmental risk contexts such as water pollution [47] and genetically modified food [48], and has helped predict public response to management strategies to reduce risk exposure. However, cultural theory has yet to be applied to wildlife poaching or other environmental crimes. This is a significant gap in knowledge for at least two reasons. First, when social bonds, or relationships between individuals and groups, are not taken into account, anti-poaching strategies (e.g., enforcement, sanctions, education) can result in unintended inverse effects. When an offender does not see the sanctioning agent as legitimate and/or fair, and there are weak social bonds between them, enforcement can result in defiance (and additional crimes) rather than deterrence [29, 49]. For example, retaliatory killing of carnivores in Tanzania has been found to be a function of both biological (e.g., nocturnal predation) and sociocultural factors [11]. In Madagascar, perceptions of conservation laws and authorities as unfair and biased towards the wealthy have been associated with noncompliance with conservation rules [50].

Secondly, poachers operate not in isolation, but in context. As Beck [46] noted in his Risk Society thesis, nature is culturally constructed; thus, to ignore the cultural specificity of wildlife-related risks is to neglect a key component of the anti-poaching equation. In Madagascar, the cultural context has produced “rules in use” that govern local interactions with wildlife and delineate unique notions of culpability, enforcement, and sanctions that can differ from the rule of law [50]. Many of these rules in use affect conservation even though they did not originate from attempts to conserve wildlife but rather from the fokonolona, a pre-colonial institution in Madagascar that engages in traditional, administrative, and environmental policing [51]. Understanding “rules in use” can help explain the lack of cultural acceptance of certain activities as crimes despite their labeling as “crimes” by the rule of law [50].

Rules in use are particularly prominent in developing nations, such as Madagascar, where there can be weak environmental laws. Despite its small geographic size, Madagascar plays a disproportionate role in the illegal wildlife market and extremely high rates of environmental exploitation and corruption. Rules in use are in essence the prescriptiveness of norms (or “grid”) referenced by cultural risk theory. This theory offers the potential to explore how the norms of different cultural institutions in Madagascar affect risk perceptions towards poaching, to contextualize these risk perceptions within particular myths of nature and cultural types, and to offer insight into how social affiliation (the “group” of cultural risk theory) can promote or hinder poaching deterrence in Madagascar. Green criminologists have detailed the extensive harm that illegal hunting and wildlife trafficking exacts on both ecosystems and on individual animals [52, 53]. Addressing these issues requires a criminological analysis that integrates local, site-specific data [54] and an understanding of the social dynamics that perpetuate environmental harms; this approach “demands that environmental issues be framed within the context of a sociological and socio-criminological imagination” [55]. In addressing the social construction of crime and the cultural significance of the environment, green criminologists have described how the media (a social institution) plays a key role in constructing and framing environmental harm [19, 56]. Through the application of cultural risk theory to poaching, we offer a site-specific and sociologically-based analysis of the intersecting contributions of various social institutions to the social construction of crime.

Case study site: Madagascar

Environmental crime in Madagascar

We selected Madagascar as a case study site to achieve our objectives; case studies are intensive studies of specific situations that permit a researcher to use a diverse array of methods and theories [57]. Madagascar is a global center of biodiversity; 75% of its species are endemic (i.e., found no place else on Earth). However, the flora and fauna of this island face numerous anthropogenic risks such as illegal logging and unsustainable hunting [20, 58]. Madagascar has been rated as having extremely high rates of environmental exploitation and natural resource trafficking, particularly wildlife and timber products [59, 60]. Considering its relatively small land area, it is remarkable that Madagascar is estimated to supply 2.5% of the world’s live reptile trade and has one of the highest rated trade volumes in the world [61]. Illegal natural resource exploitation and wildlife trafficking in Madagascar are at least in part a function of extremely rare species in high demand, corruption, high poverty rates, poor governance and prolific natural resources [1, 52, 62]. Three-quarters of Malagasy live in poverty, only 35% of the population has access to an improved water source, and the Gross National Income of $420 USD/year ranks Madagascar 178 of 184 nations [63]. Since experiencing a political crisis in March 2009, Madagascar has experienced an upsurge in resource degrading behaviors, including increases in the illegal harvesting of endangered hardwoods such as rosewood, mining for gemstones such as sapphires, and trafficking in animals such as the ploughshare tortoise for the pet trade [64]. For example, in 2012, over 550 critically endangered radiated tortoises were confiscated at the Ivato airport [65]. In 2014, authorities in Singapore intercepted a shipment of over 29,000 illegal rosewood logs from Madagascar, the largest international seizure of its kind [66]. International’s 2014 Corruption Perception Index gave Madagascar a 28/100, ranking it 133 out of 175 countries in transparency [67]; the 2013 Global Corruption Barometer survey reported only 2% of the Malagasy population sampled in a national survey felt that the level of corruption has decreased substantially over time.

