Year : 2022 | Volume
| Issue : 4 | Page : 325-335
Exploring a Comprehensive Behavioural Model to Investigate Illegal Sea Turtle Trade in Cabo Verde
Morgan Casal Ribeiro1, Juan Patino-Martinez2, Janete Agues2, Alexandra Marçal-Correia3, Ana Nuno4
1 Maio Biodiversity Foundation (FMB), Cidade Porto Inglês, Ilha do Maio, Cabo Verde; Okeanos-UAc Instituto de Investigação em Ciências do Mar, Universidade dos Açores, Rua Prof. Dr. Frederico Machado, Horta, Portugal
2 Maio Biodiversity Foundation (FMB), Cidade Porto Inglês, Ilha do Maio, Cabo Verde
3 Faculdade de Cieências da Universidade de Lisboa; MARE - ULisboa, Lisboa, Portugal
4 Centre for Ecology and Conservation, College of Life and Environmental Sciences, University of Exeter Cornwall Campus, Cornwall, UK; Interdisciplinary Centre of Social Sciences, School of Social Sciences and Humanities, NOVA University Lisbon, Lisboa, Portugal
Morgan Casal Ribeiro
Maio Biodiversity Foundation (FMB), Cidade Porto Inglês, Ilha do Maio, Cabo Verde; Okeanos-UAc Instituto de Investigação em Ciências do Mar, Universidade dos Açores, Rua Prof. Dr. Frederico Machado, Horta
Source of Support: None, Conflict of Interest: None
|Date of Submission||16-Jul-2021|
|Date of Acceptance||29-Jul-2022|
|Date of Web Publication||26-Sep-2022|
| Abstract|| |
Successful conservation outcomes often depend on changing human behaviours that negatively impact biodiversity, such as unsustainable wildlife harvesting or illegal wildlife trade (IWT). However, inclusive psychology models that examine motivations of those behaviours have been underutilised in IWT contexts. This research examines the drivers of illegal harvesting and consumption of sea turtles on Maio, Cabo Verde (West Africa), by adapting data from interviews (n=20) and questionnaires (n=325) into the Comprehensive Action Determination Model, an environmental psychology theoretical framework. Initial findings suggest local behavioural motivations have changed over time, but key beliefs remained intact. Structural equation modelling showed intention to consume turtles is influenced by positive attitudes towards consumption, but interviews suggest normative personal and social beliefs are becoming relevant to consumptive behaviour mitigation. The same seems true of harvesting, reportedly performed mostly by young men looking to sell turtle by-products. Overall, results indicate the beliefs underlying harvest and consumption behaviours are distinct, such that outreach initiatives must be designed to address each. Results demonstrate how conceptual models developed in underutilised disciplines can be adapted to expand the transdisciplinary tools available to conservation practitioners. Embracing behaviour-focused approaches is crucial to address the intricate cultural and contextual factors of IWT.
Abstract in Portuguese: https://bit.ly/3Aj9xuu
Keywords: behaviour change; conservation psychology; illegal wildlife trade; mixed methods; sea turtle conservation
|How to cite this article:|
Ribeiro MC, Patino-Martinez J, Agues J, Marçal-Correia A, Nuno A. Exploring a Comprehensive Behavioural Model to Investigate Illegal Sea Turtle Trade in Cabo Verde. Conservat Soc 2022;20:325-35
Abstract in Portuguese: https://bit.ly/3Aj9xuu
| Introduction|| |
Illegal wildlife trade (IWT) poses complex threats to biodiversity conservation and management. The importance of designing and implementing research and interventions that focus on behavioural change to reduce demand and supply of illegal wildlife is increasingly recognised (Milner-Gulland et al. 2018; Wallen and Daut 2018). Although regulations can be crucial to reduce IWT, they are most effective when informed by behaviour change principles (t' Sas-Rolfes et al. 2019). Practitioners have begun to shift emphasis from restrictions, legislation, and trade bans, to designing interventions that focus on consumer behaviours (Veríssimo and Wan 2018) and enhanced understanding of rule-breaking (Cooney et al. 2016). This is exemplified by sea turtle trade, where conservation initiatives worldwide have attempted to curb illegal harvest, use, and trade of marine turtles since at least the 1950s. Despite local declines in trade and consumption over recent years, illegal activities remain a pressing global threat to marine turtle populations (Lopes et al. 2022). The behaviours and motivations inherent to consumption or commercial usage of turtle products are specific to each regional context (Veríssimo et al. 2020). An expanded comprehension of behaviour change principles is needed to adequately tackle illicit wildlife harvest and consumption; this requires more, and improved, transdisciplinary tools and practices (Godley et al. 2020).
The relatively new field of conservation psychology seeks to understand peoples' behaviours regarding the natural environment and the conservation contexts with which they interact to promote sustainable pro-conservation behaviours (Saunders 2003). Since its emergence in the early 2000s, it has not yet penetrated mainstream conservation science and its impact is relatively small despite its potential (Wallen and Landon 2020). It is essential to gain insight into key drivers of behaviours such as illegal wildlife hunting and consumption to predict their development over time and promote strategies that reduce them by addressing their root motivations (Thomas-Walters et al. 2020). Areas such as public health and recycling have successfully tested an array of behaviour predictors (Devlin 2018; Geiger et al. 2019). However, within conservation, approaches have tended to focus on attitudes towards overarching conservation contexts, which may be inadequate predictors of specific target behaviours within those contexts (Nilsson et al. 2020). The lack of research on specific behaviours considerably limits the effectiveness of interventions and may threaten existing values that positively influence conservation outcomes (Balmford et al. 2021).
