Bananas, pesticides and health in southwestern Ecuador: A scalar narrative approach to targeting public health responses

Social Science & Medicine 150 (2016) 184e191

Contents lists available at ScienceDirect

Social Science & Medicine journal homepage: www.elsevier.com/locate/socscimed

Bananas, pesticides and health in southwestern Ecuador: A scalar narrative approach to targeting public health responses Benjamin Brisbois School of Population and Public Health, University of British Columbia, 2206 East Mall, Vancouver, BC, V6T 1Z3, Canada

a r t i c l e i n f o

a b s t r a c t

Article history: Received 6 May 2015 Received in revised form 9 November 2015 Accepted 17 December 2015 Available online 20 December 2015

Public health responses to agricultural pesticide exposure are often informed by ethnographic or other qualitative studies of pesticide risk perception. In addition to highlighting the importance of structural determinants of exposure, such studies can identify the specific scales at which pesticide-exposed individuals locate responsibility for their health issues, with implications for study and intervention design. In this study, an ethnographic approach was employed to map scalar features within explanatory narratives of pesticides and health in Ecuador's banana-producing El Oro province. Unstructured observation, 14 key informant interviews and 15 in-depth semi-structured interviews were carried out during 8 months of fieldwork in 2011e2013. Analysis of interview data was informed by human geographic literature on the social construction of scale. Individual-focused narratives of some participants highlighted characteristics such as carelessness and ignorance, leading to suggestions for educational interventions. More structural explanations invoked farm-scale processes, such as uncontrolled aerial fumigations on plantations owned by elites. Organization into cooperatives helped to protect small-scale farmers from ‘deadly’ banana markets, which in turn were linked to the Ecuadorian nation-state and actors in the banana-consuming world. These scalar elements interacted in complex ways that appear linked to social class, as more well-off individuals frequently attributed the health problems of other (poorer) people to individual behaviours, while providing more structural explanations of their own difficulties. Such individualizing narratives may help to stabilize inequitable social structures. Research implications of this study include the possibility of using scale-focused qualitative research to generate theory and candidate levels for multi-level models. Equity implications include a need for public health researchers planning interventions to engage with scale-linked inequities, such as disparities within nation-states. Finally, the prominence of the global North in explanatory narratives is a useful reminder that ‘structural factors’ prominently include inequities related to the legacies of colonialism. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Ecuador Scale Narrative Bananas Pesticides Ethnography Structural Risk perception

1. Pesticides, structural factors and qualitative health research Responses to agricultural pesticide exposure, in the global South and elsewhere, are generally divided between individual behaviour-focused and ‘structural’ approaches. As an example of the former, a risk perception study funded by the pesticide giant Syngenta stated that ‘[t]he problem is whether it will be possible to change farmers’ attitudes to improve the way they use pesticides' (Matthews, 2008, p. 845). Educational interventions, however, have been strongly criticized for their limited effectiveness when compared to upstream interventions in the workplace, or national

E-mail address: [email protected]. http://dx.doi.org/10.1016/j.socscimed.2015.12.026 0277-9536/© 2015 Elsevier Ltd. All rights reserved.

or international pesticide-control policies (Konradsen et al., 2003; Murray and Taylor, 2000). Dangerous practices have been repeatedly observed among farmers and workers with ostensibly good knowledge of pesticides and their health effects (Galt, 2013), and increasing numbers of studies support structural intervention strategies. This recognition reflects public health's broader engagement with structural factors, defined by Shannon et al. (2014) as ‘factors that are external to the individual and operate outside the locus of control of individuals’ (p. 175). Multi-level modelling and neighbourhood effects research respond directly to the challenge of understanding structural factors (O'Campo, 2003). For example, Cole et al. (2011) modelled individual/household and community determinants of pesticide-related health outcomes in the Ecuadorian Andes and found a significant effect of community-level poverty. In this and numerous other studies,

B. Brisbois / Social Science & Medicine 150 (2016) 184e191

explicit attention to structural determinants often involves political and economic explanations for health problems. Increasingly sophisticated scholarship goes well beyond simplistic ‘proximal/ distal’ distinctions in showing health to be affected by forces and actors from the genetic to the global (Krieger, 2008). Structural approaches appear to have informed a growing number of qualitative pesticide risk perception studies, often carried out to inform larger research and intervention projects cognizant of political and economic causes of pesticide risks. By employing ethnographic and other in-depth methods to the question of how pesticide-exposed people understand the risks to their health, these risk perception studies have also shown that political and economic factors affect pesticide risk perception (e.g.
n et al., 2001; Barraza et al., 2011; Hunt et al., 1999; MeraArago Orces, 2001; Orozco et al., 2009; Ríos-Gonz
alez et al., 2013; Salazar et al., 2004). Such studies show the embeddedness of pesticide risk perception in social realities such as relationships with nature, gender roles and poverty. Links between structure and risk perception are especially evident in a study on highland Ecuadorian potato production, where pesticide exposure was attributed by employers to worker carelessness, and suicide using pesticides to insanity (Mera-Orces, 2001). Workers, in contrast, characterized pesticide exposure as an occupational hazard, and suicide as an act of desperation. Research in southern Mexico similarly found attribution of responsibility for pesticide exposure to depend on position in the labour process, with landless workers, small farmers and owners of large farms tending to attribute blame differently (Rios-Gonzalez et al., 2013). Such ethnographic studies can generate hypotheses to help target quantitative studies, flesh out their results, and facilitate more effective community-based interventions (Behague, 2008; Janes et al., 1986). They can also provide a valuable corrective to top-down or paternalistic public health strategies (Trostle, 2005), helping to illustrate the cultural and political and economic logic behind allegedly ‘unscientific’ health beliefs (e.g. Briggs, 2004). Yet in applying qualitative methods to the challenge of pesticide exposure, with its acknowledged multi-scalar roots, research to date has largely missed the opportunity to document the bottomup scalar reasoning of individuals experiencing the exposures in question. Barraza et al.'s (2011) ethnographic study in a Costa Rican banana region, for example, is among the most structure-conscious of such qualitative pesticide risk perception approaches, recommending public, non-profit and private-sector collaboration to go ‘far beyond’ educational-behavioural pesticide safety intervention approaches' (Barraza et al., 2011, p. 716). The paper further highlights 'community, regional and national' levels and 'multi-national' actors in discussing the etiology of, and appropriate solutions to, pesticide exposures (p. 716). This scalar discussion appears to represent the authors' political and economic assessment, however, rather than that of study participants. Other studies similarly refer in passing to 'international macroeconomic policies'
n et al., 2001, p. 300) and 'macroeconomic forces' (Mera(Arago Orces, 2001, p. 38) in explaining pesticide exposures, but do not root such macro-scale references in the words of study participants. Qualitative pesticide risk perception studies thus frequently draw on scalar terminology in explaining structural health determinants, but have not yet examined how structural factors are divided up in scalar terms by the people experiencing the health impacts in question. In this paper, I map the scales at which pesticide-exposed residents of Ecuador's banana-producing El Oro province locate causes of, and appropriate responses to, pesticide exposures and other health problems. I draw on ‘social construction of scale’ approaches in human geography (Marston et al., 2005), showing how they can complement perceptive structural, political economic or multi-level responses to health issues using scale-

