Pii: s0277-9536(01)00277-5

Social Science & Medicine 55 (2002) 1325–1336 Researching sexual and reproductive behaviour: a peer ethnographic approach Neil Price*, Kirstan Hawkins1 Centre for Development Studies, University of Wales, Swansea SA2 8PP, UK In recent years, ethnographic research has challenged the notion within demography that fertility-related behaviour is the outcome of individualistic calculations of the costs and benefits of having children. Anthropology has furthercriticised the abstraction in demographic analysis of sexual behaviour and fertility decision-making from the socio-cultural and political context in which the individual or couple is located. Within demography itself, institutional andpolitical-economic analyses have argued strongly that sexual and reproductive behaviour must be understood withinlocally specific social, cultural, economic and political contexts. Positivist and empiricist research methods, such as thesample survey and focus groups, which continue to dominate demographic inquiry and applied research into sexual andreproductive behaviour, have been shown to be limited in their ability to inform about the process of behaviour changeand contexts within which different behaviours occur. The article introduces a new methodology for researching sexualand reproductive behaviour, called the peer ethnographic approach, which the authors have developed in an attempt toaddress some of the limitations of the methods which currently dominate research into sexual and reproductivebehaviour. The peer ethnographic methodology is discussed in detail and the results of recent field-testing are reported,which show that, although the approach has limitations, it also has the potential to make a significant contribution toour understanding of sexual and reproductive behaviour. r 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Sexual and reproductive behaviour; Qualitative research methods; Peer ethnography; Zambia vioural research. After reviewing the dominant theore-tical and methodological approaches to the study of In this article we introduce a new methodology for sexual and reproductive behaviour, we discuss the peer researching sexual and reproductive behaviour, what we ethnographic methodology in detail and present the are calling the ‘peer ethnographic approach'. We have results of recent field-testing. The article concludes with developed the approach in an attempt to address some our reflections on the peer ethnographic approach: of the limitations of the methods that currently although the approach has limitations, it also has the dominate sexual and reproductive health and beha- potential to make a significant contribution to ourunderstanding of sexual and reproductive behaviour.
*Corresponding author:Subject Leader in Population and International Health, Centre for Development Studies, Uni-versity of Wales, Swansea. Tel.:+44-1792-295975; fax:+44- The social context of sexual and reproductive behaviour E-mail addresses: [email protected] (N. Price), Demography has been criticised for paying insufficient [email protected] (K. Hawkins).
attention to the social, cultural and political-economic Formerly lecturer in anthropology and development at the contexts of sexual, and particularly reproductive beha- Centre for Development Studies, is currently a freelanceresearcher and consultant. Address:c/o Centre for Develop- viour (Greenhalgh, 1995; Kertzer & Fricke, 1997; Price ment Studies, University of Wales, Swansea.
& Thomas, 1999), and for failing to take on board 0277-9536/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.
PII:S 0 2 7 7 - 9 5 3 6 ( 0 1 ) 0 0 2 7 7 - 5 N.Price, K.Hawkins / Social Science & Medicine 55 (2002) 1325–1336 methodological advances made in cognate disciplines through the spread of Christianity and modern educa- such as anthropology (Fricke, 1997, p. 249; Greenhalgh, 1990, 1996; Kertzer, 1995, p. 29; McNicoll, 1992, p. 400).
Economic rationalism and cultural diffusionism con- One strand of criticism focuses on the limitations of ceptualise fertility transition as resulting from an economic-rationalist models, in which fertility behaviour evolutionary and unilinear process of development in is understood as driven by attempts to maximise scarce which societies transform from ‘traditional' to modern resources (Easterlin & Crimmins, 1985; Behrman & (or western) forms (Greenhalgh, 1990, 1995, 1996; Price Knowles, 1998). In their conceptualisation of the & Thomas, 1999). This understanding of development is individual or reproductive couple as the fulcrum of predicated upon modernisation theory, which has come reproductive decision-making, such models abstract the under a sustained critique in the social sciences since the decision-making process from the socio-cultural and 1960s as providing an inadequate understanding of the political context in which the reproductive unit is located diversity of development contexts around the world (see (Schneider & Schneider, 1995; Simmons, 1988) and fail Gardner & Lewis, 1996; Price & Thomas, 1999). The to consider the influences of social institutions and social Westernisation hypothesis, which forms the lynch-pin of relations that structure that context (Thomas & Price, cultural diffusion theory, has been increasingly chal- 1999). Anthropological analyses have challenged the lenged by institutional and political-economic demogra- notion that fertility-related behaviour is the outcome of phers who, drawing extensively upon anthropological individualistic calculations of the costs and benefits of theory and method, have demonstrated that sexual and having children (see Price, 1996).
fertility related behaviour is embedded within locally In an apparent move away from an economic- specific institutions and social and political contexts, rationalist understanding of fertility behaviour, cultural which both determine and constrain behaviour change diffusion has gained increasing prominence within (Greenhalgh, 1990; Hammel, 1990; Lockwood, 1995; demography over the past two decades. Diffusionism McNicoll, 1994; Price, 1996; Price, 1998; Price & understands culture as sets of attitudes and values that Thomas, 1999; Schneider & Schneider, 1995).
act as both facilitators to and inhibitors of the spread of Institutional demography (see Frank & McNicoll, knowledge and consequently behaviour change (Green- 1987; McNicoll, 1980, 1994) has sought to demonstrate halgh, 1995, p. 20). While the diffusion of new (western) that reproduction is shaped by relations of power and ideas about modern contraception to ‘traditional inequity at different levels of social organisation and, as cultures' is seen as a precursor to fertility decline, deeply such, is a highly political process. Institutional demo- rooted elements of traditional culture are understood as graphers have, for instance, shown how patrilineal inhibiting family planning programme effectiveness (see kinship systems and associated virilocal marriage for example Freedman, 1987). Diffusion theorists such patterns in Africa and India affect fertility through their as Cleland & Wilson (1987) and Lesthaeghe (1989) have construction of gender roles, by isolating women sought to identify elements of traditional cultural and economically, concentrating child-rearing costs almost belief systems which support high fertility, along with exclusively on them, and denying them any authority in the socio-cultural variables which act as ‘barriers' to decisions about family size and the practice of family contraceptive uptake, in order to facilitate the design of planning (see Thomas & Price, 1999). While institutional ‘culturally' appropriate family planning service delivery systems (Kertzer, 1995).