There are several conceptual models that incorporate behavioural predictors—adopting a positivist paradigm—of potential use in conservation. To conjugate three commonly used theories within environmental psychology (Theory of Planned Behaviour; Norm Activation Theory; and the Value Belief Norm Theory), the Comprehensive Action Determination Model (CADM) was proposed (Klöckner 2013) as a unifying framework. It seeks to avoid the weaknesses of single models, such as under-representing the impact of morality or situational constraints. And although studies comparing different models are generally limited, CADM has been shown to have a higher predictive power when compared to its integrating models (e.g., Klöckner and Blöbaum 2010; Balundė et al. 2020). It consists of a relatively large number of psychological constructs proximally or distally related to behaviour, and incorporates intentional, normative, situational, and habitual influences to explain behaviour. If simplification is called for, these can be scaled down to include only the closest determinants of behaviour for a given context (Klöckner 2013; [Figure 1]).So far, CADM has been applied to research on recycling, energy saving, clothing consumption reduction and other environmentally significant behaviours (Ofstad et al. 2017; van den Broek et al. 2019; Joanes et al. 2020). However, it has not been adopted to understand behaviours directly linked to conservation. Psychological tools applied in conservation must consider the highly contextual nature of target behaviours, such as IWT (Selinske et al. 2018).
|Figure 1: Adapted CADM framework based on Klöckner (2013). Relationships between psychological variables that predict behaviour; habit and behaviour (grey) are not incorporated in this study; habit, used in the original model, was substituted by tradition (bold), which better addresses specific research questions. Solid arrows show tested relationships; dashed arrows represent causal pathways not analysed due to illicitness of turtle consumption and harvest. For definitions of variables see Table 1|
Click here to view
One of the CADM's strengths is that it encompasses a broader range of behavioural predictors when compared to other commonly utilised theories in conservation psychology. Therefore, it could be more suitable in identifying the relevant determinants of turtle meat consumption and recognising the motives of sea turtle trade on Maio island, Cabo Verde. In this context, CADM's suitability must first be explored before it can be recommended to conservation practitioners. At the same time, notwithstanding its broad applicability, adopting the CADM should be done in a culturally sensitive manner, by taking nuanced approaches adjusted to socially specific contexts of IWT (Margulies et al. 2019). The current research aims to better understand drivers of illegal consumption and harvesting of sea turtles on Maio by employing mixed methods that combines open-ended interview responses and closed-ended questionnaire responses. The study 1) characterises the contexts of past and present consumption and harvesting of turtles on Maio; and 2) identifies key psychological factors related to turtle consumption on Maio in both past and present via the CADM framework.
| Methods|| |
The study followed a dynamic mixed methods design, combining semi-structured interviews and a structured questionnaire survey based on the CADM (Creswell and Clark 2018). An initial phase of semi-structured interviews was performed to characterise past and present activities, motivations and social contexts related to consumption and harvesting of turtles on the island. Second, an island-wide survey informed by the CADM framework was conducted to explore key psychological predictors of past and present intention to consume turtles on Maio. The iterative combination of data collection and interpretation was used with the expectation that qualitative results would inform the design of a CADM survey instrument, explain relationships found by quantitative analysis and provide data on turtle harvesting that was not specifically targeted by the survey. International best practices were followed, adhering to the guidelines by the British Sociological Association Statement of Ethical Practice (2017). When approached for participation, respondents were informed that participation was voluntary and anonymous; withdrawal was possible at any time; individual details would not be disclosed or identifiable; and information collected would be used for research purposes only.
The Cabo Verde archipelago, 500 km from the West African coast, hosts more than 95% of endangered Loggerhead Turtle Caretta caretta nests across eastern Atlantic, and possibly represents the largest nesting subpopulation in the world (Marco et al. 2012; Patino-Martinez et al. 2022). The island of Maio (Figure S1, Supplementary Information) is among the most important islands for nesting. It comprises one civil parish of 6,980 people, subdivided into 12 settlements that range from 22 to almost 3,000 inhabitants (INECV 2018).
Conservation efforts at a national scale involve governmental authorities; environmental awareness campaigns; local and international non-governmental organisations (NGO); and regulations against killing, acquiring, commercialising, exporting or consuming sea turtle meat and eggs, punishable by fine or imprisonment. The Maio Biodiversity Foundation (FMB), a local NGO, has conducted extensive work to protect sea turtles in the area since 2010. All nesting beaches are currently patrolled nightly during the nesting season by teams of local residents. Yet, harvesting and consumption of loggerheads have persisted and are responsible for the annual take of hundreds of nesting females while research regarding sea turtle trade in the country is scarce (Marco et al. 2012; Hancock et al. 2017). In addition, clandestine consumption of turtle meat, harvesting of female loggerheads on beaches, selling turtle meat locally or between islands, artisanal crafting using turtle shells, medicinal use of turtle organs and employing their penises as an aphrodisiac still occurs (Martins et al. 2015).
To acquire in-depth contextual information regarding sea turtle trade on Maio, and to contribute to the later design of a questionnaire, semi-structured interviews were conducted in October 2019. These consisted of in-person discussions guided by an interview script listing key questions. Two interview scripts specific to harvesting or consumption (defined as furtive behaviours conducted in private or in small groups) were developed and were conducted with men and women, respectively. The two interview scripts became gender-tied primarily because men's almost exclusive role in harvesting while preparation and cooking are delegated to women. Both interview scripts first asked questions about consumption/harvesting over the past 15 years, followed by open-ended questions concerning these behaviours in more recent years. This allowed participants to answer candidly and in-depth about past behaviours before choosing what to focus on when discussing recent events, which could be adequately probed by the interviewer. Key interview questions and topics addressed were: turtle conservation; turtle harvesting and consumption; body parts used and why; motivations for consumption and harvesting and their prevalence in different communities (see Supplementary Information for interview scripts). As data collection progressed, novel lines of questioning were added throughout the interviewing process (e.g., questions about inter-community variability) as iteration is one of the strengths of inductive qualitative methods (Bryman and Burgess 1994).