185

focused input from those experiencing those issues most directly. 2. Theories of scale and health Health research on structural influences e pesticide-focused or otherwise e has, to this point, emphasized the scales at which health determinants are actually located. It has had less to say on what different actors stand to gain or lose by portraying e ‘socially constructing’ e different scales as more or less important in the causation of health problems. Though ‘illness narratives’ work in anthropology (Farmer and Good, 1991) has documented social struggles in which individual-focused explanations of health problems are countered by more structural narratives, specific scalar features of structure-focused accounts in this body of work have been left largely unexamined (although see Briggs, 2004 on multi-scalar resistance to individualizing cholera-blame narratives in Venezuela). However, as human geographers have demonstrated, ‘scale politics’ e struggles to define how the world is, or should be, divided up in scalar terms e have major ramifications in terms of equity and power relationships (Smith, 1992; Herod, 2010). Swyngedouw (1997, p. 139) explains that ‘Scalar narratives … provide the metaphors for the construction of “explanatory” discourses … scale-related explanations define and suggest different ideological and political positions’. Harris (2011), for example, found that individual responsibility for environmental protection was often privileged in narratives recounted by Turkish citizens and environmental activists, resonating with the individual-focused political climate accompanying Turkey's entry into the European Union. Masuda et al. (2012), similarly, found chronic disease prevention strategies in three Canadian provinces to deploy individual or 'collectivist' accounts of responsibility for improving health, roughly corresponding to provincial ideological climates. Several studies have also examined the scale politics involved in contesting scientific credibility and public health priorities. Edge and Eyles (2014) documented how an alleged lack of laboratoryscale evidence allows scientists and regulators to dismiss the possibility of ecosystem-scale effects of the endocrine disruptor bisphenol-A. The laboratory scale also features prominently in the ‘rescaling’ of global concerns such as deforestation and urbanization into a perceived need for laboratory research by U.S. emerging disease researchers in the 1990s (King, 2004). Anderson (2014) description of such ‘scale making in biomedicine’ (p. 372), in addition, problematizes global health's tendency to gloss over messy political legacies of colonialism. Such a glossing over allows the portrayal of scientific knowledge on HIV/AIDS as ‘global’ and therefore authoritative to justify global health interventions in the lives of ‘local’ people around the world (Campbell et al., 2012). The relevance of such scalar considerations for pesticide exposure reduction is further highlighted by Harrison's (2006) analysis of pesticide drift activism in California, in which the scale where responsibility for pesticide exposure was located determined where public sector responses were targeted (individual decisions of farmers, for example, as opposed to the state's regulatory apparatus). As such analyses illustrate, scalar arguments e featuring individuals, laboratories, ecosystems, states, ‘the local’ or ‘the global’ e help to determine which responses are considered appropriate for specific health problems. 3. Methods 3.1. Setting I employed an ethnographic approach to document scalar features in narratives of health and illness voiced by pesticide-exposed

186

B. Brisbois / Social Science & Medicine 150 (2016) 184e191

individuals in Ecuador's southwestern El Oro province. El Oro accounts for approximately 1/3 of banana production in Ecuador (the world's leading banana exporter for approximately half a century) and its capital city Machala is the self-described ‘banana capital of the world’ (Astudillo Samaniego, 2009). The province is an appropriate setting for scale-focused analyses, as actors from the molecular to the global are widely recognized to impact how bananas are produced, and the well-being of the people producing them (Brisbois, 2011). Conventional banana production occurs in monocultures, where vulnerability to plant pathogens e especially the devastating ‘sigatoka negra’ fungus e necessitates heavy pesticide use. Applications with both backpack sprayers and airplanes often deviate from legislated methods, resulting in widespread pesticide exposure and suspected dermatologic, respiratory and reproductive health effects in both individual workers and surrounding communities e albeit with little conclusive evidence on population-level effects of the pesticides in current use (Harari et al., 2011). Importantly, the factors typically associated with pesticide exposures from banana production e poverty, unsafe and precarious work, and unhealthy neighbourhoods e suggest that pesticide exposure is a proxy for several social determinants pathways to poor health. While the Ecuadorian government has recently implemented reforms, banana production typically follows a precarious-labour model in which workers are hired by the day and provided with neither benefits nor safe working conditions (Martínez Valle, 2004; Striffler, 2002). This model is conditioned by the ability of transnational fruit companies to source bananas from different countries in the Americas when national policies, labour organization or agroecological conditions cut into their profits (Frundt, 2009). Numerous small-scale, independent producers, often organized into cooperatives, compete with large plantations in El Oro, in contrast to other Ecuadorian provinces and Central America where large farms are dominant (Astudillo Samaniego, 2009). El Oro thus presents a setting where structural factors at various scales influence e and are perceived to influence e local biophysical and social environments and their probable health consequences. 3.2. Ethnographic approach While ethnographic methods can gather data to inform etiologic reasoning, in this study ethnography was used in its classical sense to understand cultural features such as ways of understanding the world (Schensul and LeCompte, 2012). I began the study with a pesticide risk perception conceptual model developed from existing literature, hypothesizing that risk perception depends on factors such as gender, race, profession and class (Satterfield et al., 2004), and is consistent with overall ways of making sense of the world (Tansey, 2004). Medical anthropological scholarship further suggests that these ways of making sense of the world reflect political economic power dynamics, but not in a straightforward or deterministic manner (Farmer and Good, 1991; Kleinman and Kleinman, 1991). I used this model to inform a series of ethnographic techniques: unstructured observation and informal interviews; key informant interviews; and in-depth semi-structured interviews (Schensul and LeCompte, 2012). Research took place over 2011e2013, involving 8 months of fieldwork in Machala and surrounding banana-producing communities. Snowball and purposive sampling were used to select interviewees for both key informant and semi-structured interviews, targeting demographic categories suggested by the evolving conceptual model. The 14 participants in key informant interviews covered the following categories: health professional; clergy; large banana farm owner; labour leader; cultural organization leader; university professor; government employee; and social movement leader. Early