within social institutions and political structures, poli- Social demography has looked beyond family plan- tical-economic demographers (such as Schneider & ning programme failures to explain the persistence of Schneider, 1995; Fricke, 1995, on Sicily and Nepal high fertility. Drawing on micro-ethnographic research respectively) demonstrate the linkages between broad methods, Caldwell & Caldwell (1987) and Caldwell, macro-economic processes, social inequalities and local- Orubuloye, and Caldwell (1992), for example, have level fertility and sexual behaviour.
demonstrated the relationship between local kinship and Anthropological analyses have demonstrated that far belief systems and the demand for children in sub- from the reproductive couple performing a utilitarian Saharan Africa. In particular, these demographers have decision-making function, reproduction is negotiated focused attention on the role of intra-familial relations within gender-based power relations and within local and descent systems, which continue to imbue fertility knowledge and health systems (Angin & Shorter, 1998; with a high symbolic and moral value, in maintaining Dixon-Mueller, 1993; Oppong, 1995; Renne, 1993; high fertility. Despite the emphasis on locating fertility Rylko-Bauer, 1996). A study in the Gambia shows behaviour within its cultural context, however, many how fertility related behaviour and decision-making are social demographers (see for instance Caldwell & based upon indigenous understandings of bodily pro- Caldwell, 1992) continue to view the process of fertility cesses rather than calculations regarding the demand for decline in ‘traditional' societies as dependent upon the children (Bledsoe, Banja, & Hill, 1998). Use of contra- diffusion of western ideas, largely brought about ceptive and other sexual and reproductive health services N.Price, K.Hawkins / Social Science & Medicine 55 (2002) 1325–1336 has also been shown to be mediated by power accurately represent an empirical pattern at one point in differentials between health providers and the commu- time, whereas causality depends upon an analysis or nity they serve, based on gender, ethnicity and class understanding that goes beyond the collection of surface relations (Schuler, Choque, & Rance, 1994; Rutenberg facts and figures (Calhoun, 1995).
& Cotts Watkins, 1997). Thus, even in societies where Over the past two decades there has been a notable indigenous fertility regimes support fertility regulation, increase in the use of qualitative methods for applied utilisation of modern family planning services often policy research (Ritchie & Spencer, 1994). These remains low, with indigenous health practitioners methods have increasingly been employed in family continuing to be used in preference to bio-medical planning and sexual and reproductive health research.
services as a consequence of the significance of the social The value of qualitative methods is seen to lie in their relations implicit in the client–provider relationship potential to explore and probe more deeply into people's (Hawkins & Price, 2001).
accounts of social life than survey methods allow.
We now turn to a discussion of the methodological Nonetheless, a good deal of qualitative research remains implications of researching sexual and reproductive informed by positivism, aimed at collecting objective behaviour within specific social, cultural, and political- social facts and eliciting accounts of shared norms and economic contexts.
values. The use of the focus group is a case in point (seeCunningham-Burley, Kerr, & Pavis, 1999). In recentyears, the focus group has become the qualitative Researching sexual and reproductive behaviour: method par excellence of the sexual and reproductive methodological issues health field. Its utility is seen to lie in its ability toexplore attitudes, opinions and values, as a basis for Economic-utilitarianism and cultural diffusionism are understanding the social norms which guide the informed by functionalist theories of social organisation, behaviour of programme users and non-users (Bender in which behaviour (or social action) is understood as & Ewbank, 1994, p. 63). While the focus group is a being driven by an individualistic means-end rationality.
useful tool for eliciting discourses on dominant social Social order is considered to be achieved through the values, one of its main limitations lies in its tendency to integration of individual goal-oriented actions into a produce normative responses (Parker, Herdt, & Carbal- shared system of cultural and moral values, reinforced lo, 1991). Focus group participants are often reluctant to and maintained through sets of socially sanctioned discuss openly experiences that deviate from the norms and rules. Until recent years, mainstream accepted norms and values, and hence the method demographic and family planning research has relied provides little information about how, when and why heavily upon a positivist and empiricist research people use norms to legitimise behaviour or use methodology, the goal being to produce ‘empirical' strategies which contradict dominant norms (Parker and concrete knowledge concerning the variables that et al., 1991).
govern sexual and fertility related behaviour, utilising The faith placed by positivism on the empirical ‘verifiable' and ‘objective' methodological approaches.
efficacy of stated values and norms raises a fundamental The large-scale sample survey has remained the dilemma in social research, which can be summed up as dominant method of demographic research and has the ‘‘difficult relationship between the ‘is' and the ‘ought' provided much of the empirical basis for family planning in social action'' (Cohen, 2000, p. 82). While positivist and other sexual and reproductive health programme approaches may successfully elicit socially and morally design, monitoring and evaluation, most notably prescribed principles for behaviour, they are limited in through the use of knowledge, attitude and practice their ability to lay bare the concrete reality of everyday (KAP) surveys. The high value placed on the survey actions that often deviate from these norms and values.
