Data collection in the field for a ethnographic research

1. ‘Facts’ and ‘reality’

Trained biologists looking at cells under a microscope can come up with accu­rate descriptions of the components of those cells. If they have looked at many cells over the course of time, they can determine which are the intrinsic features of a cell belonging to a certain plant or animal, and which are random deviations. Moreover, there is an assumption that any trained biologist would come to the same conclusions.

Ethnographers can rarely operate with such objective certitude. While we may strive for accuracy, we must always keep in mind that the ‘facts’ of human behav­iors, values, and interactions are sometimes in the eye of the beholder. They can be manipulated, deliberately or otherwise, by the people being studied. The ‘real­ity’ we perceive as ethnographers is thus always conditional; we cannot take it for granted that another ethnographer, looking at the same set of ‘facts’ at a dif­ferent time, will come to exactly the same conclusions.

Some scholars (such as the ‘postmodernists’ discussed earlier) would take the position that striving for ‘accurate’ depictions of social ‘reality’ through the col­lection of objective ‘facts’ is an inherently futile exercise. Statements of reality, they contend, must always be ‘deconstructed’ in order to discern who the observer was, and what his or her biases may have been that caused the conclu­sions to take the form they did. Still other scholars take the position that society is a kind of elaborate game in which observer and observed create ‘reality’ as they interact (much as participants in a game of football play with the objective rules of the game and so come up with a somewhat different game every time); as such their intention is not so much to characterize some sort of timeless ‘real­ity’ but to chronicle a particular snapshot of that reality. They may even be more interested in analyzing the process by which ‘players’ strategize their way through the ‘game’ than with the presumed outcome of the ‘game’.

My remarks in this section are not meant to take one position or the other on these theoretical issues. I will operate on the assumption that whatever interests an ethnographer may have in analyzing his or her data, there is still the necessity to collect data in a systematic form so as to best support his or her arguments.

2. A note on applied ethnography

When a researcher wants to use the results of his or her fieldwork to make rec­ommendations on public policy, or to contribute to the formation and mainte­nance of organizations or agencies that serve the community under study, then he
or she is said to be conducting applied ethnography (see Chambers, 2000, for a complete review of this field). Unlike academic researchers, who can consider the ‘postmodern’ possibilities of ambiguity and deception raised in the previous section, applied ethnographers must proceed from a position of relative certitude. Why, after all, would anyone pay attention to their recommendations for action unless they could back up their assertions with clearly delineated, more or less objective data? So the potential for participant observation research making a real contribution to the world at large is dependent on the ethnographer being able to convince the relevant audience that he or she really does know what is going on in the study community.

In Trinidad, my research on alcoholism in the Indian community led me to rec­ommend to government health planners that public monies would be better spent on public education campaigns designed to encourage people with a problem to seek out their nearest AA group. Spending limited public monies on expensive hospital-based treatment facilities would be a waste, since most Indians would not consider anything that went on in such a setting to be legitimately therapeu­tic. The AA group, based in the kin group and the local village, was for that com­munity a more appropriate setting for recovery.

In the deinstitutionalization study, I was able to use my data to convince pro­gram managers to include sexuality training as part of the habilitation plan. I advocated against making much of the mechanics of sex (basic anatomy, etc.) since such information would likely not be absorbed by the clients. I recom­mended instead that the training focus on relationships, and suggested that ‘classes’ be structured not as didactic lectures but as ‘role-playing’ sessions in which the clients could try out styles of behavior and comment on what they had seen and participated in.

3. Three key skill areas

Although there are, as we shall soon see, a great many specific data collection techniques available to ethnographic researchers, all of them fit into three large categories representing the key skill areas that must be part of the repertoire of all fieldworkers: observation, interviewing, and archival research.

4. Observation

Observation would seem to be the most objective of ethnographic skills, since it seems to require little or no interaction between the researcher and those he or she is studying. We must, however, remember that the objectivity of our five senses is not absolute. We all tend to perceive things through filters; sometimes these filters are an intrinsic part of the research method (e.g. our theories or ana­lytic frameworks), but sometimes they are simply artifacts of who we are: the pre­conceptions that come with our social and cultural backgrounds, our genders, our relative ages, and so forth. Good ethnographers strive to be conscious of – and therefore to set aside – these latter factors, which constitute a perspective we call ethnocentrism (the assumption – conscious or otherwise – that our own way of thinking about and doing things is somehow more natural and preferable to all others). But we can never banish them completely.