unstructured observation and key informant interviews helped me refine an interview guide for in-depth semi-structured interviews, the primary source for this paper's analysis, with fieldnotes and key informant interview notes allowing triangulation of results. I conducted 15 semi-structured interviews with a total of 30 participants (four of the interviews had multiple participants). These participants were a mix of banana industry employees (16) and small or medium-scale farm owners (14), including both men (23) and women (7), with two participants residing in a different bananaproducing province. Initial questions asked respondents to explain the meaning of health to them, and to explain why they, and the people around them, were healthy or not. Specific questions on pesticides asked whether they had health or environmental effects, and also what factors caused or led to any such effects. I also asked what should be done in response to these problems, and whose responsibility it was to carry out these responses. These questions were asked in a way that was intentionally agnostic as to scale, letting participants identify the scales at which problems, their causes, and their potential solutions were located. Preliminary fieldwork and key informant interviews had made it clear, however, that actors in the global North e supermarket chains, fruit multinationals, fairtrade organizations, etc. e play major, and widely-recognized, roles in banana production in El Oro. As a result, I included a question near the end of interviews asking if there were ‘changes at the international level that could help with the challenges facing people in the Ecuadorian coast’. Semi-structured interviews were recorded and transcribed in Spanish (by the bilingual author), with selected quotes translated into English for inclusion in this paper, and data analysis employed QSR Nvivo, V10. The study was approved by the University of British Columbia's behavioural research ethics board. 3.3. Methodological concerns Coding and interpretation of data were attentive to how participants divided the world in scalar terms and the interrelations between, and political connotations of, scalar elements. Concern has been raised that scale-focused analyses tend to reify particular scalar configurations as static and/or actually ordered in rigid hierarchies, as opposed to fluid and discursively ordered by social actors (Marston et al., 2005). Methodological innovations responding to these criticisms include recognition of scale politics as performative, with social actors simultaneously describing a scalar situation, and attempting to make that description true (Kaiser and Nikiforova, 2008). The primary ‘stage’ for performative citation of scale in this study was a research interview, meaning that the results simplify a more complex politics of scale as enacted in daily life. However, social scientists have increasingly come to acknowledge that 'the performance of narrative, and its reading, is a collaborative endeavour' (Garro and Mattingly, 2000, pp. 29e30). Data analysis was cognizant of both interview circumstances and possible effects of my social location as a white, male, Northern graduate student. Interview participants tended to be slightly or significantly older than me (between 21 and 80þ, with most between 40 and 70 years of age), and less formally educated (16 participants had no more than primary education). Most were less affluent and all were Ecuadorian, although sometimes with personal or professional ties to Europe or North America through the banana industry. In keeping with best practice in ethnographic research (Schensul and LeCompte, 2012), I used a field journal to carefully document my thoughts, feelings, and interactions with people in El Oro. Additional steps taken to ensure methodological rigour (cf. Golafshani, 2003; Morse et al., 2002) included triangulation between the different types of ethnographic data and between types

B. Brisbois / Social Science & Medicine 150 (2016) 184e191

of participant; feedback from interview participants on preliminary results; and concurrent data collection and analysis to ensure iterative interaction between data and analytically-derived theories. The use of snowball and targeted sampling means that my findings do not represent all voices in El Oro. On average, participants were more educated and affluent than the majority of banana farm workers. I was unable to conduct interviews with some vulnerable sectors of the banana workforce, such as women employed on large farms. Sustained discussion of gender, race and ethnic difference was also beyond the scope of this study. 4. Results Narratives drawn from interview data exhibited several recurrent scalar themes concerning the causes of good or poor health, and appropriate responses to pesticide exposure and other health problems. Individual-focused explanations were common, but not dominant. Other prominent scalar elements included farms; the cooperatives into which small banana farmers often organize; Ecuador and other nation-states; and the banana-consuming world (global North). All emphasis in quotes used to substantiate these themes has been added to illustrate specific scalar narrative elements. 4.1. Individuals and their bodies Many descriptions of health and its determinants attributed responsibility primarily to individuals. Health or illness in such narratives was produced by behaviours such as healthy eating and avoiding drugs, alcohol and cigarettes. As one male banana worker explained, ‘For me, health begins with the personal hygiene of each person … My health is good. I'm a person who doesn't destroy my body [mi organismo].’ The female university-trained accountant for a small producer cooperative also referred to ‘things that can damage our body, like alcohol and drugs‘. Such usages, which were voiced across demographic categories, specifically highlight (‘construct’) the body as a scale over which an individual person has jurisdiction. Consistent with this construction, pesticide exposure was attributed by many participants to carelessness, including improper use of personal protective equipment (PPE) such as gloves and facemasks. While a small number of participants admitted to being careless with PPE themselves, many cited the carelessness of others as a reason for those others' pesticide exposure. In the words of a small banana farmer with some university education: Sometimes, the guy going to apply pesticides doesn't want to, because it bothers him, because it's suffocating, because it's uncomfortable with the safety glasses, no? So, it's also a question of culture, no? They know that it does damage, but they don't protect themselves. And why? Because there's no culture. There's no education about this. In this and other evocations of culture, individual failure to take health-protecting actions is interpreted as reflecting a lack of education (‘cultura’ in Spanish often signifies ‘education’ or ‘refinement’). Importantly, this farmer's explanations of pesticide exposure also included a number of other scalar elements. He discussed the problems facing small producers due to international market conditions and a lack of Ecuadorian government support, and spoke heatedly of the tendency for industrialized countries (such as my own) to export pesticides that had been banned in the global North to 'underdeveloped' countries such as Ecuador. This farmer's focus on ‘culture’ was therefore not a blanket attribution of