method has centred on its apparent ‘scientific' legitimacy The importance of making the conceptual distinction and the objectively verifiable and therefore generalisable between stated social norms and everyday behaviour has data produced. However, the limitations of survey data been raised in a number of recent anthropological in producing the knowledge necessary to understand the critiques of functionalist social demography (Carter, complexity of social institutions and relations in which 1995; Greenhalgh, 1995; Hammel, 1990; Lockwood, sexual and reproductive behaviour are contextualised 1995). These critiques suggest that a conceptual and has been highlighted by a number of methodologists methodological distinction needs to be made between (Baum, 1995; Greenhalgh, 1990; Hammel, 1990; Hauser, what people say they will do or should do, and what 1993; Lockwood, 1995; Warwick, 1982). While a well- they actually do. Because of the implicit assumptions constructed statistical sample can provide important made by functionalism regarding the normative nature data on trends in social behaviour, it cannot necessarily of social behaviour, this distinction is largely overlooked inform about the contexts in which different behaviours in the main body of demographic research (Lockwood, occur or indeed about causality. Statistical data can only 1995). Furthermore, the views of different sets of actors N.Price, K.Hawkins / Social Science & Medicine 55 (2002) 1325–1336 (including the most marginalised) are often missing from which is presented and recorded in PRAs. At the very demographic accounts of sexual and reproductive least, where critical debate in public is not an established convention, we should avoid unwar- These shortcomings allude to the need for a metho- ranted assumptions about the accountability of dological approach to researching sexual and reproduc- publicly processed informationy PRAs tend to tive behaviour that enables a more rigorous engagement emphasize the general over the particularytend with the realities of the everyday lives of poor and towards the normative (‘what ought to be' rather marginalised people:far from being a static set of norms than ‘what is'), and towards a unitary view of and expectations, culture is continually being con- interests which underplays differencesy The ten- structed and negotiated in social interactions and dency to give normative information may be everyday practice. A starting point for such a methodo- encouraged by faulty interviewing techniquesy but logical approach is to develop an ‘actor-centred' view, often the very structure of the PRA sessionsFgroup which investigates how actors strategise and deploy activities leading to plenary presentationsFassumes different social norms in different concrete situations and encourages the expression of consensus. (Mosse, (Lockwood, 1995). Investigating actors' views of their social world has largely been the domain of theanthropologist, using ethnographic fieldwork and parti- Methodologies such as case study and social network cipant observation over substantial periods of time to analysis are increasingly being advocated by anthropol- reach below the surface of reported social norms, to ogists for researching sexual and reproductive behaviour observe concrete actions, and to collect the discourses (Hammel, 1990; Lockwood, 1995; Smith, 1993). Wolf and narratives of people as they go about their daily (1998) reports on the evaluation of youth health programmes through exploring young people's social The recognition by development agencies and practi- networks. Eyre (1997) suggests using ‘vernacular term tioners of the limitations of positivist research methods interviews' to elicit social knowledge related to sex for generating valid and appropriate information on among young people, on the basis that conventional social behaviour and of the time constraints in conduct- research on adolescent sexual behaviour has provided ing in-depth ethnographic research has led to an little understanding about the context in which sex takes increased interest in employing rapid assessment meth- place, because ‘ymost adolescent sex research has been ods (Manderson & Aaby, 1992). Participatory rural (or conducted using only surveys and focus groups rather rapid) appraisal (PRA) and participatory learning for than interviews, which allow the greatest probing of action (PLA) methods2 have been used effectively by individual knowledge' (Eyre, 1997, p. 9). Recent development agencies to conduct community based research on the cultural construction of young gay analyses from an actor-centred perspective. Nonetheless, men's HIV vulnerability has utilised peer researchers to experience has also shown that far from being rapid, the carry out interview and group discussions (Trussler, effective use of PRA/PLA tools often requires an initial Perchal, Barker, & Showler, 1999).
extensive input of time and resources for building Building upon some of the above methodological relations of trust with the community and to generate approaches, we have recently developed and field-tested an understanding of local social contexts and social a method that we refer to as peer ethnography.
networks. Without this initial understanding of com-munity dynamics, PRA/PLA tools may produce nor-mative statements which unknowingly favour the The peer ethnographic method discourses of the powerful and elite, while the perspec-tive of socially marginalised groups may remain unheard The peer ethnographic method is derived from the (Mosse, 1994). One of the major shortcomings of PRA is anthropological approach of ethnographic fieldwork, that it often represents an activity far removed from which holds that building a relationship of trust and informal, everyday life, hence promoting the production rapport with the community is a prerequisite for of a consensus view and normative discourse from researching social life. The ethnographic approach used by anthropologists is based on the premise that whatpeople say about social life and behaviour changes It seems highly probable that this social formality according to the level of familiarity and trust established imposes a selectivity on the kind of information between the researcher and researched. In recognition ofthe extensive timeframe required for ethnographic research, the peer ethnographic method has been PRA and PLA are two of a number of participatory designed to be carried out by what we are calling ‘peer approaches that developed out of the earlier rapid ruralappraisal, notably rapid assessment procedures, beneficiary researchers', who are already recognised members of the assessment, and community diagnosis.
community. We are not using the term ‘peer' to refer N.Price, K.Hawkins / Social Science & Medicine 55 (2002) 1325–1336 exclusively to young people, whose peers are often of the We included in the tools a set of conversational same age, but to refer to membership of any significant prompts for each of the interviews, to assist the peer social network, such as friends, neighbours, work-mates, researchers to initiate conversations and to follow up on and kinsfolk.
key issues. Some of the prompts covered the same issue The peer researchers conduct in-depth and unstruc- through different ways of asking, in order to allow tured interviews with individuals selected by them from probing and changes in conversational context. The their own social networks. The peer researchers in effect prompts were presented to the peer researchers as a become key informants by virtue of their recognised framework for their conversations, rather than as an status as community members and their local knowledge interview script. Interviewees were often also asked by (see Manderson & Aaby, 1992, p. 842). Rather than a the peer researchers to tell a story about someone they large sample of people being interviewed once only, a know (without giving names) in a particular situation series of in-depth interviews are conducted with a small (for example, someone who had sought treatment for a sample of individuals, selected from the same social STI), and to describe what happened to them.
network, on the basis that data produced by intensive Prior to the fieldwork, we provided participatory exploration of a few cases produce a more thorough training to the peer researchers in the use of the tools.
understanding of social life than the superficial explora- During training the peer researchers suggested signifi- tion of many cases (Hammel, 1990, p. 471).
cant changes to the prompts, refining them so that they A further basic tenet of the approach is that the peer translated easily into appropriate local language and researchers have an established relationship of trust with context. Some of our initial prompts were found to have the people they are interviewing. As a result, the little meaning in the context of young people's lives in fieldwork does not require the same amount of time urban Lusaka (where the method was field-tested), while for rapport building as conventional anthropological other important issues and topics had not been covered ethnography or some PRA exercises. In designing the in our initial framework. The peer researchers field- approach we have also taken account of the fact that tested different ways to raise the same issue with social networks are not made up of consensus groups, different interviewees, as the way something is talked but include relationships of conflict and mistrust. It is about varies even within one social context, according to anticipated that by using an approach which does not the age, gender, and ethnicity of the interviewee.
ask the interviewee to talk directly about themselves Following the training the peer researchers piloted the (and therefore make themselves vulnerable to their tools in the community for a week, during which time we peers), but about ‘other people like themselves', differing provided them with support and supervision. Following and conflicting perspectives will emerge in the narra- the field-testing, the prompts were refined further, so tives. The aim of the interviews is not to collect that the tools that the peer researchers finally took to the demographic or social ‘facts' through individual ac- field were locally specific versions of the generic tools counts of personal experience, but to elicit the meanings that we had initially designed.