These trends paint a bleak picture for Madagascar’s degraded environment and the people whose well-being depends on natural resources [59, 68]. Because many of the wildlife involved in trafficking originate in biodiversity-rich developing countries such as Madagascar, insights from research conducted there may have practical implications for other biodiversity hotspots around the world. In Madagascar and elsewhere, poaching can be a highly localized activity that directly feeds into the global wildlife trafficking supply chain.

Cultural institutions and conservation in Madagascar

Currently, various institutions address wildlife poaching and conservation risks in Madagascar, including the MNP and the fokonolona, or a group of people living within the same indigenous community [51]. In response to high rates of deforestation and unsuccessful federal conservation protection practices, Madagascar has attempted to decentralize forest management by transferring some conservation enforcement to local people and integrating modern (i.e., state-based) and traditional (i.e., fokonolona-based) law. For example, an effort was made to amend dinas, traditional fokonolona regulations governing community assets, for integration into state-based forestry regulations [69]. In 1997, the Malagasy government adopted a community-based natural resource policy, known as GELOSE, that governs forests, water sources, and wildlife. GELOSE allowed for the negotiation of contracts between the state (i.e., the forest service) and a voluntary association of community residents (i.e., the fokonolona). Under these regulated contracts, local communities manage natural resources through dinas [70]. However, there is still much centralized control of wildlife, as the terms of the contracts are heavily influenced by the state and by international non-governmental organizations (NGOs) [69, 70].

The fokonolona may create and enforce rules and sanction offenders autonomously from but in concert with the federal government. All members of the fokonolona are born and live in the same communal territory. Migrants who establish themselves in that territory by marriage or who have their descendants born into the community are considered to be part of the fokonolona. The fokonolona is the supervisor of all natural resource management zones within its geographic area. Within this geographic space, the fokonolona has the power to recognize new or different zones within its territory that are beholden to traditional and legal norms of the fokonolona [51]. The fokonolona also engages in community-level policing of conservation and non-conservation rules [51], whereas the MNP is specifically tasked with conservation rules. At the local level in Madagascar, fady (i.e., taboos or rules in use) range considerably, but often include temporal, life history, and species-specific restrictions on the taking of wildlife. For example, tailless tenrecs should only be hunted after reproduction but before hibernation and eastern woolly lemurs, considered by some Malagasy to be ancestors in animal form, should not be hunted (Jones et al. 2008). These taboos demonstrate how the conceptualization of animal victimhood is socially constructed. Although our survey measured generalized attitudes towards poaching in order to standardize responses, information about individual fady can be found in Jones et al. [21]. While many of these social norms originate from taboos unconnected to conservation concerns, they have been shown to have tangible effects on the protection of wildlife in Madagascar [21].

Madagascar is an ideal location for applying a cultural theory of risk to poaching since: (a) it is a biodiversity “hotspot” with unique species such as lemurs threatened by poaching [20], (b) it has distinct cultural institutions such as the fokonolona, and (c) the strong role of culture in conservation among the Malagasy people is well known [21]. Within Madagascar, our study site was the broader Analamazaotra Special Reserve region of east central Madagascar and we presented our affiliation as being with Michigan State University. We selected Analamazaotra based on our collective knowledge about environmental crime in Madagascar, permission of relevant traditional authorities, willingness of communities to participate, available research resources, and accessibility (Fig. 2).