187

causality to individual factors. It nevertheless evokes a common characterization I encountered in interviews, and in numerous other conversations over the course of my fieldwork with educated or affluent Ecuadorians. In this depiction, individuals who work in bananas e often poor, rural and/or young people e abuse alcohol and drugs, and do not adequately care for their children or protect themselves from pesticides. As expressed by another small banana producer, a woman with some university education: Near where I live, it's full of banana farms. That's the zone … pure delinquency. You pay people their week, but people don't know how to make use of their money. They just devote themselves to drinking, and their children also end up drinking. You learn what you see. And they don't pay attention to feeding themselves, and they don't protect themselves, because it's necessary to use certain chemical products, right? Over time, this affects your health. You give them gloves and everything, but because of their ignorance, they don't protect themselves. This focus on individual characteristics such as ignorance and a tendency not to protect oneself led seamlessly into a call for educational responses: Design some talks, so they can be guided, told that if they keep drinking, alcohol can give them cirrhosis. Talks to motivate them to have better health. And these talks would be, practically speaking, for people in the countryside. Lots of times people don't listen to the advice you give them. They say: what's this got to do with me, why are you interfering in our lives? So I don't know how to arrive at these people, to save them, no? Of the 11 semi-structured interview participants who attributed pesticide exposures and other health problems to the individual choices of less-privileged others, and suggested education as a response, 9 owned farms and/or had some postsecondary education. The two exceptions, both banana workers, also had at least a measure of relative privilege in that one was a part-time health promoter and the other a labour leader. Thus the scale of an individual with appropriate jurisdiction over their body's health was frequently performed in interviews in ways that appear linked e though not straightforwardly e to relations between social classes.

4.2. Farms Another scale frequently arising in explanations of pesticide exposure and other problems facing Orenses was that of the farm. Numerous interviewees implicated actions (or lack of actions) by large farm owners and managers with control over farm-scale decisions and processes. Some participants e often landless workers e highlighted the failure of banana farm owners and managers to provide PPE and training for employees. One worker without land or postsecondary education described the exposure resulting from aerial fumigations near populated areas and in farms while workers were still in the fields, an illegal but common practice. When the airplane fumigates, people are inside the plantations. You have to leave for an hour or a half-hour, but it's not being observed. Why? It's clearly related to the interests of the owners of the plantations e especially the plantations that fumigate with airplanes. It isn't convenient for the boss that his people lose an hour or two, so they have to stay in the plantations when the airplanes are fumigating. This quote employs a scalar vocabulary in which ‘the plantation’

188

B. Brisbois / Social Science & Medicine 150 (2016) 184e191

is a container for social processes (‘people are inside the plantations’), but is also an agent (‘plantations that fumigate’). Importantly, both of these usages have embedded political content, in that the plantation is controlled by ‘the owner’ or ‘the boss’. In contrast to this focus on large farms or plantations, however, the frequent self-identification of many participants as small-scale ~ os productores] invoked a specifically banana farmers [pequen small farm scale: The big farm owners, they can certify their farm [with GlobalGAP, a good-agricultural-practices certification]. But when you're small, how do you do this? We don't have enough. There's a lot of poverty here, and they're making us [certify our farms]. It looks to me like it's a political question, that the owners, the big farm owners, they want the small producers to disappear. This description identifies a farmer with farm size (‘when you're small’), and clearly invests the scale of the farm with political significance: big farm owners want the small producers to disappear. References to small banana farms were common in interviews, news stories, conversations, and in the specific names and composition of various organizations. The exact size of a ‘small’ ~ o] farm appears to vary: for example, one Ministry of [pequen Agriculture document classifies farms up to 50 ha (ha) as small, but offers technical support to those under 30 ha (Ministerio de Agricultura, Ganadería, Acuacultura y Pesca, 2015). Many small producers I encountered owned less than 5 ha, and some had as little as a single hectare. In spite of this diversity, enough commonality existed to motivate political strategies. Numerous newspaper articles decried the adverse market conditions facing small producers, while the existence of numerous small-producer cooperatives and organizations such as the ‘Front for the Salvation of Small-Scale Producers’ further illustrates the political importance of the small farm scale. With respect to this paper's goal of mapping the scalar dimensions of pesticide risk perception, it is clear that farm-level processes are important elements of Orense scalar narratives. 4.3. Cooperatives In particular, the small farm scale's political content involved linkages to the scale of the cooperative. While construction of the small farm scale typically employed measures of areal extent (i.e. farm size), the cooperative represents a scalar configuration composed of multiple small farmers, and e by extension e the combined areal extent and banana production capacity of their farms. The importance of this scale, and its relationship to allegedly individual entities such as personal health, are suggested by one female small producer: ‘When I'm healthy, I can work. I can have motivation for anything: to keep living, to help with the organization [cooperative]'. Several small producers also mentioned the ‘motivation’ provided by their cooperatives' secure fairtrade contracts in the face of recurrent economic crises in the open banana market. The cooperative, therefore, represents a scalar element with considerable positive political significance for many producers. 4.4. Ecuador and other nation-states Another frequently-invoked scale in interviews was that of the nation-state, typically Ecuador. Several participants, for example, expressed their opinion that Ecuador's bananas were both bettertasting and ‘less contaminating’ than those of rival countries. The national scale was also invoked by one landless worker's identification of ‘the president of the republic’ as the most appropriate