that actors attribute to the social behaviour of their We did not intend that the peer researchers should peers. An important aspect of the method, therefore, is record a detailed script of each conversation or produce that all interviews are conducted in the third person, in vast quantities of authentic conversational narrative or an attempt to elicit narrative accounts of how inter- qualitative research data. The data collection prompts viewees conceptualise the social behaviour of ‘others' in were designed to assist the researcher to record phrases their networks, not accounts of their own behaviour or and/or events given most importance by the interviewee normative statements about how they ‘ought' to behave.
during the course of the conversation. The interviewee In the design of the data collection tools we have also participates in the data recording process through been aware that it is not possible to observe the confirming that the phrases and events recorded by the behaviour or to record the narratives of others without peer researcher were the most important ones in their filtering the data through an analytical framework, narratives and explanations. The peer researchers involving some level of meta-analysis. The tools that we recorded the key words, phrases and data from each designed and subsequently field-tested in Zambia (see interview on a sheet, and during data analysis these below) were therefore structured around several key sheets were used to show how different themes had analytical issues, which we considered to be central to emerged in different conversational contexts.
the social analysis of sexual and reproductive behaviour.
The framework consisted of a set of five conversationalinterviews aimed at eliciting perceptions of social Field-testing the peer ethnographic method identity and social networks, health and illness, sexualknowledge and sexual behaviour, reproductive beha- The peer ethnographic tools were field-tested in viour and fertility decision-making, and access to and collaboration with CARE International in Zambia quality of health services.
through their Partnership for Adolescent Sexual and N.Price, K.Hawkins / Social Science & Medicine 55 (2002) 1325–1336 Reproductive Health (PALS) project in Lusaka, which ethnic groups, as well as from a range of social seeks to improve the sexual and reproductive health of young people. The initial design of PALS was informed The interview narratives provided rich data from by a PLA appraisal process, involving 10,000 young which to build up an understanding of young people's people (see Shah, Zambezi, & Simasiku, 1999). Follow- perceptions and experiences of sexual and reproductive ing the PLA exercise, the PALS project was designed health. The section which follows provides a summary with the following key components:the establishment of of selected key issues emerging from the interviews.
youth-friendly counselling corners in government clinics;a cadre of trained youth educators to lead discussionswith young people in the community and provide Key issues emerging from the interviews counselling in the youth-friendly corners; and theestablishment of community agents to promote and sell Categorisation of social groups contraceptives to young people.
The peer ethnographic method was field-tested over The interviews showed a surprising consistency in the four months, between October 1999 and January 2000.
terms and categories used by young people to describe The field-testing was undertaken by eight of the PALS the different groups of people that live in their youth educators:four female and four male, with three compounds. For young people, ethnicity was not an still in school/college, two in part-time employment, and important category in describing their social world.
three being unemployed. Each of the peer researchers Rather, categorisations were based on perceived social interviewed between four and six young people, with a behaviours, placed within a predominantly moral total of 30 young people interviewed. The PALS youth framework. The most common categories used by educators underwent the initial training as peer interviewees to describe the people in their compounds researchers in early October 1999. We conducted the were churchgoers, school-goers, self-employed, business- training, in conjunction with CARE programme staff, men, poor/unemployed, drunkards, thieves, prostitutes who then supervised the peer researchers over the and gangsters (yos). Several interviews also referred to subsequent three-month data-collection period. Follow- wizards (particularly in Chilenje and Mandevu com- ing the three months of data collection, a data review and analysis process was facilitated over a three-week Within young people's narratives, by far the most period (described in more detail below), during which we important aspect of social organisation and identity is collected data from the peer researchers' interviews and church membership, and each compound has a multi- from their experience of using the tools, by individually plicity of churches including Catholic, Evangelical, interviewing each of the peer researchers.
Jehovah's Witness, Seventh Day Adventists, and Spir- The field-testing was carried out in three compounds itualist. Church membership provides structure to young in Lusaka:Chilenje, Mandevu and Chipata. Each people's daily lives, both in terms of a moral framework compound is served by a government clinic, in which a of behaviour to which they aspire and access to a social youth-friendly corner has been established. While the support network. Many out-of-school and unemployed compounds are relatively close to each other (Mandevu youth talk about the church as a major social resource in and Chipata border each other), they are also quite their community and a focus for their daily activities.
distinct in character and layout. Chilenje, the most Although the church provides young people with a set of affluent, was constructed by the government to provide ideals of moral behaviour, in young people's narratives housing to state employees, but during the 1990s being a church-goer does not preclude being a member housing passed from government to private ownership.
of groups whose behaviours are described as immoral, Chilenje has a higher proportion of residents in employ- bad or unchristian, such as drunkards, thieves and ment than either of the other two compounds and a relatively well-developed infrastructure.
Mandevu and Chipata compounds differ significantly Local belief systems from Chilenje, with much higher levels of unemploy-ment and poverty. Houses in Chipata and Mandevu are Wizards, in young people's narratives, are those mainly one-room buildings constructed by residents members of the community who practise witchcraft. It from mud bricks with corrugated iron (or cardboard) is beyond the scope of this case study to go into detail roofs, in contrast to the largely brick built two-three about indigenous witchcraft beliefs and practices.
room houses of Chilenje. While Chilenje is built around However, wizards are pervasive throughout the com- an infrastructure of roads, Mandevu and Chipata are pounds, and are active mainly at night. Wizards have an accessible only by dirt tracks. All three compounds are ambivalent place in young people's belief systems, ethnically mixed:residents include Bemba, Shona and illustrative of their ambivalent position vis-"a-vis their Nyakusa. The peer researchers were also from different traditional culture. Most young people describe wizards N.Price, K.Hawkins / Social Science & Medicine 55 (2002) 1325–1336 as being older people, who by practising witchcraft show to be a more pervasive activity throughout bars and themselves as unchristian:the churches teach that nightclubs. While young women's narratives tend to believing in witchcraft is the equivalent to worshipping associate commercial sex with vulnerability and poverty, the devil. Although young people reject witchcraft as a young men's discourses relate such activities to ‘bad false and traditional (and hence unchristian) belief behaviour', with moral judgements and condemnations system, most young people confirmed that there are of commercial sex often referring to young men's many wizards in the compounds, but being Christians vulnerability to HIV/AIDS: they are protected by the blood of Christ from the effectsof witchcraft. In other words, young people associate Prostitutes talk about how to make money. There witchcraft with the traditional belief systems of their was this girl who, before she was married, was a parents' and grandparents' generation from which they prostitute because she had nothing to do at home and are socially and culturally separated. These belief no money. She joined a small group of girls who used systems have important implications for young people's to go in the street and hang out around the cabs. She was lucky because she did not get HIV and found herhusband while she was a prostitute. She told him Social identity and livelihoods about being a prostitute and he offered to marryhery now she says she will never be a prostitute The categories of social behaviour used to describe the groups of people living in their compounds are alsoclosely linked to young people's perceptions of poverty Prostitutes are different to people like me. They don't and livelihood strategies. Those described as being have good moralsFeven though some of them do go businessmen, are mainly young men who own small to churchy Prostitutes cause worry because some- shops or bars in the compounds or who drive or work as times they go round knocking on single men's doors conductors on mini-buses, the informal transport looking for business. Some men are not strong services that link the compounds to the town. The enough to resist. (boy).