Interviews were conducted in private spaces familiar to participants, with only researchers' presence to minimise bias. Because harvesting is a particularly sensitive subject, privacy and confidentiality were of utmost importance. Men were interviewed individually, and women were interviewed in groups of three to six, so the outspoken participants encouraged the apprehensive ones to open up. Women within the same group were neighbours, friends, or family. Participants were selected through targeted sampling whereby gatekeepers, such as community leaders or charismatic local women, referred researchers to consenting adults willing to talk about their direct experiences with the subject matter. A quota sample of 20 interviews (10 with individual men and 10 with groups of women) in the 12 settlements of Maio was set. Once reached, data was reviewed to confirm if theoretical saturation was achieved, i.e., a point when further data collection provides little new information towards understanding the research questions (Corbin and Strauss 2015).
Interviews were administered in a mixture of Portuguese and Maiense creole, by two Portuguese non-local researchers, one acting as facilitator while another took notes to aid communication and data collection (recordings were taken via smartphone devices). Interviews began informally, stating the purpose of the research, explaining confidentiality, and recording consent. Two pilot interviews were performed in the same settlement with one man and three women, which did not produce changes in interview scripts.
Questionnaires—quickly administrable structured questions that allow for testing relationships between variables (Bryman 2012)—were employed to explore the relationships between the various CADM predictors for consumption [Table 1]. Questionnaires could not feasibly follow the framework through to behavioural outcomes due to the illegality of turtle consumption, instead we focused on intention—the immediate antecedent of the behaviour (Newth et al. 2021; in addition, during preliminary data collection, intention was a less contentious topic than having admitted turtle consumption). A reduced version of the CADM, with variables most proximal to behaviour, was adopted as suggested by Klöckner (2013). Given the illicit nature of turtle consumption and its dependency on annual nesting seasons, it would be inappropriate to use 'habit' as a variable. As such, to maintain the general framework of the model while considering the regional context, 'tradition' was used instead. Traditional values can contribute greatly towards conservation, favourably or otherwise (Mutalib et al. 2013). Because harvesting represents an extremely sensitive topic, reportedly engaged in by far fewer people than consumption, and only performed by men, this behaviour was not addressed in questionnaires.
|Table 1: Definitions for each CADM behavioural predictor (adapted from Klöckner 2013; Mutalib et al. 2013)|
Click here to view
Questionnaires (Supplementary Information) were composed of three sections: 1) sociodemographic information; 2) items for consumption behaviour predictors in the present and 3) in the past (approximately 20 years ago). Survey administration was conducted after turtle nesting season, so items for the present timeframe were worded around next year's nesting season. The past timeframe, which substituted the interview's 15-year timeframe, emerged from interview data that suggested it more accurately conveyed the general period when turtle protection was non-existent. Sociodemographic variables were also drawn from preliminary interview results: gender, age, settlement size, geography (east/west), household size, level of education, monthly salary and time living in Maio (see Table S1, Supplementary Information). Questions in the last two sections consisted of single or multiple items for each CADM variable, using five-point semantic differential or Likert-type scales. Except for timeframes, past and present items were worded in the exact same way to allow for comparison (Table S2, Supplementary Information).
Surveys were conducted across all Maio settlements in November 2019, using a systematic sampling protocol for some areas and census protocol for others. Sampling units were defined as any person 18 years or older who resided on Maio. The questionnaire was reviewed by two residents and a field pilot was administered (n = 13) in an area of the largest community, resulting in the rewording of two items (pilot study data from those items was not used in subsequent analyses). Surveys were carried out separately by two Portuguese non-local researchers in Maiense creole using digital forms on ODK Collect v.1.25.1 (Hartung et al. 2010). After removing five questionnaires with missing data resulted in a sample size of 325 (our target sample size was n = 363 based on a power analysis at 95% confidence level and 5% margin of error).
Analysis of semi-structured interviews
Following verbatim transcription, a single coder followed a descriptive coding process with instances of in vivo coding when pertinent, creating a codebook (Supplementary Information). Interview transcripts were coded inductively in two cycles, with codes developed from the data and revised until common codes were organized into categories and overarching themes (Saldaña 2021). Categories were saturated according to Charmaz (2006). Notes taken later during unstructured conversations with survey participants were included in this analysis. Concept maps were generated throughout coding to aid analysis [Figure 2]; coding was performed using MAXQDA 2020 (VERBI Software 2019).
|Figure 2: Conceptual map of motivations and deterrents for past/present consumption and harvesting, using qualitative data collected in interviews and field notes. Motivations/deterrents are displayed as major or minor according to participants' perceptions of these factors rather than frequency of their coding|
Click here to view
Analysis of CADM questionnaires
Preliminary descriptive analysis was performed to explore sociodemographic data and reverse code certain items so that low/high scores indicated favourable/unfavourable responses towards turtle consumption. Participants often attributed intense consumption/harvesting to the west or the east side of the island, therefore this was included as a sociodemographic variable. Cronbach's alpha, composite reliability and average variance extracted tests were conducted to estimate the reliability and validity of questionnaire items relating to the CADM. Scales that did not reach the customary threshold of >0.85 were further analysed during confirmatory factor analysis (CFA) (Said et al. 2011). Mardia's multivariate normality tests were run for CADM items from both past and present datasets; as multivariate normality was not verified (z-scores >5), the generally employed Maximum Likelihood estimation method was substituted by Maximum Likelihood estimation with robust standard errors (MLR) for all subsequent analyses (Gana and Broc 2019). Structural equation modelling (SEM) was used to assess the network of relationships between CADM variables by testing hypothesised relations among both observed (measured) and unobserved (latent) variables (Kyle et al. 2020). Measured variables are items such as the ones in the questionnaire, while latent variables (LVs) are a construct defined by multiple items (Table S2, Supplementary Information).
The common SEM approach was followed, first specifying the model, then determining model identification and estimating its parameters, and finally assessing model fit (Kyle et al. 2020). The indicators that measure LVs (measurement model) were estimated through CFA, followed by determining the interrelations between variables. Sociodemographic variables were included to control for any potential effects on intention. As the same model was specified for past/present, measurement invariance was established to test mean differences across time (Putnick and Bornstein 2016). Finally, latent growth modelling was employed to answer three questions: 1) did predictors of intention and intention itself change significantly from past to present; 2) are there differences in their initial levels and rates of change between participants; and 3) do any sociodemographic variables influence these changes? Statistical analyses were conducted using RStudio 1.1.546 and lavaan 0.6-6 (Rosseel 2012; R Core Team 2020).