actor in addressing health problems. The challenges involved in relying on the government, however, were emphasized by another landless worker (also invoking the plantation-as-container scale, with its political linkages to farm owners): Here it would be the authorities, the government, who would get involved in the [problem of aerial fumigations], ideally with some drastic measures, especially where there are settlements inside or near the plantations. They should make a space of some 200 or 300 m, and not fumigate there. You can say it, but because you're not the farm owner … they buy justice, buy the authorities. It's difficult. The rich have the power. The relationship of the nation-state scale to small producer cooperatives was also illustrated by ongoing negotiations involving the Ecuadorian government, small producer cooperatives and large banana companies to fix the official minimum price of a case of bananas. Another link is illustrated by one participant's suggestion that ‘big countries’ such as my own bypass the large banana companies and buy directly from Ecuadorian cooperatives. Invocations of the nation-state by small farmers similarly often involved discussions of power and inequity: The government should first focus on work, on agriculture. Because from that comes the food of a whole people, and exports also. Here the banana exporting companies do whatever they feel like. They pay, they don't pay. They raise the price when they want to, when they feel like it. There's no respect for our [small producers'] signed contracts. They don't pay the official price, which is in the contract. And the government can't do anything, because they're millionaires. And they pay money under the table, and the poor people … we're always screwed [jodidos]. Importantly, such scalar narratives link small producers and their cooperatives to the Ecuadorian government in a politically meaningful story featuring the nation-state as an appropriate-butineffective actor (transnational banana exporting companies are also featured, a theme I take up in the following section). In other cases, however, the Ecuadorian state was portrayed in a negative light by some farmers because of their difficulties meeting new requirements that they make social security payments for their workers. In any case, the role (or absence) of the Ecuadorian state in influencing poverty and toxic exposures was a common feature of participants' narratives. 4.5. The banana-consuming world As suggested by the above reference to ‘big countries’ such as my own, participants often also had well-developed ideas concerning actors identified with the banana-consuming world (a region that largely mapped onto the global North). Establishing organic or fairtrade market niches for Ecuadorian producers to bypass the large banana exporters is a common transnational strategies of cooperatives, for example, combined with ‘educating’ foreign banana consumers in a bid to influence market, health and environmental conditions in Ecuador. One landless worker, similarly, urged me to apply pressure on my own government to improve conditions in Ecuador: Listen, I think that it would be more practical if you would take this message direct to your government. And, in turn, your government would have contact with the Ecuadorian government, so that the two parts … one selling and the other exporting … then we'd see a better exchange of ideas, to improve these things.

B. Brisbois / Social Science & Medicine 150 (2016) 184e191

In this passage, a sellers/exporters scalar divide is employed, with different national governments (and me, the foreign researcher) linked to different sides of the divide. While some representations of the global North emerged in response to my specific questions about it, this scalar element was also evident in numerous settings outside of my control as an interviewer, such as my unstructured observation. Indeed, my very presence in El Oro appeared to invoke the scale of the banana-consuming world, for example in conversations with participants prior to the start of interviews. Judging by both interviews and my reception as a foreign researcher in Machala, the banana-consuming world is populated with wealthy individuals who have the ability to provide various kinds of assistance to individuals in El Oro, through their banana purchasing decisions but also as conduits to a generally better-resourced part of the world. I have an idea, now that you're here. You come from another country. We want to have a factory here. If you support us, the small producers, so much the better! With machinery, we could make juice, sweets, so many things, and give them to the schools. I always tell the friends who come from other countries, other gringos. Gringos, that's what we call you. You should give some support here. With the president [of the cooperative], we're also supporting you! Here I am classified by a primary school-educated small producer about my own age as belonging to the privileged collection of countries where gringos originate, and where much of the power in the banana industry resides (he also identifies himself as a small producer in a cooperative). My education and age likely also informed this portrayal of me as a foreigner with ‘more knowledge’, ~ ero]. On other occasions, local newsbut also as a ‘friend’ [compan papers would highlight the nationality of foreign attendees at events, including me on one occasion. Foreigners so highlighted were often from Europe or North America, but also from slightly larger and wealthier (and banana-importing) Latin American countries such as Argentina or Chile. Importantly, the specific composition and political role of the scale of the banana-producing world differed between participants, in ways that e as the interaction described above suggests e likely interacted with factors such as gender, age and education. In contrast to the request for help made by the relatively marginal small producer quoted immediately above, for example, another small producer with leadership experience in his cooperative bristled at the implication that a gringo such as myself might offer him an incentive to participate in my study (this was a misunderstanding, as I offered no incentives and said so during the informed consent process). Possible reasons for this intense reaction include the fact that, in addition to coming from the global North, I was about a decade younger than him and only slightly more educated (he had some university training and seemed to pride himself on his authoritative knowledge of the banana industry). Whether the role of actors from the banana-consuming world was seen as positive or negative, however, the banana-consuming world clearly emerged as a scale of importance in the daily lives of many, if not most, Orenses. 5. Discussion The above examples highlight several prominent scalar elements in the narratives of people experiencing pesticide exposures related to banana production. Individual-scale explanations were common, but also a variety of more-than-individual scales: farms; the cooperatives into which small farmers organize; the Ecuadorian nation-state; and the banana-consuming world. Such scalar