Fig. 2
figure 2

Map of study area within the Analamazaotra Special Reserve, east-central Madagascar

Full size image

Methodology

Ethics statement

Michigan State University’s Committee on Research Involving Human Subjects IRB# ×10–394 reviewed and approved all methods and procedures used in this research. A committee-approved informed consent was obtained in verbal form due to potential participant illiteracy. In instances where participants approved use of a digital voice recorder, consent was documented digitally. In all instances, participants had to verbally consent to participate in the study before data collection commenced.

Data collection

We used quantitative methods to achieve our research objectives. Quantitative inquiry is one valid mechanism for the application of a cultural theory of risk to poaching. Quantitative methods have been widely used and validated in the literature on cross-cultural risk perception research [71–73]. Although our methodology is consistent with the risk perception literature, we acknowledge our study does not provide elaborate descriptive and contextual information underlying participants’ responses to survey questions. Future research might apply qualitative methods to inquiry, and would likely produce complimentary insight concerning the cultural dimensions of risk perception explored herein. Using a five-member research team that consisted of three Caucasian female Americans and two male Malagasy individuals, we conducted quantitative surveys with Malagasy residents of Analamazaotra in December 2013. Our survey was based on five years of collaborative risk perception research and teaching experience in Madagascar and the United States. Many study questions have since been replicated and published [44]. Within Analamazaotra, we used purposeful sampling to intentionally recruit participants at least 18 years of age within 8 villages who had experience with the central concepts of our research questions in order to capture diverse perspectives [74]. Because we were interested in exploring conceptual ideas and not generalizing results to the entire Malagasy population, we did not use a probability sample to collect data. Opportunity sampling is pertinent when researchers are attempting to gain access to potentially difficult data sources (e.g., geographically remote populations) and have a limited timeframe with which to collect data. The technique relies on the local knowledge and attributes of the researcher to identify a sample. We minimized threats to external validity (i.e., the generalizability of results) by not connecting results to the broader population or claiming representativeness of results. We maximized internal validity (i.e., truthfulness of results) by posing various questions pertaining to the same concept and calculating reliability coefficients using statistical software [57]. Our approach was intended to capture as much detail as possible about the complex phenomena of environmental crime in Madagascar [75].

Since poaching is a potentially sensitive topic, we developed our methodology in consultation with the literature on data collection of sensitive information [27, 76]. We minimized the potential for response bias to sensitive issues by utilizing established techniques such as not anchoring questions directly to the respondent; during the survey, we notified participants that we were not interested in their personal behavior, but rather knowledge of other people’s behavior. The survey was primarily administered in Malagasy with some French and English, accommodated dialectical variation based on ethnicity, and promoted high translator autonomy. Prior to commencing data collection, three Malagasy conservation researchers independently reviewed the survey guide for content validity and salience. In-field instrument translation from English into Malagasy occurred during face-to-face discussions among all research team members before data collection commenced. These discussions involved word-by-word verbal translation of all terms; the team had French, English, and Malagasy dictionaries at their disposal and came to consensus on the translation of questions. After defining study objectives, introducing visual aids to participants, and discussing affiliation, anticipated research outcomes, intended use of data, and definitions of key issues in surveys (e.g., poaching versus legal hunting), informed consent was obtained. The survey had two parts: a relative risk-ranking activity where participants were asked to hierarchically arrange photographs of eight environmental risks and quantitative four-point survey questions.

To maximize our capacity to conduct statistical analysis on cross-cultural risk perception data, we used a quantitative visual 4-point scale to record answers [77]. Respondents were asked their level of agreement with statements such as “the fokonolona deters poaching in the area” and used the visual scale to indicate their answer (see Fig. 3). Methodological scholarship on risk perception has determined that quantitative surveys with limited (fewer than 7) response categories produce valid, consistent data on attitudes towards risk while reducing the risk of surveyor bias [71]. Cross-cultural measurement of human attitudes towards wildlife can pose a number of challenges, such as respondent illiteracy and language barriers between researchers and respondents; visual scales can help mitigate these methodological challenges [77]. For example, visual scales reduce the potential for culturally influenced bias toward neutral or extreme response categories [78, 79] and address the limitations that traditional Likert scales have for non-readers [80].