category of self-employed includes those who work asmarket traders (such as women who sell fruit and Health and health-seeking behaviour vegetables) and young men or ntemba boys (who sellsweets and biscuits from small kiosks at the side of the Poverty is described in many of the narratives as the road). While market traders of their parents' generation major determinant of health. The rich or well-off are are not necessarily perceived as being rich, ntemba boys identified as those who have three meals a day, can are perceived by their peers as being well-off having their afford the costs of medicines, are in good health, and own independent source of income. Gangsters (yos), look fat and confident. Those who are poor eat one meal identified by their style of dress (T-shirts and wide a day (and sometimes none), cannot afford to buy trousers) and their following of black American cult medicines, and consequently suffer from poor health.
music figures, are also categorised by other young The main illnesses about which young people are people as self-employed or businessmen.
concerned are sexually transmitted infections (STIs) Bars and nightclubs emerge as an important part of and HIV/AIDS, malaria, cholera, diarrhoea, and social life in the compounds, particularly for those young people who are out-of-school. The category The narratives also revealed the very ambivalent ‘drunkards' refers to those young-men who spend their nature of young people's health-seeking behaviour. Few days around the many small bars in the compounds.
young people trust nyangas (traditional healers), who These young men usually earn money from piecework are referred to variously as ‘witches', ‘cheats', and (casual labour). One interviewee described drunkards as ‘tricksters', and seen as ‘unreliable and expensive'.
those who ‘‘after earning a small amount from piece- Narratives describe nyangas as causing distrust in the work, spend it in the bars and on chibuku (locally community because of their association with witchcraft.
brewed beer) and dagga (marijuana)''. While bars are a Young people consider bio-medical health professionals focal point for social networking they are also important (doctors and nurses) as having the appropriate knowl- contact points for piecework which is usually contracted edge and skills to treat their illnesses. Nonetheless, there through these informal social networks.
is an apparent contradiction between discourses on ‘Prostitute' was a significant category referred to in health knowledge and belief systems and narratives the interview narratives. These narratives indicated that describing health-seeking behaviour. Peer researchers in the nature of commercial sex activity varies between the their discussions observed that while young people say different compounds. In Chilenje commercial sex they do not trust nyangas, many stories indicate that a appears to be more organised (in a ‘red-light area'), significant proportion of young people still visit nyangas, whereas in Mandevu and Chipata commercial sex seems in particular for treatment of STIs. This is largely N.Price, K.Hawkins / Social Science & Medicine 55 (2002) 1325–1336 attributed to lack of accessibility, availability and and those who have the ‘four C's' (a car, a crib/house, affordability of such treatment from government clinics cash and a cell-phone). However, men who have access and private doctors.
to income also figure largely in the narratives as thosemost likely to have multiple sexual partners and to Sexual relationships engage in high risk sexual activity, especially aroundbars and night-clubs. Young men talk about sexual One of the key findings of the interviews is that young relationships as central to their esteem and social status, people have a high level of awareness and accurate but also refer to sexual relationships as difficult when knowledge on sexual health, and many have been they do not have money to buy gifts for girlfriends.
successfully reached with information on STIs and Hence, for young men securing an income appears to be HIV/AIDS. Most young people were able to identify as much about being able to secure a girl-friend as it is accurately transmission routes, how to protect against about access to a livelihood. For young women, a sexual infection, and where to go for treatment for a sexually relationship is as much about being able to secure a transmitted infection. Nonetheless, such infections are a livelihood as it is about achieving sexual and emotional major concern of young people and figure prominently security. Hence, sexuality and livelihood strategies are in their narratives as worries, problems and illnesses closely intertwined in young people's construction and experienced among their social networks.
negotiation of sexual relationships and the social Narratives on sexual relationships also indicate that contexts in which they engage in high-risk sexual sexuality presents itself as a major area of conflict in young people's lives. The majority of young people citeabstinence as the preferred option for protection against Reproduction and fertility HIV. Abstinence is the message received by youngpeople from the church as the ‘moral standard' to which While knowledge of HIV and other STIs appears to they should adhere. Sticking to one partner and using be high, the interviews indicate that many young people condoms are mentioned in the interviews largely as lack knowledge of family planning and reproductive second options, if a person is unable to abstain.
health. Unwanted pregnancy emerged as a major However, narrative accounts of sexual relationships concern for young women and the majority of pregnan- present a very different reality:sexual relationships are cies described by young people in their social network an integral part of most young people's lives and an were unplanned.
important source of social and emotional support and While abortion is legal in Zambia (requiring the self-esteem. Sexual relationships are described as being signature of three doctors) only one of the stories important because they are ‘fun', ‘enjoyable' and ‘help referred to a young woman accessing a safe abortion you experience new things' and because it is ‘not normal through the University Teaching Hospital in Lusaka.
to not have a sexual relationship'.