In the ideal, observation begins the moment the researcher enters the field set­ting, where he or she will strive to set aside all preconceptions and take nothing for granted. It is sometimes said that the ethnographer becomes like a little child, to whom everything in the world is new. As a result, the process of observation begins by taking everything in and recording it in as much detail as possible, with as little interpretation as possible. (For example, one might observe, ‘The people at the temple were chanting and swaying to the beat of a drum’ rather than ‘The people at the temple were carried away by religious ecstasy.’) Gradually, as the researcher gains more experience in the field site, he or she can begin to discern matters that seem to be important and to concentrate on them, while paying pro­portionately less attention to things that are of lesser significance. It is vital to the outcome of the research that the ethnographer also come to recognize patterns – behaviors or actions that seem to be repeated so that they can be said to be typical of the people being studied (as opposed to unique and perhaps random occurrences).

We may think we all have a natural facility for observing and describing the people and events that surround us. But in fact, what we usually have is a well- developed screening process. When we are functioning in our own everyday worlds, it would simply be inefficient if we paid complete and objective attention to everything, even things very familiar to us. In our own worlds, we learn to focus. That which we do not ‘see’ is almost always greater than that which we do. Notwithstanding the weight we grant to ‘eyewitness’ accounts, the fact is eye­witnesses can be quite unreliable because most of us have gotten used to tuning out most descriptors. So ethnographic observation cannot depend solely on our ‘natural’ facilities. We have to work hard to really and truly see all the many details of a new situation – or (as in the case of the deinstitutionalization study) to see familiar situations through the eyes of those who are in many ways ‘strangers’ to those situations.

Some observational techniques are said to be unobtrusive, which has tradition­ally meant that those under study do not know that they are being observed. Modern standards of ethical research, which include procedures for ‘informed consent’ (which will be discussed in a later chapter), have greatly restricted the scope of truly unobtrusive observation. It is still possible, however, to observe people in public places where you as a researcher can just blend in (e.g. making notes about how people seat themselves in an airport waiting room or a Department of Motor Vehicles office); it is not necessary to explain oneself or obtain permission from people so observed. The study of such spatial relation­ships is known as proxemics; the related study of people’s ‘body language’ is technically known as kinesics (see Bernard, 1988, pp. 290-316 for an extended discussion of unobtrusive techniques). Researchers must, however, be sensitive to matters of privacy even in ‘public’ spaces. It is not likely that anything very inti­mate will happen in an airport waiting area. But conducting proxemic observa­tions in a public restroom might certainly be questionable.

Careful, reasonably unobtrusive observations of proxemic and kinesic behavior can tell us a great deal about the unspoken assumptions of cultures. Among the Trinidad Indians, there is a restricted sense of private space as compared with North Americans. Houses of the more tradition-minded people often lack doors or other partitions demarcating sleeping areas from other living areas. On the other hand, the people are quite distant and reserved in terms of interpersonal space: there is very little hugging, hand-holding, or other forms of emotional expression, at least in public places. Keeping ‘good posture’ seems to be important and children are some­times explicitly reprimanded for ‘slouching around’. A rather formal conversational distance is maintained in most circumstances. Indians sometimes express disdain for non-Indian Trinidadians who, they say, ‘are all over you all the time’.

Adults with mental retardation often have not mastered the nuances of expected proxemic and kinesic behavior typical of the mainstream in the United States. Indeed, among the most important cues marking people as ‘retarded’ are those having to do with improper use of spatial and body language. People with mental retardation tend to be very vigorous touchers and huggers – they often seem to ‘invade the space’ of others. On the other hand, they seem to have a par­adoxically highly developed sense of their own personal space. If one of the men in the program had his own room – or even his side of a shared room – he would defend it passionately and sometimes fly into a rage if anyone came into it with­out being explicitly invited.

There are other kinds of unobtrusive research that are still ethically defensible. For example, behavior trace studies are very much like archeological excava­tions, but among the living. There has been much publicity about ‘garbology’ projects – research based on sifting through people’s trash in order to find clues about how they live. One might question how truly ‘unobtrusive’ such a project could be (I, for one, would definitely notice teams of researchers picking through my garbage and perhaps even give some second thoughts to what I am throwing away), but even if the subject knows that he or she is being studied and gives per­mission for researchers to proceed, there need be no further interaction between researchers and subjects.

Given the ethical concerns about absolutely ‘unobtrusive’ observation (as even the most innocuous project could be considered ‘deceptive’ under some circum­stances), ethnographers rely much more frequently on observations of settings in which they are known to the participants and in which they may well be engaged in the activities themselves (participant observation). But just because the behav­ior of people in a research setting unfolds in a seemingly haphazard fashion (or so it may seem to the ‘little child’ researcher at the beginning of a field study), this does not mean that the observational process itself should be haphazard. Good ethnographic observation necessarily involves some degree of structure. At minimum, researchers should cultivate the habit of taking well-organized field notes that include:

• a statement about the particular setting (e.g. school, home, church, store);
• an enumeration of the participants (number, general characteristics, e.g. ages, genders);
• descriptions of the participants (rendered in as nearly objective a form as pos­sible: ‘The man wore a torn, dirty pair of pants’, not ‘The man looked poor’);
• chronology of events;
• descriptions of the physical setting and all material objects involved (in great detail, taking nothing for granted);
• descriptions of behaviors and interactions (avoiding interpretations: ‘The man was weeping and repeatedly struck his head with his fist’, not ‘The man looked deranged’ – particularly if video recording equipment is not possible);
• records of conversations or other verbal interactions (as near to verbatim as possible, particularly if it is not feasible or desirable to have a tape recorder running).