Fig. 3
figure 3

The quantitative four-point visual scale used in our survey. Respondents indicated their level of agreement with each survey question: none (0), low (1), medium (2), or high (3)

Full size image

Measurement

Our survey asked respondents about:

  1. (a)

    Awareness of rules and laws against poaching, informed by psychometric theories of risk perception stating that risk perception is influenced by a variety of factors, including awareness [81, 82];

  2. (b)

    The actions (i.e., ignore, deter, adapt, welcome) of four cultural institutions (i.e., family, village, fokonolona, and MNP, see Fig. 4) in response to poaching, informed by cultural risk theory’s proposal about variation in public opinion of risk management responses [83];

  3. (c)

    The perceived effects of poaching (i.e., predictable, unknown, random) on the environment and local people, informed by cultural risk theory’s mythologies of nature [39]; and

  4. (d)

    Motivations to poach and motivations not to poach (motivations to poach and motivations not to poach are two separate variables (see [25]). Motivations to poach included food, income, fun, poaching trees for homes, poaching animals for pets, and poaching to protest rules. Motivations not to poach included fear of getting caught, the strong enforcement of local rules, severe punishment, respect for rules, shame at getting caught, fair enforcement of rules, and the belief that rules are doing the right thing.

Fig. 4
figure 4

Cultural institutions in Madagascar with varying degrees of authority to manage natural resources

Full size image

The survey instrument was developed in English, translated into Malagasy, verbally translated back into English and conducted in Malagasy. This approach has worked successfully for us in the past in similar research contexts [77]. Most questions were easily translated into Malagasy. When words were unclear (e.g., politics, “better off”) the American field researcher provided more specific definitions and synonyms, the two Malagasy researchers offered ideas and all three reached an agreement on the appropriate word choice for the survey. Conceptual triangulation was achieved through team decision-making to ensure survey clarity and cross-checks of translation [84]. The field team debriefed each evening after surveys concluded, continuously reevaluating the consistently of methods to maintain reliability and validity [85]. The team used the same protocol to maximize consistency and minimize the potential for surveyor bias that might occur through inconsistent order or wording of questions [86, 87]. A three-year history of research collaboration among the field research team aided the validity and reliability of data collection along with use of visual scales and aids.

Analysis

Responses were recorded on paper and entered into STATA 13.0 for analysis. Parametric statistics were used because our tests [88] for skewness and kurtosis fell within boundaries of a normal distribution. We calculated normality using the Shapiro-Wilk statistic [88]. Normality is indicated by closeness to one. This statistic requires that the sample size is greater than or equal to 7 and less than or equal to 2000. We first used bivariate correlations [89] to examine the relationships between the institutional response variables (i.e., whether each institution was perceived to be deterring or ignoring poaching) and motivations to poach. Bivariate correlations were repeated with motivations not to poach and questions about the effects of poaching on the environment and local people. Multiple variables were correlated with the beliefs that the effect of poaching on the environment is predictable and the effect of poaching on people is predicable. For these variables, we conducted progressive adjustment using nested multivariate ordinal regression models to determine which model best predicts these variables [90]. We structured our progressive adjustment so that fokonolona variables preceded MNP variables. The fokonolona predates the MNP by hundreds of years and the rules of the fokonolona encompass but are not limited to environmental crime. Thus, we theorized that the fokonolona would yield more robust prescriptiveness of norms than the MNP, and that the MNP would be the intermediary institution (or mediator in statistical terms) of the effects of the fokonolona. We used the likelihood-ratio (LR) test to determine if adding a new deterrence variable improved the previous model [89]. The LR test requires that all nested models have the same N; therefore, we dropped any cases that were missing data on the dependent or independent variables. In all instances of ordinal regression discussed in this paper, the demographic variables of age, gender and years lived in area were not statistically significant and an LR test indicated that adding these demographic variables to a null model did not improve the model; therefore these variables were not included in the final analyses [89].