All the other narratives concerning abortion described As was highlighted in the initial PLA exercises unsafe abortions, procured through insertion of instru- undertaken as part of the PALS project design process, ments into the vagina or consuming large quantities of economic exchange forms an integral part of young chloroquine. Several narratives refer to unsafe abortion people's sexual relationships, and the giving of gifts by leading to death and to young women disposing of new- boyfriends is talked about by young women as an born babies in pit latrines.
important part of a caring sexual relationship. Whileyoung men referred to women's physical appearance as Access to sexual and reproductive health services being one of the most important concerns in choosingwith whom to have a sexual relationship (eg ‘she is Cost of services emerged as a major barrier to young pretty', ‘she wears nice clothes', etc.), young women people's access to health care. As with the interviews on refer to men's economic status and sexual behaviour. A health-seeking behaviour, interviews concerning access boy who expects a girl to have sex with him and who and quality of services indicated that young people does not make her feel good by giving her nice gifts such consider clinic personnel as having the most appropriate as biscuits, body lotion and make-up, is considered not skills and knowledge to meet their needs. The majority only undesirable but also immoral. For many young of young people indicated that personnel at government women, boyfriends are their only means of accessing clinics would be their first choice of service provider, desirable consumer goods, as well as money to pay for while those with access to income tend to seek services essentials such as school fees.
from private doctors.
Economic criteria figure largely in young women's Nonetheless, interviews identified some significant descriptions of their ideal boyfriends. In particular barriers to young people's access to treatment and young women describe ideal partners as being ntemba services. Most young people know that STI services are boys (kiosk owners), bus drivers and bus conductors, provided free of charge at government clinics and select N.Price, K.Hawkins / Social Science & Medicine 55 (2002) 1325–1336 clinics as their first option for treatment. However, while has, therefore, necessarily been filtered through an young people consider that clinic staff are the most analytical framework, the theoretical background to appropriate providers to diagnose infections effectively which has been outlined at the start of this paper.
and accurately, government clinics are often short of At a second level of analysis, we are also aware that drugs, and young people are unable to afford prescrip- the peer researchers filtered their understanding of the tions. As a consequence many young people turn to self- data they collected through their own world view. The medication, using drugs sold by peddlers (which are initial participatory training session served to highlight usually expired), or turn to traditional healers.
the normative discourses of peer researchers regarding Condoms are provided free as part of the government young people's sexual and reproductive behaviour. One family planning programme, available at clinics. Inter- of our concerns in field-testing the method was that by views indicate that barriers to young people accessing interviewing people from their own social networks the family planning services include lack of knowledge of peer researchers may only access perspectives that their availability, negative attitudes of nurses, and fear reflected their own normative frameworks. In order to of being seen at clinics. Public sector family planning partially safeguard against this, we asked the peer services are provided within the maternal and child researchers to conduct the interviews with young people health clinic, and young men in particular describe in their networks who were not youth educators in the feeling self-conscious about visiting such female envir- CARE PALS project. During the data analysis process, onments to ask for condoms.
the contradictions between the peer researchers' norma- Socially marketed condoms are available at kiosks, tive frameworks and the narrative accounts of behaviour bars and pharmacies in the compound (three condoms collected in the interviews became evident.
retail for the price of a glass of chibuku). Lack of access The peer researchers also conducted their own data to economic resources was identified as a major concern analysis process, in which they worked as a group to for all young people interviewed. The lack of availability identify what they considered to be the key issues of free condoms at community-based outlets located emerging from the interviews and of most importance to close to centres of high-risk sexual activity (such as bars feedback to the CARE programme. One of the key and night-clubs) emerged in discussions with peer observations of the peer researchers was that while the researchers as a major barrier to access. As one peer young people using the youth-friendly corners saw them researcher put it:‘‘If a young man is at a bar, and he has as very valuable, certain groups of young people been drinking beer, and he meets a girl, if he has a choice remained almost entirely unserved by these services. In between spending his money on condoms or buying particular, the most marginalised groups of young another chibuku, he will buy another chibuku''.
people, such as out-of-school youth, those engaged incommercial sex, young men who spend their time inbars, and mini-bus drivers and conductors, were not The data analysis being reached. The peer researchers' analysis highlightedthe need for the PALS project to develop a more in- The data presentation above amounts to what could depth understanding of the social networks and be considered as our ‘meta-analysis', in which we have behaviour of marginalised youth. Furthermore, the peer drawn on the peer researchers' narratives as our primary researchers suggested that peer ethnography could ethnographic data. In undertaking this analysis the peer provide useful means for reaching these more margin- researchers became our key informants. We comple- alised groups and increasing their participation and mented our in-depth interviewing of the peer researchers dialogue with health care providers. Several peer (as our key informants) with participant observation researchers observed that young people they had talked that we carried out in the compounds during the to who had never used the clinics had expressed an training, supervision and data-collection process. Dur- interest in carrying out interviews with others in their ing our visits to the compounds we were also able to social networks, in order that their views may also be conduct informal interviews with community members, including friends and kin of our key informants, as well Peer ethnographers also utilised the data to identify as with other significant community members (such as key areas in which the PALS project could be improved nyangas, church leaders, bar owners, and health care or strengthened from young people's perspective. These providers). Hence, through the peer researchers we were included:utilising peer networks to reach young people able to gain rapid access to the community, including to with information and condoms in places where they community members who were not an immediate part of meet such as bars, night clubs, and on mini-buses the key informants' social network. However, we (especially for out-of-school youth); using peer networks recognise that it is not possible to interpret what people to make free condoms more easily available; and say and do in a way that is uninfluenced by theory increased advocacy on behalf of young people, for (Lockwood, 1995, p. 14). Our interpretation of the data example, to increase their access to effective drugs for N.Price, K.Hawkins / Social Science & Medicine 55 (2002) 1325–1336 treating STIs, and to information (for example on as a purely ‘oral' method with non-literate groups, such treatment of complications of abortion). Through an application of the method remains to be developed the 30 interviews conducted during the field-testing, and field-tested.
the method was able to confirm many of the findings of We are aware that the approach is vulnerable to the the PLA research which was carried out with 10,000 criticism that the sample of respondents may be biased young people during the initial PALS project design and unrepresentative. We recognise the importance of ensuring a locally representative sample, while alsorecognising that it is extremely difficult to use randomsampling techniques to find representative informants.