Following current recommendations, several fit indices are reported for each model (Table S3, Supplementary Information). However, because of sample size and non-normal distributions on chi-square and χ2/df values, incremental and absolute fit indexes are more relied upon (Putnick and Bornstein 2016).
| Results|| |
Practices, motivations, and barriers of consumption/harvesting
The main categories that emerged from semi-structured interviews fit into three overarching themes, all divided into past and present: 1) narratives about consumption, 2) about harvest, and 3) motivations/deterrents determining both behaviours [Figure 2].
Past consumption was described in detail, with explanations about how turtle parts were cooked or used in crafting. Participants unanimously acknowledged that barring rare exceptions, the entire population ate turtles before they were protected. No consensus existed as to when protection began, but participants agreed that the implementation became stricter after the FMB began its activities and after the criminalisation of turtle consumption in 2017. Consumption was reported to have decreased greatly thanks to new legislation and awareness. People who persist were said to have adopted strategies to avoid discovery; due to its intense smell, cooking meat undetected is challenging. Some participants disclosed that if cooked turtle meat were offered to them by friends or family, they would indulge in opportunistic consumption.
I don't catch any (turtles). But if someone cooks some, I might have some. My neighbour sometimes cooks them. She uses lemon and wine so it doesn't smell, and cooks them (Male, Vila do Maio).
My (redacted familiar) helps protect turtles. Next year I want to go on patrols too. (…) I don't catch turtles, no. But if someone else offers me turtle from their pan, I'd accept (Female, Alcatraz).
I don't eat (turtle) because I haven't got a fisherman at home. There's no fisherman at home, but If there was, I might (eat turtle) (Female, Vila do Mail).
Regarding past harvesting, participants mentioned that men would leave their communities in small groups to patrol beaches until they found turtles. Turtles' plastrons (ventral surface of the shell) would be removed and meat/organs cut away; the animal remained alive for almost the entire process. Various participants reflected on how sad this method currently made them feel. Groups were reported to ordinarily catch more than one turtle per night, sometimes up to five or ten. Some participants asserted harvesting was a daily activity, while others said it occurred once or twice a week, suggesting potential variability in practices. Concerning present harvesting, some participants either denied its existence or affirmed it has become rare in recent years. Others explained it still occurs, sometimes recounting new strategies harvesters use to evade detection by the beach guards.
As to motivations, past consumption/harvesting were associated with several incentivising factors but not a single deterrent, being recounted as just a part of everyday life [Figure 2]. Participants conveyed consumption was mostly prompted by turtles' delicious meat and the lack of food, with infrequent mentions of medicinal purposes (although discussions did not provide additional information on these). Harvesting was said to be driven by consumption itself while also to “fill the house” with other foods, by trading turtle meat for rice, potatoes, and such staples. Current behaviours showed both incentives and deterrents [Figure 2]. Turtle-meat flavour was mentioned as the near-exclusive motivation for consumption. Most participants observed that consumption can no longer be attributed to poverty, as other food is more readily available and turtle nesting is seasonal. Harvesting is now reportedly largely connected to selling turtle meat, but occasionally shells and penises (an aphrodisiac); the exchange of meat for money reportedly only started after the illegalisation of consumption/harvesting, unlike earlier when turtles could be freely caught. Meat is assertedly sold to locals or on the nearby island of Santiago for higher prices (Table S4, Supplementary Information). Participants blamed harvesting on young men unable to find jobs or men addicted to drugs/alcohol/cigarettes. Recorded deterrents for present consumption and harvesting coincided. Fear of law enforcement and the dependence of local incomes on turtle conservation are major obstacles; the self-perceived sadness of the way turtles were harvested, and religious dietary restrictions linked to turtles being considered unclean animals unfit for consumptions (Adventists have taboos on eating turtles) being minor ones.
Turtles have a sad death. In those times, you would kill turtles with only an axe. It's sad… for my part, I say don't eat turtle anymore. No, turtle not even when offered (Female 1, Morro).
The problem was, in those times, we didn't understand what a turtle was. Because we used to say anything that came from the sea was welcome (Female 2, Morro).
Me, now… I couldn't do it anymore. A turtle's death, they have a sad death. People beat them with axes (Male, Figueira).
Difficulty to covertly cook turtle meat (e.g., cooking at night or using herbs/spices to avoid being noticed due to intense smell) was a barrier exclusive to consumption. One community where harvesting in recent years had been particularly intense (Barreiro) was said to have improved substantially by diverse participants from all communities; due in large part to the collaboration of the much-respected local football team with the FMB for year-round conservation activities.
After 2012, people even stoned me at the beach. Harvesters would go and throw rocks and take turtles by force. It was a kind of violence(…) (Male, Barreiro).
Interviewer: Why has there been no harvest this year?
Because, with culture, with the FMB, with the football club… They entered a deal. The club helped people, by campaigning. (…) Turtles haven't been caught this year, but only thanks to great efforts by the club and FMB (Male, Barreiro).
Interviewer: Was there anywhere with more harvesting, or less?
These last years, after 2014, 2015, in Barreiro. (…) But now not so much. Because the FMB has had help from the football club, they have had help protecting turtles (Male, Figueira).
Finally, not discussed necessarily as a deterrent, existing turtles were commonly perceived as smaller and less tasty than in the past, due to coming from “nurseries”, “laboratories”, “captivity”, “freshwaters”, “mixed with foreign species” or “injected with medicine” (in these cases participants seem to have confused turtle tagging for medicinal injections).
Turtles raised in freshwater have another taste than those from the sea. That's why nowadays people say you can get ill from eating turtles (Male, Pilão Cão).