189

elements combined in complex ways that appear linked to social class, but not in a straightforward manner. Narratives locating responsibility for the health problems of (poorer) others primarily in individual characteristics, for example, were largely but not exclusively recounted by landowners with at least some postsecondary education. These individuals would often also implicate structural factors in explaining their own problems, however, while narratives of landless workers would sometimes highlight individual characteristics. The finding that individual and structural explanations for health problems were deployed differently across gradients of income and education (‘class’) is consistent with several comparable qualitative pesticide risk perception studies done in Latin America (e.g. Barraza et al., 2011; Mera-Orces, 2001; Rios-Gonzalez et al., 2013). Narratives highlighting limits to farmer and worker agency are also congruent with Hunt et al.'s (1999) ethnographic risk perception study in southern Mexico. Similarly, Phillips (2004) observed individual-scale emphases on personal responsibility (hygiene) in official public health responses to the 1991 cholera epidemic in coastal Ecuador, in contrast to peasant perspectives highlighting structural determinants (i.e. poverty). In Briggs (2004) study of the same cholera pandemic's appearance in Venezuela, narratives of Venezuelan public health officials tended to attribute blame for cholera to individual behaviours linked with Indigenous 'culture'. Interpretations circulating among Indigenous communities in the region, in contrast, indicted dominant actors such as the mestizo-dominated Venezuelan state, and even global forces such as US imperialism. Garro (1995) also found that explanations for diabetes in a Canadian Anishinaabe community typically fell on a continuum between individual factors (e.g. diet and exercise) or more 'societal' ones such as colonial changes to traditional food acquisition practices. Narratives in the present study fit this pattern, but also highlight specific scalar elements along such an individual-to-societal continuum and their apparent relation to the social class of the people recounting them. Public health implications of these scalar elements relate to both research and practice. While the goal of the study was to map perceptions of pesticide risk, the detailed ‘lay multi-level models’ of numerous participants suggest that scalar narratives may provide at least possible causal mechanisms and candidate levels for multilevel epidemiologic studies. While social and environmental epidemiologists draw on various theoretical and empirical sources to construct and test multi-level models (O'Campo, 2003), this study suggests that people actually experiencing the health impacts in question may be fruitful sources of etiologic hypotheses. The scalar political implications of such narratives may also provide equityfocused guidance for interventions, such as the participatory pesticide-health projects increasingly advocated in the literature (Galt, 2013; Orozco et al., 2009). Numerous participants, for example, described struggles among large-scale banana farms (and farmers), small producers and their cooperatives, the Ecuadorian state, and transnational actors such as fruit multinationals and supermarket chains. Interventions to support the efforts of small producer cooperatives to challenge inequitable market conditions, however, would need to acknowledge that the political content of the cooperative scale is not uniformly progressive. Many depictions of landless workers as ignorant or careless in the present study were voiced by educated small producers in cooperatives, who would attribute their own difficulties to more structural factors such as transnational banana companies or corruption in government. Consistent with a documented reluctance of Ecuadorian fairtrade cooperatives to support worker unionization (Frundt, 2009), furthermore, no small farmer participants mentioned unionization as a possible response to unsafe or precarious work. The ‘victim-blaming’ narratives of several educated small farmers

190

B. Brisbois / Social Science & Medicine 150 (2016) 184e191

therefore illustrate potential obstacles to equitable participatory interventions, and suggest the need to draw on perspectives from e at the very least e small farmers and landless workers. Similarly problematic are the political implications of national scalar elements, consistent with the observation that taking the scale of the nation-state for granted may obscure its role in generating inequities (Wainwright, 2005). National unity in Ecuador, with its numerous cultures and prominent regional differences, has historically been promoted by elites as a way of advancing their economic and political interests (Whitten, 2003). Profits from banana production in rural coastal areas of Ecuador have disproportionately flowed to metropolitan centres such as Guayaquil and Quito, fuelling development of bureaucracies and infrastructure in the capital (Larrea et al., 1987). As an example of the relevance of such historical context to the present study, I attended a workshop in Machala in which a visiting Ministry of Public Health physician from Quito presented a new national approach to reducing pesticide exposures through measures that included improving the ‘self esteem’ [autoestima] of banana labourers e a group left impoverished by historical transfers of resources from the banana-producing coast to the capital. This point holds special relevance for the field of global health, which prides itself on equitable partnerships with Southern partners (a category that could include public health officials from Quito) but which may frequently underestimate how regional, class and racial inequities within countries of the global South complicate this picture (Crane, 2013). Implications for research and practice also stem from the demonstrated awareness of the world as divided into bananaproducing and -consuming countries. One such implication stems from the request I received to support small banana farmers (who, they pointed out, were supporting me). Fieldwork in Latin America by North American and European researchers inevitably evokes over a century of inequitable political economic relationships (Sundberg, 2003). Concerns such as these led me to constantly reflect on whether my time in El Oro was likely to improve the lives of the people with whom I was working. I worried that I might gain career advancement by reporting their stories to an academic community, while leaving untouched the structures creating their poverty, and my ability to do research. Such questions are informing a growing critical literature describing the problematic dependence of global health research on NortheSouth resource asymmetries (e.g. Crane, 2013; Janes and Corbett, 2009). As described above, furthermore, these dependencies relate to inequities not only in NortheSouth interactions, but also within the countries that make up both North and South. Highlighting the residual and ongoing effects of colonialism is an important reminder that perceptive work on structural determinants of health can nevertheless pass over, and potentially naturalize, these ongoing inequities (cf. Anderson, 2014). This paper has illustrated how preparatory ethnographic work can provide scalar nuance to inform analytic strategies and equityfocused interventions. In addition, an intriguing and disturbing hypothesis is that narratives blaming poor people for their own poverty and ill health constitute an upstream factor that stabilizes unfair political economic structures. Swyngedouw defines hegemony as ‘the capacity of a dominant group (or an alliance of class factions) to impose a series of social practices at a particular spatial scale that are to its (their) advantage’ (1997, p. 146). Educational pesticide safety interventions, for example, ‘fail forward’ in that they fail to protect farmers and workers from pesticides, but effectively preserve the political economic status quo by undermining possibilities for more structural change (Galt, 2013). The need to support and not undermine such structural change remains an open challenge in public health.

Acknowledgements
sThis paper was improved by feedback from Marta Berbe
zquez, Anne-Emanuelle Birn, Tanya Chung Tiam Fook, Donald Bla
ndez Cervantes, Brie Cole, Mathieu Feagan, Leila Harris, Tania Herna McAloney, Patricia Polo, Jerry Spiegel and Patricia Spittal. Fieldwork costs and living expenses during the study were covered by a Canadian Institutes of Health Research Doctoral Research Award.