Conclusion: reflections on the method Local experience is needed to know what representa-tiveness means, using researchers who are able to locate The peer ethnographic method was designed to a point of entry into the local social and cultural system.
address some of the limitations of other applied research The peer ethnographic approach endorses this principle, methods. However, it does have limitations of its own, using peer researchers as key informants, who are as well as considerable strengths. One of the most strategically placed by virtue of their membership and important outcomes of the data collection and analysis understanding of the communities in which the research process was the ensuing dialogue that it generated, both is undertaken. The field-testing confirmed that collec- between the peer researchers and the CARE programme tively the peer researchers had talked to young people staff and between the peer researchers and other young from different sections of the local community, with people in their social networks. In this respect, we whom they made contact through their own social consider that the peer ethnographic method has networks and through ‘friends of friends'. Diversity of potential to facilitate greater participatory dialogue interviewees was further facilitated by the fact that the between programme implementers and the groups in peer researchers were a mix of male and female, and the community that they are aiming to reach. As we came from a range of different ethnic groups and socio- field-tested the method, its potential as a dynamic and flexible approach, which can be continually adapted and A more substantial potential criticism surrounds the re-designed as part of the research process, also became question of validating the data, by ensuring that the peer evident. Indeed, following its field-testing by the peer researchers and interviewees are indeed telling us the researchers, the final version of the tools was quite ‘truth' about their social lives and behaviour. It is, different to the one we had initially designed, and which perhaps, at this point that the method makes its greatest had been based upon our initial (possibly ethnocentric) departure from positivism. Indeed, in our view the key assumptions. In this sense our ownership and control strength of the method lies in its very lack of claims to over the method and the tools diminished in the course any positivist criteria for the collection and presentation of it being used in the field.
of ‘objective' or ‘standardised' data. The main focus of A key issue that we have sought to address in the peer the method is on analysing contradiction and difference ethnographic method is in the domain of language and in the discourses of different people within a social interpretation. Nuance is of the essence in understanding network, rather than on gathering ‘social facts'. A and collecting local narratives. We therefore considered further aim is to look at different ways in which the same that it was of particular importance that the interviews issue is talked about at different stages in the relation- were conducted in local languages, particularly as the ship between the researcher and the actor and how this research focused on such intimate issues as sexuality, varies according to conversational context.
health, illness and reproduction. In our view this is one One of the key social dynamics that the method draws of the key strengths of the peer ethnographic method:it upon is that of gossip (see Hammel, 1990). The facilitates interviews to be carried out using local contradiction between consensus statements made by colloquialisms and nuances. The major linguistic and groups and what individuals gossip about provides interpretative problems arise at the level of ‘meta- important insights into social context. Gossip is the local analysis'. In our case we had to interview the peer currency of social networks, providing invaluable insight researchers (our key informants) in English, as we did into relationships of power and vested interests that not have the time in the field that conventional operate under the surface. Instead of discarding gossip anthropological ethnography would allow for learning as invalid because it derives from biased accounts, the the local language. A further limitation of the method is peer ethnographic method embraces and analyses gossip that the peer researchers must be sufficiently literate to as an essential component of ethnographic data. In our be able to record the conversational interview data. This use of the peer ethnographic approach we are not requirement may, therefore, preclude the most margin- seeking social ‘truths', as positivist methods may alised groups from becoming peer researchers. While we understand them, but a rich and dynamic social consider that the approach has potential to be developed commentary, in which subjective identity is created N.Price, K.Hawkins / Social Science & Medicine 55 (2002) 1325–1336 and in which sexual and reproductive behaviour is given Dixon-Mueller, R. (1993). The sexuality connection in repro- ductive health. Studies in Family Planning, 24(5), 269–282.
Easterlin, R., & Crimmins, E. (1985). The fertility revolution: A supply demand analysis. Chicago:University of Chicago Eyre, S. (1997). The vernacular term interview:Eliciting social The authors acknowledge with gratitude funding for knowledge related to sex among adolescents. Journal of our current research which was provided by the Adolescence, 20, 9–27.
Department for International Development, UK. We Frank, O., & McNicoll, G. (1987). An interpretation of fertility are also grateful to CARE International in Zambia for and population policy in Kenya. Population and Develop- their cooperation in field-testing our peer ethnographic ment Review, 13(2), 209–244.
Freedman, R. (1987). The contribution of social science tools in the context of their PALS (Partnership for research to population policy and family planning pro- Adolescent Sexual and Reproductive Health) project in gramme effectiveness. Studies in Family Planning, 18(2), 57– Lusaka. However, the opinions expressed in this article are ours alone, and we accept full responsibility for any Fricke, T. (1995). History, marriage politics, and demographic errors or shortcomings.
events in the central Himalaya. In S. Greenhalgh (Ed.),Situating inquiry (pp. 202–224). Cambridge:Cambridge UniversityPress.
Fricke, T. (1997). Culture theory and demographic process: Angin, Z., & Shorter, F. (1998). Negotiating reproduction and Toward a thicker demography. In D. Kertzer, & T. Fricke gender during the fertility decline in Turkey. Social Science (Eds.), Anthropological demography: Toward a new synthesis and Medicine, 47(5), 555–564.
(pp. 248–277). Chicago:University of Chicago Press.
Baum, F. (1995). Researching public health:Behind the Gardner, K., & Lewis, D. (1996). Anthropology, development and the post-modern challenge. London:Pluto Press.
Science and Medicine, 40(4), 459–468.
Greenhalgh, S. (1990). Toward a political economy of fertility: Behrman, J., & Knowles, J. (1998). Population and reproduc- Anthropological perspectives. Population and Development tive health:An economic framework for policy evaluation.
Review, 16(1), 85–106.
Population and Development Review, 24(4), 697–737.
Greenhalgh, S. (1995). Anthropology theorises reproduction: Bender, D., & Ewbank, D. (1994). The focus group as a tool for Integrating practice, political economic, and feminist health research:Issues in design and analysis. Health perspectives. In S. Greenhalgh (Ed.), Situating fertility: Transition Review, 4(1), 63–80.
Anthropology and demographic inquiry (pp. 3–28). Cam- Bledsoe, C., Banja, F., & Hill, A. (1998). Reproductive mishaps bridge:Cambridge University Press.
and western contraception:An African challenge to fertility Greenhalgh, S. (1996). The social construction of population theory. Population and Development Review, 24(1), 15–57.
science:An intellectual, institutional, and political history of Caldwell, J., & Caldwell, P. (1987). The cultural context of high twentieth century demography. Comparative Studies in fertility in sub-Saharan Africa. Population and Development Society and History, 38(1), 26–66.
Review, 13(3), 409–437.
Hammel, E. A. (1990). A theory of culture for demography.
Caldwell, J., & Caldwell, P. (1992). Family systems:their Population and Development Review, 16(3), 455–485.
viability and vulnerability. In E. Berquo, & P. Xenos (Eds.), Hauser, P. (1993). The limitations of KAP surveys. In M.