In the old days, turtles were tastier. They were of the land, nicer. Now turtles are raised at beaches, mixed with foreign turtles (Female, Ribeira Dom João).
Psychological factors driving turtle consumption
Of the 325 survey participants, 270 considered themselves able to answer questions concerning past consumption (i.e., 20 years ago). On average, 9.2% of residents were sampled in each of the 12 settlements (100% of the island's settlements; 4.7% of the island's inhabitants), having a median age of 45-54 and being two-thirds women (Table S5, Supplementary Information).
Both measurement and structural models for past and present data showed good or acceptable fit, standardised factor loadings were satisfactory, and models were over-identified (Tables S6 and S2 respectively, Supplementary Information). For past and present data, attitude was the only LV to significantly predict intention (B = 0.98, P-value = 0.001 and B = 1.15, P-value = 0.04 respectively; [Figure 3]). Although unrelated to intention, normative personal beliefs were affected similarly by normative beliefs in the past and present (B = 1.29, P-value = 0.03 and B = 3.07, P-value = 0.01 respectively; [Figure 3]) but were determined by perceived behavioural control (PBC) only in the present (B = -0.59, P-value = 0.01; [Figure 3]). Sociodemographic variables showed no significant relationships with intention in either model, suggesting these effects are widespread across groups.
|Figure 3: SEM results, showing the relative influence of statistically significant variables using unstandardized structural coefficients (*p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; Bollen 2014). Non-bold/bold numbers correspond to past/present models, respectively; ovals/rectangles represent multiple/single indicator variables respectively|
Click here to view
The latent growth model for attitudes showed excellent fit, and after correcting a Heywood case due to sampling fluctuation (Dillon et al. 1987) so did the intention model (Table S6, Supplementary Information). Slope and intercept estimates were significant in both models, indicating change in mean attitudes and intention over time. Their respective variances were also significant, suggesting that inter-individual differences exist in the initial mean levels and change over time of both attitudes and intention, that could be explained by sociodemographic variables. Of the sociodemographic variables tested, both models' intercepts were related to household size, while slopes were affected by household size and geography (Table S7, Supplementary Information).
Overall, latent growth models indicate three things: attitudes and intention changed significantly over time; larger households were related to significantly lower/higher scores in past attitudes/intention respectively but changed more over time; past attitudes/intention did not differ significantly according to geography, but changed considerably more in eastern settlements than western ones [Figure 4].
|Figure 4: Linear change in mean attitudes (top) and intention (bottom), from past to present, according to household size groups, in west (left) and east (right) settlements|
Click here to view
| Discussion|| |
Reducing IWT entails understanding and reforming the complex behaviours at its centre (Wallen and Daut 2018). The purpose of this research was to characterise past and present consumption and harvesting of sea turtles on Maio, and to identify the psychological factors of potential significance to past and present turtle consumption behaviour. Results from both qualitative and quantitative approaches suggest negative attitudes have been the largest cause of turtle consumption. Other factors did not significantly predict consumption intention in the CADM, but interview data suggests PBC, normative personal beliefs and normative social beliefs may be consequential to consumption and harvesting and could be useful for practitioners designing outreach programmes.
Attitudes as the main motivators for sea turtle trade
Findings suggest attitudes towards sea turtle consumption and harvesting have been important drivers of both behaviours in the study area, from the present dating back to before sea turtle conservation was established on Maio. Regarding consumption, it seems clear that attitudes play a vital role. It was the only CADM variable shown to affect intention across time, while major consumption motivations drawn from qualitative data could be interpreted as affective components of attitudes that involve feelings and emotions (McLeod 2018). Thomas-Walters et al. (2020) offer a useful guide to classify motivations responsible for wildlife consumption, whose definitions are followed here. Participants often described that in the past, large quantities of easily obtainable food provided by turtles offered them peace of mind, but these nutritional motivations no longer exist. The most prevalent attitude in the data seems to have endured from past to present: a sensory motivation of pleasure seeking, owing to turtle meat's appreciated flavour. Consuming expensive turtle penises seems to be multifaceted, conforming to medicinal, ritualistic, and reputational motivations, but uncommon on Maio and more prevalent on neighbouring Santiago. Penis consumption may be more frequently discussed in the present due to it being considered a display of wealth/status by emigrants returning to Cabo Verde from European countries. Historically, semi-structured interviews and latent growth models show attitudes/intentions in eastern populations were more in favour of consumption but presently are similar to western populations. Interviews suggest this change is due to conservation activities targeting a more hostile eastern settlement, involving a local football team and its supporters to noticeable effect, further substantiated by Patino-Martinez et al. (2020). Furthermore, individuals from larger households showed greater increases in attitudes/intention perhaps due to illegal behaviours being harder to furtively accomplish compared to people who live alone. Interview data suggests interfamily conflicts can arise within the same household due to turtle consumption, but this was not expanded upon by participants.
Despite their major role, altering attitudes alone may not result in a behaviour change that decreases consumption. As strong and stable constructs, they may remain accessible to an individual's motivations even after seemingly being neutralised (Petty et al. 2006; Bohner and Dickel 2011). Best practices from conservation psychology seem to be shifting from attitude campaigns to well-rounded programmes that also address other relevant behavioural components (Nilsson et al. 2020) and the same approach is suggested for Maio. Interestingly, the belief that turtles are now less tasty because they are “raised in nurseries” or “injected with medicine” could hold value to conservation, as it goes against the predominant sensory motivation. Rumours concerning protected species are common in certain fields of conservation, but often hinder species' protection (Skogen et al. 2008). In this case, not taking deliberate action to deconstruct these beliefs might be the best approach for turtle conservation, as they seem to help protect turtles from being consumed (although the ethics of doing so should obviously be considered).
Although addressed solely through interviews, harvesting appears to be similarly motivated by attitudes. In the past these were likewise nutritional and sensory, but presently are motivated by immediate profit (Thomas-Walters et al. 2020) comparable to other Cabo Verde islands (Hancock et al. 2017). When exported to Santiago, a single turtle can generate several months' worth of salaries (Supplementary Information).