References Anderson, W., 2014. Making global health history: the postcolonial worldliness of biomedicine. Soc. Hist. Med. 27 (2), 372e384. http://doi.org/10.1093/shm/ hkt126.
n, A., Arago
n, C., Tho € rn, Å., 2001. Pests, peasants, and pesticides on the Arago northern Nicaraguan Pacific plain. Int. J. Occup. Environ. Health 7 (4), 295e302. http://dx.doi.org/10.1179/107735201800339281. Astudillo Samaniego, C., 2009. Sociedad Orense en el Siglo XX [Orense society in the
n, Machala, 20thCentury]. Casa de la Cultura Ecuatoriana Benjamín Carrio Ecuador. Barraza, D., Jansen, K., van Wendel de Joode, B., Wesseling, C., 2011. Pesticide use in banana and plantain production and risk perception among local actors in Talamanca, Costa Rica. Environ. Res. 111 (5), 708e717. http://dx.doi.org/10.1016/ j.envres.2011.02.009. Behague, D.P., Goncalves, H., Victora, C.G., 2008. Anthropology and epidemiology: learning epistemological lessons through a collaborative venture. Cien. Saude Colet. 13 (6), 1701e1710. Briggs, C.L., 2004. Theorizing modernity conspiratorially: science, scale, and the political economy of public discourse in explanations of a cholera epidemic. Am. Ethnol. 31 (2), 164e187 doi:5/ae.2004.31.2.164. Brisbois, B.W., 2011. Pesticides, people and power in Ecuador's banana industry: participatory epidemiology and political ecology approaches to occupational health and safety. Antipode 43 (1), 176e179. http://doi.org/10.1111/j.1467-8330. 2010.00870.x. Campbell, C., Cornish, F., Skovdal, M., 2012. Local pain, global prescriptions? Using scale to analyse the globalisation of the HIV/AIDS response. Health Place 18 (3), 447e452. Cole, D.C., Orozco, F.A., Ibrahim, S., Wanigaratne, S., 2011. Community and household socioeconomic factors associated with pesticide-using, small farm household members' health: a multi-level, longitudinal analysis. Int. J. Equity Health 10, 54. http://dx.doi.org/10.1186/1475-9276-10-54. Crane, J.T., 2013. Scrambling for Africa: AIDS, Expertise, and the Rise of American Global Health Science. Cornell University Press, Ithaca, NY. Edge, S., Eyles, J., 2014. The socio-spatial (re)configuration of legitimacy, knowledge, and practice in chemical risk governance: integrating boundary-work and scale-frame analytics. Environ. Polit. 23 (2), 282e301. http://dx.doi.org/10.1080/ 09644016.2013.818300. Farmer, P., Good, B.J., 1991. Illness representations in medical anthropology: a critical reviewand a case study of the representation of AIDS in Haiti. In: Skelton, J.A., Croyle, R.T. (Eds.), Mental Representation in Health and Illness. Springer, New York, pp. 132e162. Frundt, H.J., 2009. Fair Bananas! Farmers, Workers, and Consumers Strive to Change an Industry. University of Arizona Press, Tucson, AZ. Galt, R.E., 2013. From Homo economicus to complex subjectivities: reconceptualizing farmers as pesticide users. Antipode 45 (2), 336e356. http://dx.doi.org/10.1111/ j.1467-8330.2012.01000.x. Garro, L.C., 1995. Individual or societal responsibility? Explanations of diabetes in an Anishinaabe (Ojibway) community. Soc. Sci. Med. 40 (1), 37e46. http:// dx.doi.org/10.1016/0277-9536(94)00125-D. Garro, L.C., Mattingly, C., 2000. Narrative as construct and construction. In: Mattingly, C., Garro, L.C. (Eds.), Narrative and the Cultural Construction of Illness and Healing. University of California Press, Berkeley and Los Angeles, pp. 1e49. Golafshani, N., 2003. Understanding reliability and validity in qualitative research. Qual. Rep. 8 (4), 597e607.
n bananera: Impacto en la Harari, R., Harari, H., Harari, N., Harari, F., 2011. Produccio salud y el ambiente [Banana production: Impacts on health and the environment]. FENACLE, FOS, IFA, Quito. Harris, L.M., 2011. Neo (liberal) citizens of Europe: politics, scales, and visibilities of environmental citizenship in contemporary Turkey. Citizsh. Stud. 15 (6e7), 837e859. http://dx.doi.org/10.1080/13621025.2011.600095. Harrison, J.L., 2006. ‘Accidents’ and invisibilities: scaled discourse and the naturalization of regulatory neglect in California's pesticide drift conflict. Polit. Geogr. 25 (5), 506e529. http://dx.doi.org/10.1016/j.polgeo.2006.02.003. Herod, A., 2010. Scale. Routledge, New York, NY. Hunt, L.M., Ojanguren, R.T., Schwartz, N., Halperin, D., 1999. Balancing risks and resources: applying pesticides without using protective equipment in Southern Mexico. In: Hahn, R.A. (Ed.), Anthropology in Public Health. Oxford University Press, New York, pp. 235e254. Janes, C.R., Corbett, K.K., 2009. Anthropology and global health. Annu. Rev. Anthropol. 38, 167e183. http://doi.org/10.1146/annurev-anthro-091908164314.