Results

Sample

Our instrument captured quantitative data and our analyses were quantitative in nature. Thus, our results are not presented with the substantial breath and description that typically accompanies qualitative data. Our results do, however, offer measurement precision and reliability of data. We conducted 49 surveys and encountered 2 response refusals; not all participants answered all questions; this suggests that participants did not feel pressured to answer potentially sensitive questions [57]. The sample included more males than females (67% male, n = 33, 33% female, n = 16) and the average age was 40.9 years (range = 19 to 77). At least 6 ethnic groups were denoted in our sample, with Betsimisaraka (n = 28, 56%) and Bezanozano (n = 13, 26%) most commonly represented. Respondents reported an average of 4.5 children (range = 0 to 12). The mean length of time residing in the area was 28.8 years (range = 1 to 77). Respondents reported 13 different professions, with farmer (n = 27, 54%) and local guide (n = 5, 10%) most frequently noted (Table 1).

Table 1 Professions of respondents
Full size table

Objective 1: beliefs about poaching rules at different group levels

Our first objective was to explore cultural beliefs about laws and rules affecting poaching at four different levels of institutional affiliation: family, village, fokonolona, and MNP. The majority of respondents agreed that people in their community were aware both of Malagasy laws (75%, n = 36) and fokonolona rules (84%, n = 41) associated with poaching. However, when we asked respondents about the different responses to poaching demonstrated by their family, village, the fokonolona and MNP, differences emerged according to the level of group affiliation. The majority of respondents (63%, n = 29) agreed that the MNP actively deterred or discouraged poaching in the area. However, only approximately one third of respondents agreed that the family (32%, n = 16), village (36%, n = 18), or fokonolona (38%, n = 19) deterred poaching. Therefore, although participants demonstrated mostly consistent knowledge about the laws and rules (i.e., the grid) against poaching, there was variation in perceptions of institutional responses to poaching. Results do not indicate that participants believed these institutions ignore, overlook, or discount poaching. Over half of the respondents disagreed with the assertion that the fokonolona (57%, n = 27), village (62%, n = 29), and family (64%, n = 29) ignore poaching. Simply, these institutions are perceived as not being as actively engaged in deterring poaching as the MNP.

Objective 2: effects of social institutions on deterrence

Our second objective was to examine the effects of social institutions’ responses to poaching on motivations to and not to poach. Participants indicated institutional responses were associated with different deterrence beliefs. Respondents who agreed that the MNP deterred poaching thought that people poach for food (r = .62, p < .01) or income (r = .49, p < .01), whereas respondents who agreed that the village deterred poaching believed that people poach for fun (r = .13, p <. 05).

Awareness of Malagasy laws and fokonolona rules had different effects on perceptions of deterrence. There was a positive correlation between awareness of Malagasy poaching laws and agreement with the statement “people do not poach because local rules are well enforced” (r = .26, p < .05) but no such correlation emerged for awareness of fokonolona rules. Furthermore, whereas awareness of Malagasy poaching laws had a positive correlation with agreement that people do not poach because they respect local rules (r = .30, p < .1), awareness of fokonolona rules had a negative correlation with agreement that people do not poach because they respect local rules (r = −.05, p < .05). Although deterrence of poaching based on enforcement of and respect for rules differed by institution, deterrence vis á vis fear of being caught was homogenous across institutions. Belief that any level of group affiliation (e.g., family, village, fokonolona, and MNP) was deterring poaching was associated with deterrence vis á vis fear of being caught (Table 2). These results can be interpreted as meaning that, while the aversion to being caught is a universal deterrent across social institutions, respect for rules (and any associated deterrent effect) depends upon the source of the rule.

Table 2 Bivariate correlations between institutional deterrence and perceptions that people do not poach due to fear of getting caught from analyses of community responses in Analamazaotra Special Reserve, Madagascar
Full size table

Objective 3: effects of societal institutions on perceptions of how poaching impacts people and the environment

Our third objective was to examine the effects of social institutions’ responses to poaching on perceived impacts of poaching on people and the environment. The majority of participants agreed that poaching posed a threat to the environment (73%, n = 36) and the lives of local people (63%, n = 31). Perceptions of institutional responses to poaching affected beliefs about the predictability of poaching impacts (Tables 3 and 4). Analysis indicated that among study respondents, (a) MNP deterrence had a robust effect on perceptions of how poaching affects nature and (b) actions of the fokonolona had a stronger effect on beliefs about the risks of poaching to people than on beliefs about risks of poaching to the environment.