Law enforcement as a behavioural deterrent
Interview results indicate law enforcement as a major deterrent of consumption/harvesting on Maio. Of the limited studies conducted in Cabo Verde on the topic, Hancock et al. (2017) also found that legal measures were important in influencing turtle trade, generally people did not willingly comply with sea turtle trade bans, and strict enforcement was needed to prevent harvesting. Similarly, present data indicates a considerable number of inhabitants on Maio do not engage in consumption/harvesting for fear of being caught by beach patrols. This would be in line with participant's accounts of increased harvesting when patrol periods end and reports from recent research (Patino-Martinez et al. 2020). Such opportunistic behaviour may point to the importance of perceived controllability: the belief that one's behaviour is volitional, a component of PBC (Ajzen 2002; Bandura 1991). People seem most likely to engage in consumption/harvesting on Maio when they perceive they will successfully carry out these behaviours without being impeded—whether because they believe they can avoid detection by patrols, or because patrols have ended altogether. Because PBC can have a direct effect on behaviour (unmediated by intention) when conditions change before the behaviour is performed, this could help reconcile why PBC did not significantly affect consumption intention in the CADM while being described as a major deterrent in interviews. In future, other under-utilised scientific frameworks could further enrich our understanding of these behavioural deterrents, such as conservation criminology (Gibbs et al. 2010)
Norms of potential value to conservationists
In this study, tradition, normative personal beliefs and normative social beliefs were not found to meaningfully shape intentions. While no participant from interviews or questionnaires explicitly attributed consumption/harvesting to tradition, minor deterrents taken from interviews could be related to normative personal beliefs or normative social beliefs. For some participants, recent exposure to live turtles sparked sadness over the manner of their harvesting, and for others, mentions of Bible passages declaring turtles unsuitable for consumption (Deuteronomy 14, verses 9-10) elicited moral reflections. The first could be interpreted as an ascription of responsibility (feeling responsible for negative consequences) generated by newfound awareness of consequences (being aware of negative consequences for others), two components of normative personal beliefs according to the CADM and its integrating theories (De Groot and Steg 2009; Klöckner 2013). The second seems to be a religious motivation (Thomas-Walters et al. 2020), and the desire to practice spiritual beliefs seems in line with the desire to act in ways consistent with one's values of morality and sense of self-worth (Schwartz 1977). Interventions focusing on norms require understanding of the norm generation and activation process. Therefore, conservation programmes on Maio could reinforce these existing norms through interaction with turtles and incorporating religious passages in their messaging, and subsequently measure their effectiveness (Clements et al. 2009).
Outreach initiatives conducted by the FMB that partnered with a local football team in the most consumption/harvesting intense community (in the east) were attributed substantial success by participants. Although a geographically confined effect, the community has transitioned from being mostly indifferent or hostile to turtle conservation with negligible social pressure on harvesters/consumers to stop their behaviours, to a community with more unified values against these behaviours. Unlike Maio as a whole, where normative beliefs were not significant in either interviews or the CADM questionnaires, this settlement's participants point to outreach being the primary reason for behavioural decline. This supports recommendations about using collective action in IWT outreach to shift or create normative beliefs (Bujold et al. 2020) and has been shown effective in similar contexts (Steinmetz et al. 2014).
Limitations and methodological considerations
The findings of this study should be viewed in the light of some limitations. First, a smaller-than-ideal sample size may be a concern, especially regarding analysis of past psychological factors. Second, fewer indicators than recommended were used for each LV that examined consumption (Bollen 2014), with PBC and tradition being direct measures—although high reliability with direct measures is possible (Ajzen 2002). A smaller number of indicators was considered a necessary compromise for several reasons, mainly: the sometimes-subtle differences between different indicators for the same LV can be difficult to express in Maiense creole; as questionnaires repeated the same items for past and present consumption, repetitiveness and participant fatigue was an issue. Scales used to measure these items were not existing validated scales, instead being created in Maiense specifically for this research. Additionally, items for normative social beliefs included both injunctive and descriptive aspects. Future research on Maio should make sure that injunctive and descriptive items are distinct LVs, which can also be conceptualised as empirical and normative expectations (Bicchieri et al. 2014). Third, a lack of previous research on past consumption/harvest required the use of participants' long-term recall. Retrospective perceptions of past ecological conditions must be cautiously interpreted (Jones et al. 2020), but research shows recollections of high-value/low-frequency events (such as seasonal turtle harvesting) can be less prone to this bias (Thurstan et al. 2016). Sensitive questioning techniques were initially considered to minimise social desirability bias, but ultimately were not utilised due to the larger sample sizes required and the potential increased complexity for participants (Nuno and St John 2015). Finally, it is necessary to acknowledge that analysis of relationships between variables provides a static view of social life (Bryman 2012). This was at least partially offset by complementing behavioural modelling with interviews, which helped provide a more complete answer to research questions.
| Conclusion|| |
Conservation psychology has become increasingly prominent in literature (Wallen and Landon 2020), and in recent years has come to the fore of IWT consumer research ('t Sas-Rolfes et al. 2019). Here, a conceptual model developed in environmental psychology was applied to an IWT context, which complemented by qualitative methods provided a better understanding of drivers of sea turtle harvesting and consumption on Maio, Cabo Verde. Combining qualitative and quantitative research elements is fundamental to address specific issues that would not be sufficiently understood using one standalone approach (e.g., turtle consumption/harvesting), and to obtain prior knowledge that provides in-depth contextual information for developing educated questionnaire designs. Efforts on the island should use acquired knowledge to design appropriate outreach programmes. In large part these initiatives should: tackle the predominant sensory motivation regarding consumption; increase the perceived amount of beach protection; reinforce people's ascription of responsibility regarding turtle conservation; and when appropriate, use specific community partnerships to help shape normative beliefs. Future research on Maio should continue to explore the social psychology of sea turtle trade and rigorously evaluate the impacts of conservation actions on psychological constructs. This would support the need for identifying which behaviours are central to IWT, who engages in them, what primarily drives them, and how to properly engage in behaviour change.