B. Brisbois / Social Science & Medicine 150 (2016) 184e191 Janes, C.R., Stall, R., Gifford, S.M., 1986. Anthropology and Epidemiology: Interdisciplinary Approaches to the Study of Health and Disease, vol. 9. Kluwer Academic Publishers, Norwell, MA. Kaiser, R., Nikiforova, E., 2008. The performativity of scale: the social construction of scale effects in Narva, Estonia. Environ. Plan. D Soc. Space 26 (3), 537e562. http://dx.doi.org/10.1068/d3307. King, N.B., 2004. The scale politics of emerging diseases. Osiris 19, 62e76. Kleinman, A., Kleinman, J., 1991. Suffering and its professional transformation: toward an ethnography of interpersonal experience. Cult. Med. Psychiatry 15 (3), 275e275. Konradsen, F., van der Hoek, W., Cole, D.C., Hutchinson, G., Daisley, H., Singh, S., Eddleston, M., 2003. Reducing acute poisoning in developing countries e options for restricting the availability of pesticides. Toxicology 192 (2e3), 249e261. http://dx.doi.org/10.1016/S0300-483X(03)00339-1. Krieger, N., 2008. Proximal, distal, and the politics of causation: what's level got to do with it? Am. J. Public Health 98 (2), 221e230. http://dx.doi.org/10.2105/ AJPH.2007.111278. Larrea, C., Espinosa, M., Charvet, P.S., 1987. El banano en el Ecuador: transnacionales,
n y subdesarrollo [The banana in Ecuador: transnationals, modernizacio
n Editora Nacional, Quito, modernization, and underdevelopment]. Corporacio Ecuador. Marston, S.A., Jones, J.P., Woodward, K., 2005. Human geography without scale. Trans. Inst. Br. Geogr. 30 (4), 416e432. Martínez Valle, L., 2004. Trabajo flexible en las nuevas zonas bananeras de Ecuador [Flexible work in Ecuador's new banana zones]. In: Korovkin, T. (Ed.), Efectos
n : Petro
leo, banano y flores en Ecuador. Ediciones sociales de la globalizacio Abya-Yala [u.a.], Quito, pp. 129e156. Masuda, J.R., Robinson, K., Elliott, S.J., Eyles, J., 2012. Chronic disease prevention and the politics of scale: Lessons from Canadian health reform. Soc. Work Public Health 27 (7), 639e657. http://dx.doi.org/10.1080/19371910903182880. Matthews, G.A., 2008. Attitudes and behaviours regarding use of crop protection products - a survey of more than 8500 smallholders in 26 countries. Crop Prot. 27 (3-5), 834e846. Mera-Orces, V., 2001. Paying for survival with health: potato production practices, pesticide use and gender concerns in the Ecuadorian highlands. J. Agric. Educ. Ext. 8 (1), 31e40. http://dx.doi.org/10.1080/13892240185300061. Ministerio de Agricultura, Ganadería, Acuacultura y Pesca, 2015. Proyecto de ~ os productores bananeros. [Project desarrollo de la productividad de los pequen to develop the productivity of small banana producers]. Retrieved from. http:// portalbananero.agricultura.gob.ec/index.php/2013-10-16-16-40-12/programade-desarrollo-de-la-productividad. Morse, J.M., Barrett, M., Mayan, M., Olson, K., Spiers, J., 2002. Verification strategies for establishing reliability and validity in qualitative research. Int. J. Qual. Methods 1 (2), 13e22. Murray, D.L., Taylor, P.L., 2000. Claim no easy victories: evaluating the pesticide

191

industry's global safe use campaign. World Dev. 28 (10), 1735e1749. Orozco, F.A., Cole, D.C., Forbes, G., Kroschel, J., Wanigaratne, S., Arica, D., 2009. Monitoring adherence to the international code of conduct: highly hazardous pesticides in Central Andean agriculture and farmers' rights to health. Int. J. Occup. Environ. Health 15 (3), 255e268. O'Campo, P., 2003. Invited Commentary: advancing theory and methods for multilevel models of residential neighborhoods and health. Am. J. Epidemiol. 157 (1), 9e13. http://doi.org/10.1093/aje/kwf171. Phillips, L., 2004. Changing health moralities in the tropics: ethics and the other. In: Gabriel, B., Ilcan, S. (Eds.), Postmodernism and the Ethical Subject. McGillQueen's University Press, Montreal, pp. 254e272.
lez, A., Jansen, K., Sa
nchez-Pe
rez,, H.J., 2013. Pesticide risk perceptions Ríos-Gonza and the differences between farmers and extensionists: towards a knowledgein-context model. Environ. Res. 124, 43e53. Salazar, M.K., Napolitano, M., Scherer, J.A., McCauley, L.A., 2004. Hispanic adolescent farmworkers' perceptions associated with pesticide exposure. West. J. Nurs. Res. 26 (2), 146e166. http://dx.doi.org/10.1177/0193945903259579. Satterfield, T., Mertz, C., Slovic, P., 2004. Discrimination, vulnerability, and justice in the face of risk. Risk Anal. 24 (1), 115e129. http://dx.doi.org/10.1111/j.02724332.2004.00416.x. Schensul, J.J., LeCompte, M.D., 2012. Essential Ethnographic Methods: a Mixed Methods Approach. Rowman Altamira, Lanham, MD. Shannon, K., Goldenberg, S.M., Deering, K.N., Strathdee, S.A., 2014. HIV infection among female sex workers in concentrated and high prevalence epidemics: why a structural determinants framework is needed. Curr. Opin. HIV AIDS 9 (2), 174e182. http://dx.doi.org/10.1097/COH.0000000000000042. Smith, N., 1992. Contours of a spatialized politics: homeless vehicles and the production of geographical scale. Soc. Text. 33, 55e81. http://dx.doi.org/10.2307/ 466434. Striffler, S., 2002. In the Shadows of State and Capital: The United Fruit Company, Popular Struggle, and Agrarian Restructuring in Ecuador, 1900e1995. Duke University Press, Durham, NC. Sundberg, J., 2003. Masculinist epistemologies and the politics of fieldwork in Latin Americanist geography. Prof. Geogr. 55 (2), 180e190. Swyngedouw, E., 1997. Neither global nor local: “Glocalization”and the politics of scale. In: Cox, K.R. (Ed.), Spaces of Globalization: Reasserting the Power of the Local, pp. 140e142. Tansey, J., 2004. Risk as politics, culture as power. J. Risk Res. 7 (1), 17e32. http:// dx.doi.org/10.1080/1366987042000151188. Trostle, J.A., 2005. Epidemiology and Culture. Cambridge University Press, New York. Wainwright, J., 2005. The geographies of political ecology: after Edward Said. Environ. Plan. A 37, 1033e1043. Whitten, N.E., 2003. Symbolic inversion, the topology of El Mestizaje, and the spaces of Las Razas in Ecuador. J. Lat. Am. Anthropol. 8 (1), 52e85.