Table 3 Odds ratios of higher agreement with the effects of poaching on nature are predictable from ordered logit analyses of community responses in Analamazaotra Special Reserve, Madagascar
Full size table
Table 4 Odds ratios of higher agreement with the effects of poaching on local people are predictable from ordered logit analyses of community responses in Analamazaotra Special Reserve, Madagascar
Full size table

The most parsimonious model for predicting the belief that poaching risks on the environment are foreseeable included fokonolona and MNP-based deterrence (Table 3, Model 2). Our nested models demonstrated that the effect of fokonolona deterrence on this belief was mediated through the deterrence actions of the MNP (the actions of the MNP were the mechanism through which fokonolona deterrence operated). Thus, our analysis indicated robust deterrence effects of the MNP on perceptions of risks from poaching to the environment. Even after controlling for fokonolona-based deterrence, a respondent who more strongly agreed that the MNP deterred poaching was 125% more likely to agree that poaching risks on the environment were predictable (Table 3, Model 2).

Perceptions about poaching-related responses from the fokonolona had a more direct effect on beliefs about the predictability of the risk poaching posed to people. The best model for predicting this belief incorporated three variables: perceptions about whether the fokonolona ignored poaching, whether the fokonolona deterred poaching, and whether the MNP deterred poaching (Table 4, Model 3). In this saturated model (Table 4, Model 3), the effects of the fokonolona remained significant; a respondent who strongly agreed that the fokonolona ignored poaching was 55% less likely to think that the effects of poaching on people lives are predictable. This result implied that when the fokonolona responded to poaching, respondents were more likely to think that poaching risks to people are predictable. The deterrence actions of the MNP also had a significant effect on perceptions of the predictability of poaching risks on people (Table 4, Model 3). Collectively, results indicated that the institutional response of the MNP strongly influenced beliefs about the predictability of poaching risks on the environment, whereas both the MNP and the fokonolona influence perceptions of the predictability of poaching risks on people.

Discussion

As biodiversity exploitation and other forms of environmental crime assume higher priority positions on policy agendas, it is increasingly important to examine the causes and consequences of poaching [1]. Green criminologists have begun to recognize how the integration of cultural analysis can enrich our understanding of environmental crime, with some even calling for a green-cultural criminology [56]. This work discusses how cultural institutions such as the media influence the public’s perceptions of and responses to environmental crime. For example, environmental harms ignored by the media and other “institutional sponsors” often fail to be socially constructed as environmental problems [19, 56]. Our work expands upon this scholarship by examining how different social institutions converge and diverge in their institutional sponsorship of poaching as an environmental problem that demands deterrence. We also offer novel insight about the role of culture, specifically group affiliation and prescriptiveness of norms, in poaching. Results confirm that culturally-informed perceptions of institutional responses to poaching risks can have individual-level effects. Our work also extends cultural risk theory by illustrating that the prescriptiveness of norms is not a one-dimensional construct, but rather can vary according to the type of group involved in prescription. In confirming the relevance of cultural risk theory to poaching, we identify an opportunity to broaden our thinking about theory and practice related to poaching and other environmental risks. Below, we discuss the most noteworthy findings from this work and their implications; we begin our discussion by focusing on conservation practice.

First, in our case study of Madagascar, it appears that the fokonolona and the MNP conform to different cultural types and thus exert distinctive influences on individual risk perceptions of poaching. Congruent with Beck’s [46] hypothesis that postmodernity is characterized by risk, risk was a dominant paradigm among our respondents; the majority of participants agreed that poaching posed a threat to the environment (73%, n = 36) and the lives of local people (63%, n = 31). Although Madagascar is developing and arguably may not be conceptualized as fully postmodern, its status as a site of endemic and threatened species attracted international attention beginning in the 1980s. Global institutions such as environmental NGOs exert a considerable influence on Madagascar’s contemporary wildlife policies and have likely influenced the risk paradigms assumed by national organizations such as the MNP [70].