Author contributions statement
M.C.R., A.N., J.P.M., and J.A. conceived and designed the research; M.C.R. collected the data; M.C.R. analysed the data; M.C.R. led the drafting of the manuscript. All authors contributed critical, intellectual content to the drafts and gave final approval of the version to be published.
Special acknowledgements to Raquel Amador for her input in study design, and to Inês Carneiro and Inês Mota for fieldwork assistance. Further thanks to local leaders for facilitating data collection and to all interview and survey participants. A.N. acknowledges the support of the European Union's Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement SocioEcoFrontiers No. 843865.
Declaration of competing/conflicting interests
The authors declare no competing interests in the conduct of this research.
This research was funded by MAVA Fondation pour la Nature and NOAA USFWS.
Research ethics approval
This research did not require ethical approval by the organization of the lead author. Nevertheless, it adopted the ethical research guidelines published by the European Commission, DG Research and Innovation, and followed legislation delineated in Article 31 of the Decree Law no. 58/2019 and in Regulation (EU) 2016/679 of the European Parliament. Specifically, when approached for participation, appropriate locally relevant social conventions were followed and respondents were informed that: participation was voluntary and fully anonymous; withdrawal from the study was possible at any time; individual details would not be disclosed or identifiable; and information collected would be used for research purposes only.
The datasets presented in this article are not readily available because all data used for this study is strictly confidential given the illegal nature of some practices being reported. Requests to access the fully anonymised datasets should be directed to M.C.R.
| References|| |
Ajzen, I. 2002. Perceived behavioral control, self-efficacy, locus of control, and the theory of planned behavior 1. Journal of applied social psychology
Balmford, A., R.B. Bradbury, J.M. Bauer, et al. 2021. Making more effective use of human behavioural science in conservation interventions. Biological Conservation
Balundė, A., L. Jovarauskaitė, and M.S. Poškus. 2020. Exploring adolescents' waste prevention via value-identity-personal norm and comprehensive action determination models. Journal of Environmental Psychology
Bandura, A. 1991. Social cognitive theory of self-regulation. Organizational behavior and human decision processes
Bicchieri, C., J.W. Lindemans, and T. Jiang. 2014. A structured approach to a diagnostic of collective practices. Frontiers in Psychology
Bohner, G. and N. Dickel. 2011. Attitudes and attitude change. Annual review of psychology
Bollen, K.A. 2014. Structural equations with latent variables
. Volume 210. New York, NY: John Wiley & Sons.
Bryman, A. 2012. Social research methods
. 4th edition. New York, NY: Oxford University Press.
Bryman, A. and R.G. Burgess. 1994. Analyzing qualitative data
. 1st edition. London: Routledge.
Bujold, P. M., K. Williamson, and E. Thulin. 2020. The science of changing behavior for environmental outcomes: a literature review. Rare Center for Behavior & the Environment and the Scientific and Technical Advisory Panel to the Global Environment Facility.
Charmaz, K. 2006. Constructing grounded theory: a practical guide through qualitative analysis
. Thousand Oaks, CA: Sage Publishing.
Clements, R., R. Foo, S. Othman, et al. 2009. Islam, turtle conservation and coastal communities. Conservation Biology
Cooney, R., D. Roe, H. Dublin, et al. 2016. From poachers to protectors: Engaging local communities in solutions to illegal wildlife trade. Conservation Letters
Corbin, J. and A. Strauss. 2015. Basics of qualitative research: Techniques and procedures for developing grounded theory
. 4th edition. Thousand Oaks, CA: Sage Publishing.
Creswell, J. and Clark, V. 2018. Designing and conducting mixed methods research
. Thousand Oaks, CA: Sage.
De Groot, J.I. and Steg, L. 2009. Morality and prosocial behavior: the role of awareness, responsibility, and norms in the norm activation model. The Journal of social psychology
Devlin, A.S. 2018. Environmental psychology and human well-being: effects of built and natural settings
. Cambridge, MA: Academic Press.
Dillon, W.R., A. Kumar, and N. Mulani. 1987. Offending estimates in covariance structure analysis: comments on the causes of and solutions to Heywood cases. Psychological bulletin
Gana, K. and G. Broc. 2019. Structural equation modeling with lavaan
. London and Hoboken, NJ: John Wiley & Sons.
Geiger, J.L., L. Steg, E. van der Werff, et al. 2019. A meta-analysis of factors related to recycling. Journal of environmental psychology
Gibbs, C., M.L. Gore, E.F. McGarell, et al. 2010. Introducing conservation criminology: towards interdisciplinary scholarship on environmental crimes and risks. The British Journal of Criminology
Godley, B.J., A.C. Broderick, L.P. Colman, et al. 2020. Reflections on sea turtle conservation. Oryx
Hancock, J.M., S. Furtado, S. Merino, et al. 2017. Exploring drivers and deterrents of the illegal consumption and trade of marine turtle products in Cape Verde, and implications for conservation planning. Oryx
Hartung, C.A., Y. Lerer, C. Anokwa, et al. 2010. Open data kit: tools to build information services for developing regions. In: Proceedings of the 4th ACM/IEEE international conference on information and communication technologies and development
. December 13–16. New York, NY: Association for Computing Machinery.
INECV (Instituto Nacional de Estatística de Cabo Verde). 2018. Estatísticas das famílias e condições de vida
. Praia, Cabo Verde: Instituto Nacional de Estatística. Inquérito multi-objetivo contínuo 2018.