Family Systems and Cultural Change (pp. 46–66). Oxford: Bulmer, & D. Warwick (Eds.), Social research in developing Clarendon Press.
countries (pp. 65–69). London:UCL Press.
Caldwell, J., Orubuloye, I., & Caldwell, P. (1992). Fertility Hawkins, K., & Price, N. L. (2001). From international policy decline in Africa. Population and Development Review, 18(2), to local reality. Analysing the social contexts of reproduc- tive health:a case study from Bolivia. In S. Tremayne (Ed.), Calhoun, C. (1995). Critical social theory. Oxford:Blackwell.
Managing reproductive life: Cross-cultural themes in sexu- Carter, A. (1995). Agency and fertility:for an ethnography of ality and fertility (pp. 52–70). Oxford:Berghahn Books.
practice. In S. Greenhalgh (Ed.), Situating Fertility: Kertzer, D. (1995). Political-economic and cultural explana- Anthropology and Demographic Inquiry (pp. 55–85). Cam- tions of demographic behavior. In S. Greenhalgh (Ed.), bridge:Cambridge University Press.
Situating fertility: Anthropology and demographic inquiry Cleland, J., & Wilson, C. (1987). Demand theories of the (pp. 29–52). Cambridge:Cambridge University Press.
fertility transition:An iconoclastic view. Population Studies, Kertzer, D., & Fricke, T. (1997). Toward an anthropological demography. In D. Kertzer, & T. Fricke (Eds.), Anthro- Cohen, I. (2000). Theories of action and praxis. In B. The pological demography: Toward a new synthesis (pp. 1–35).
Blackwell Companion to Social Theory Turner (Ed.),, (2nd Chicago:University of Chicago Press.
ed.) (pp. 73–112). Oxford:Blackwell.
Lesthaeghe, R. (1989). Production and reproduction in Sub- Cunningham-Burley, S., Kerr, A., & Pavis, S. (1999). Theoris- Saharan Africa:an overview of organising principles. In R.
ing subjects and subject matter in focus group research. In Lesthaeghe (Ed.), Reproduction and social organization in R. Barbour, & J. Kitzinger (Eds.), Developing focus group sub-Saharan Africa (pp. 13–59). Berkley:University of research: Politics, theory, practice (pp. 186–199). London: California Press.
Sage Publications.
N.Price, K.Hawkins / Social Science & Medicine 55 (2002) 1325–1336 Lockwood, M. (1995). Structure and behaviour in the social province, Kenya. Studies in Family Planning, 28(4), 290– demography of Africa. Population and Development Review, 21(1), 1–32.
Rylko-Bauer, B. (1996). Abortion from a cross-cultural Manderson, L., & Aaby, P. (1992). An epidemic in the field? perspective:An introduction. Social Science and Medicine, Rapid assessment procedures and health research. Social 42(4), 479–482.
Science and Medicine, 37(7), 839–850.
Schneider, P., & Schneider, J. (1995). High fertility and poverty McNicoll, G. (1980). Institutional determinants of fertility in Sicily:Beyond the culture vs rationality debate. In S.
change. Population and Development Review, 6(3), 441–462.
Greenhalgh (Ed.), Situating fertility: Anthropology and McNicoll, G. (1992). The agenda of population studies:A demographic inquiry (pp. 179–201). Cambridge:Cambridge commentary and complaint. Population and Development University Press.
Review, 18(3), 399–420.
Schuler, R., Choque, M., & Rance, S. (1994). Misinformation, McNicoll, G. (1994). Institutional analysis of fertility. In K.
mistrust, and mistreatment:Family planning among Boli- Lindahl-Kiessing, & H. Lundberg (Eds.), Population, vian market women. Studies in Family Planning, 25(4), economic development and the environment (pp. 199–230).
Oxford:Oxford University Press.
Shah, M., Zambezi, R., & Simasiku, M. (1999). Listening to Mosse, D. (1994). Authority, gender and knowledge:Theore- young voices: Facilitating participatory appraisals on repro- tical reflections on the practice of participatory rural ductive health with adolescents. Washington DC:Focus on appraisal. Development and Change, 25, 497–526.
Young Adults with CARE International.
Oppong, C. (1995). A high price to pay:For education, Simmons, O. (1988). Perspectives on development and population subsistence or a place in the job market. Health Transition growth in the third world. New York:Plenum Press.
Review, 5(Suppl.), 35–56.
Smith, H. (1993). On the limited utility of KAP-style survey Parker, R., Herdt, G., & Carballo, M. (1991). Sexual culture, data in the practical epidemiology of AIDS, with reference HIV transmission, and AIDS research. The Journal of Sex to the AIDS epidemic in Chile. Health Transition Review, Research, 28(1), 77–98.
3(1), 1–16.
Price, N. L. (1996). The changing value of children among the Thomas, N., & Price, N. L. (1999). The role of development in Kikuyu of Central Province Kenya. Africa, 66(3), 411–437.
global fertility decline. Futures, 31(8), 779–802.
Price, N. L. (1998). Institutional determinants of fertility in Trussler, T., Perchal, P., Barker, A., Showler, G. (1999).
Gwembe Tonga society. Social Sciences in Health, 4(1), 25–44.
Cultural construction of young gay men's HIV vulner- Price, N. L., & Thomas, N. (1999). Continuity and change in ability. Fourth International Conference on AIDS impact: the Gwembe Valley Tonga family and their relevance to Biopsychosocial aspects of HIV infection, Ottawa, Canada, demography's nucleation thesis. Africa, 69(4), 510–548.
Conference Paper Abstract.
Renne, E. (1993). Gender ideology and fertility strategies in an Warwick, D. (1982). Bitter pills: Population policies and their Ekiti Yoruba village. Studies in Family Planning, 24(6), implementation in eight developing countries. Cambridge: Cambridge University Press.
Ritchie, J., & Spencer, L. (1994). Qualitative data analysis for Wolf, R. C. (1998). Social networks in adolescent peer promotion applied policy research. In A. Bryman, & R. Burgess (Eds.), programs: Developing methods for monitoring and evaluation.
Analyzing qualitative data (pp. 173–194). London:Routle- Focus on Young Adults (with the Centre for Population and Development Activities and Johns Hopkins School of Rutenberg, N., & Cotts Watkins, S. (1997). The buzz outside Hygiene and Public Health), Washington DC. Unpublished the clinics:Conversations and contraception in Nyanza

Source: http://www.options.co.uk/sites/default/files/peer_article_ssm.pdf