Our results indicate that, in Madagascar, pre-colonial (i.e., fokonolona) and post-colonial (i.e., MNP) institutions promote different cultural paradigms of risk. In cultures where individuals are more incorporated into bonded units or groups, what cultural risk theory refers to as “high group,” groups strongly influence individual risk perceptions [41]. Both the MNP and the fokonolona exerted statistically significant effects on individual risk perceptions of poaching and thus can be considered to be high group cultural institutions. However, these institutions differed along the grid dimension, also known as the prescriptiveness of norms. Based on our survey data, the MNP can be classified as a high group and high grid, or hierarchical, cultural type. A high grid is classified by sharply delineated and enforced rules. The majority of respondents agreed that people in their community were aware of Malagasy poaching laws (75%, n = 36) and that that MNP actively deterred poaching in the area (63%, n = 29). Awareness of Malagasy poaching laws was associated with the belief that people do not poach because local rules are well enforced. The hierarchical cultural type accepts risk if such risk is justified by experts, who are needed to delineate the boundaries of nature’s resilience. Congruent with this notion, the deterrence effects of the MNP had a robust effect on the perceived predictability of poaching’s effects on the environment and local people. Belief that the MNP deterred poaching was correlated with a paradigm of risk in which the risks of poaching are present but well-defined (e.g., predictable) due to the expertise of the MNP.

In contrast, the fokonolona adhered to a high group and low grid, or equalitarian, cultural type. An equalitarian cultural type is one in which there are strong affiliations among individuals (and the group influences individual risk perceptions) but where regulations are relatively undefined and/or unenforced. Although the majority of our respondents were aware of fokonolona poaching rules (84%, n = 41), awareness of fokonolona rules had a negative correlation with agreement that people do not poach because they respect local rules. Furthermore, only one third of respondents (38%, n = 19) agreed that the fokonolona deterred poaching. Although the fokonolona affected the perceived predictability of poaching’s impacts, in the case of environmental impacts this effect became insignificant once the MNP was added to the model. Perceptions about poaching-related responses from the fokonolona had a more direct effect on participants’ belief that poaching risks to people were predictable. This is congruent with the equalitarian cultural type, which is concerned about risks to people in future generations and distrusts risks promoted by elites and governmental authorities.

These differences can be analyzed within the unique framework of environmental law and governance in Madagascar. The structure and sociocultural importance of the fokonolona was threatened by French colonialism in Madagascar, and the fokonolona evolved in response to the risks posed by colonialism, such as centralized control over natural resources [51]. Centralized rules that emerged under colonialism, such as the banning of lemur hunting and of tavy (i.e., slash and burn agriculture) led to social resistance and increased deforestation. Tavy can be sustainable under certain conditions, and is representative of a traditional form of sustenance; a full ban elevated tavy to a symbol of defiance of colonialism [70]. During the colonial era, the fokonolona was known as antily, or “guardian of all risks,” and participation in the fokonolona was mandatory and viewed as essential for resisting centralized control. In contrast, the MNP emerged as a postcolonial institution with the aim of using federal mechanisms to protect biodiversity.

Our results suggest that among study respondents, interventions designed to reduce poaching risks to people and the environment in Madagascar can, and likely should, be implemented by different, and multiple, institutions depending upon the targets of deterrence. For example, our findings indicate that the deterrence activities of the MNP strongly influenced beliefs about poaching risks to both people and the environment, whereas deterrence activities by the fokonolona only influenced risk-related beliefs associated with people. These results imply that it may be unrealistic to expect the fokonolona, or the non-federal institution, to be able to fully deter risks to both people and the environment. Importantly, this paper relies on culturally specific data; these intervention guidelines may not be wholly transferable to other biodiversity hotspots or regions encountering high levels of environmental crime. What is broadly relevant is the notion that utilizing culturally-specific risk data in interventions will likely be most effective in reducing risks because such interventions will be congruent with cultural predispositions [91] and facilitate local involvement in conservation initiatives, which is widely considered essential for the success of community-based natural resource management [10, 92, 93]. Within our study site, delineating these cultural dimensions of environmental crime provided information that may help predict support for laws and deterrence behavior. If laws associated with conservation policies contradict social norms, transgressions are likely [94]; however, the prescriptiveness of social norms may vary according to the cultural type. Egalitarian institutions such as the fokonolona may be ill-suited for the policing of centralized conservation regulations that are based on the risk perceptions of government elites, but may be appropriate vehicles for conservation education that focuses on how poaching negatively affects the livelihoods of local people or violates fady (informal social norms not dependent upon state enforcement). It is important to consider whether or not conservation policies contradict social norms, but it is equally important to understand how social norms are disseminated and incorporated into risk perceptions about poaching, and to design anti-poaching interventions accordingly. In Madagascar, fady have benefited wildlife on both an individual and species level by prohibiting certain hunting practices. Although not informed by expert scientific knowledge, many of these norms are congruent with science-based conservation practices [21]. In Madagascar, the erosion of these informal taboos through a unilateral emphasis on formal conservation rules can lead to negative effects on wildlife [21].