Joanes, T., W. Gwozdz, and C. A. Klöckner. 2020. Reducing personal clothing consumption: a cross-cultural validation of the comprehensive action determination model. Journal of Environmental Psychology
Jones, L.P., S.T. Turvey, D. Massimino, et al. 2020. Investigating the implications of shifting baseline syndrome on conservation. People and Nature
Klöckner, C.A. 2013. A comprehensive model of the psychology of environmental behaviour—a meta-analysis. Global environmental change
Klöckner, C.A., and A. Blöbaum. 2010. A comprehensive action determination model: toward a broader understanding of ecological behaviour using the example of travel mode choice. Journal of Environmental Psychology
Kyle, G., A. Landon, J. Vaske, et al. 2020. Tools for assessing the psychometric adequacy of latent variables in conservation research. Conservation Biology
Lopes, L., A. Paulsch, and A. Nuno. 2022. Global challenges and priorities for interventions addressing illegal harvest, use and trade of marine turtles. Oryx
Marco, A., E . Abella, A. Liria-Loza, et al. 2012. Abundance and exploitation of loggerhead turtles nesting in Boa Vista island, Cape Verde: the only substantial rookery in the eastern Atlantic. Animal Conservation
Margulies, J.D., R.W.Y. Wong, and R. Duffy. 2019. The imaginary 'Asian Super Consumer': a critique of demand reduction campaigns for the illegal wildlife trade. Geoforum
Martins, S., F. Rocha, E. Rodrigues, et al. 2015. The use of sea turtles in traditional medicine in the Cape Verde Archipelago, West Africa. African Sea Turtle Newsletter
Milner-Gulland, E.J., L. Cugniere, A. Hinsley, et al. 2018. Evidence to action: research to address illegal wildlife trade. SocArXiv. https://osf.io/preprints/socarxiv/35ndz/
. Accessed on October 4, 2020.
Mutalib, A.H.A., N. Fadzly, and R. Foo. 2013. Striking a balance between tradition and conservation: general perceptions and awareness level of local citizens regarding turtle conservation efforts based on age factors and gender. Ocean & coastal management
Newth, J. L., McDonald, R. A., Wood, K. A., et al. 2021. Predicting intention to hunt protected wildlife: a case study of Bewick's swan in the European Russian Arctic. Oryx
Nilsson, D., K. Fielding, and A.J. Dean. 2020. Achieving conservation impact by shifting focus from human attitudes to behaviors. Conservation Biology
Nuno, A. and John, F.A. S. 2015. How to ask sensitive questions in conservation: a review of specialized questioning techniques. Biological Conservation
Ofstad, S.P., M. Tobolova, A. Nayum, et al. 2017. Understanding the mechanisms behind changing people's recycling behavior at work by applying a comprehensive action determination model. Sustainability
Patino-Martinez, J., L. Dos Passos, I. Afonso, et al. 2022. Globally important loggerhead sea turtle conservation refuge: Maio island, Cabo Verde. Oryx
Patino-Martinez, J., L. Dos Passos, E. Dos Reis, et al. 2020. Integrating local leaders in efforts to improve biodiversity conservation. African Sea Turtle Newsletter
Petty, R. E., Z.L. Tormala, P. Brinol, et al. 2006. Implicit ambivalence from attitude change: an exploration of the PAST model. Journal of personality and social psychology
Putnick, D.L. and M.H. Bornstein. 2016. Measurement invariance conventions and reporting: the state of the art and future directions for psychological research. Developmental review
R Core Team. 2020. R: a language and environment for statistical computing [computer software]. Version 1.1.546. Vienna, Austria: R Foundation for Statistical Computing.
Rosseel, Y. 2012. lavaan: an R package for structural equation modeling. Journal of Statistical Software
Said, H., B.B. Badru, and M. Shahid. 2011. Confirmatory factor analysis (CFA) for testing validity and reliability instrument in the study of education. Australian Journal of Basic and Applied Sciences
Saldaña, J. 2021. The coding manual for qualitative researchers
. Thousand Oaks, CA: Sage Publishing.
Saunders, C. D. 2003. The emerging field of conservation psychology. Human ecology review
't Sas-Rolfes, M., D.W. Challender, A. Hinsley, et al. 2019. Illegal wildlife trade: scale, processes, and governance. Annual Review of Environment and Resources
Selinske, M., G. Garrard, S. Bekessy, et al. 2018. Revisiting the promise of conservation psychology. Conservation Biology
Schwartz, S.H. 1977. Normative influences on altruism. In: Advances in experimental social psychology
(ed. L. Berkowitz). Volume 10. Pp. 221–279. Chicago, MA: Academic Press.
Skogen, K., I. Mauz, and O. Krange. 2008. Cry wolf!: narratives of wolf recovery in France and Norway. Rural Sociology
Steinmetz, R., S. Srirattanaporn, J. Mor-Tip, et al. 2014. Can community outreach alleviate poaching pressure and recover wildlife in South-East Asian protected areas? Journal of Applied Ecology
Thomas-Walters, L., D. Veríssimo, E. Gadsby, et al. 2020. Taking a more nuanced look at behavior change for demand reduction in the illegal wildlife trade. Conservation Science and Practice
Thurstan, R.H., S.M. Buckley, J.C. Ortiz, et al. 2016. Setting the record straight: assessing the reliability of retrospective accounts of change. Conservation Letters
van den Broek, K.L., I. Walker, and C.A. Klöckner. 2019. Drivers of energy saving behaviour: the relative influence of intentional, normative, situational and habitual processes. Energy Policy
VERBI Software. 2020. MAXQDA 2020 [computer software].Version 2020.2. Berlin, Germany: VERBI Software.
Veríssimo, D., S. Vieira, D. Monteiro, et al. 2020. Audience research as a cornerstone of demand management interventions for illegal wildlife products: demarketing sea turtle meat and eggs. Conservation Science and Practice
Veríssimo, D. and A.K. Wan. 2018. Characterizing efforts to reduce consumer demand for wildlife products. Conservation Biology
Wallen, K.E. and E. Daut. 2018. The challenge and opportunity of behaviour change methods and frameworks to reduce demand for illegal wildlife. Nature Conservation
Wallen, K.E. and A.C. Landon. 2020. Systematic map of conservation psychology. Conservation Biology
[Figure 1], [Figure 2], [Figure 3], [Figure 4]