Theoretically, cultural notions of risk that incorporate group-grid typologies have analyzed cultural preferences on two dimensions: the level of affiliation between individuals (group) and the prescriptiveness of norms (grid) [40]. The group-grid approach does not take into account cultural specificity or the institutional level at which “group” is defined. Our research provides insight into these important cultural nuances. For example, the grid dimension traditionally refers to the system of laws and rules that affect individuals in the course of their daily interactions [42]. Our work builds on this theoretical premise by enhancing understanding of how rules, or the prescriptiveness of norms, operate at different levels of group affiliation. The data demonstrate that the deterrence actions of various groups had differential effects on motivations to poach, not to poach, and beliefs about risks from poaching. Clearly, although perceptions about deterrence activities of various institutions can prevent poaching, the reason why they deter poaching differs. Thus far, cultural risk theory has measured group as a binary variable; our research successfully parsed the group variable into multiple parts (looking at different institutions or groups separately) and proved fruitful in explaining cultural responses to poaching.

Our theoretical contribution portends exciting opportunities for future application of cultural theory because the majority of extant research uses versions of a quantitative survey methodology developed by Dake [95] and Marris et al. [43]. These questionnaires are composed of general and context-free questions and fail to incorporate analysis of concrete social relations (specific groups). Examining specific contexts and social relationships, and tailoring quantitative analyses of group-grid theory to the particular context, appears relevant for wholly benefiting from the group-grid dimensions of cultural theory [37, 96, 97]. It is possible that this lack of cultural specificity is weakening the predictive power of cultural risk theory [98]. Furthermore, attempts to revise the measurement of group-grid theory [37] do not consider that individuals may be part of several bonded units or groups. The extent to which an individual is incorporated into groups may depend upon the scale of analysis; for example, the level of incorporation at the family level may be different than the level of incorporation at the town or country level.

Conclusion

Although our study lacks broad generalizability to other contexts due to the sampling protocol, methods herein could be replicated with other sites so as to map cultural norms and responses to poaching alongside other data, including assessed poaching risks and spatially-explicit perceptions of poaching risk [10, 99, 100]. Our survey illustrates the interactions between the deterrence actions of numerous social institutions within the same cultural milieu. Our results suggest that, in conservation sites that include both pre-colonial and post-colonial social institutions, efforts can be made to classify institutions by cultural type so as to understand how different conceptualizations of risk may be undermining collaboration. At our field site in Madagascar, the fokonolona adhered to an egalitarian cultural type, whereas the MNP functioned under a hierarchical cultural model. These types are associated with different myths of nature and risk perceptions. Efforts to include these institutions in anti-poaching initiatives should take into account how these different institutions interpret risk. For example, experts (such as wildlife biologists in the scenario of poaching) are viewed with distrust by equalitarian institutions but are highly valued in hierarchical institutions. Thus, whereas experts might be useful communicators of poaching risks in a hierarchical institution, an equalitarian institution might be better suited to the dissemination of informal social norms such as fady. The continued application and extension of group-grid theory can illuminate how culture influences attitudes about the causes and consequences of wildlife poaching and how anti-poaching interventions might be tailored to specific social institutions and contexts.