Some projects involving multiple members in a team approach rely on finely tuned and standardized note-taking processes. But even if you are on your own, you should train yourself to be as meticulous as possible in recording data. The more nearly your records of observations at selected sites contain the same infor­mation, the more efficient it will be to retrieve and compare data.

My research on alcoholism as a factor in the lives of modern Trinidad Indians led me to numerous observations at meetings of Alcoholics Anonymous, which had been imported to the island from the United States in the 1960s. Keeping structured notes enabled me to readily answer such questions as: What is the average age of the ‘recovering’ Trinidad Indian alcoholic? (45-50 years). Is there a particular order of speakers? (yes, those only briefly sober speak first, building up to those with many years of documented sobriety whose ‘witness’ is therefore surrounded by greater solemnity). Are Indians the island’s only alcoholics? (no, but they are – with extremely rare exceptions – the only ones who attend AA meetings). What is the role of women? (they provide refreshments, but do not speak). I was not, strictly speaking, a ‘participant’ observer at AA meetings since I am not a recovering alcoholic. But I was brought to my first meetings by informants who did fit that category, and who introduced me to the membership at large. After a while my presence became an accepted thing.

I spent several years as a participant observer (as a volunteer tutor) in class­rooms where adults with mental retardation were being taught basic skills. Since I was ‘participating’, I had less opportunity to make detailed notes while on site, and therefore had to cultivate the skill of reconstructing my observations as soon as possible thereafter. The keeping of structured notes was especially helpful when I began observing other programs in my own area (e.g. those that served only people with mental retardation as opposed to the program that dealt with ‘dually diagnosed’ clients) and programs in other states where somewhat different laws and standards of care were in operation. The structured observa­tions in all these settings made it possible to compare and contrast behaviors and interactions that seemed to be dependent on factors beyond the clients’ control, such as the requirements of various bureaucratic systems (e.g. criminal justice, education).

A note on notes

The importance of keeping structured and organized field notes when conducting observational research cannot be overemphasized, whether one is doing the research solo or as part of a team. It is worth keeping in mind the following points about maintaining field notes:

• Make sure that every note ‘card’ (or whatever format you find most congen­ial for recording) is headed by the date, place, and time of observation.
• Be sure to record as many verbatim verbal exchanges as possible; nothing conveys the sense of ‘being there’ more than the actual words of the participants.
• Use pseudonyms or other codes to identify participants in order to preserve anonymity and confidentiality – you never know when unauthorized people might try to sneak a peek. One piece of advice from bitter personal experi­ence: don’t make your code system so complex and obscure that even you can’t reconstruct the cast of characters.
• Be sure to record events in sequence; some researchers find it helpful to divide their notepad (the same advice goes for those who take notes directly onto laptop computers) into hours or even minutes so that they can precisely place actions in order.
• Keep all descriptions of people and material objects on an objective level; try to avoid making inferences based solely on appearances (See Adler and Adler, 1994, and Angrosino and Mays de Perez, 2000, for more comprehen­sive reviews of the theory, methods, and ethical ramifications of observa­tional research.)

5. Interviewing

The hallmark of observational research, as noted several times in the previous section, is to record details in as nearly objective a descriptive manner as possible, avoiding interpretations and inferences, and setting aside one’s own preconceptions. The ethnographer ultimately comes to a point of recognizing or inferring meaningful patterns in observed behaviors. But the inevitable next question is: what, exactly, do those behaviors mean? At that point, it is necessary to start asking questions of knowledgeable people in the community or group under study. Interviewing thus grows logically out of observation.

We have noted that while observation seems to be nothing more than what we do in everyday life, it really requires a heightened degree of consciousness, awareness of fine-grained detail, and the careful recording of structured, organ­ized data in order for it to be useful as a research tool. In a similar fashion, we might be tempted to think that interviewing, which is a kind of conversation after all, is something we can all do. Moreover, we see ‘interviews’ all the time on TV – it all looks so effortless. Why, then, would anyone call the sort of in-depth, open-ended interviewing typical of ethnographic research ‘the most technically challenging and, at the same time, the most innovative and exciting form’ of data collection? (This is a position taken by Stephen Schensul, Jean Schensul, and Margaret LeCompte in the comprehensive, multi-volume, widely used Ethnographer’s Toolkit.) It is clear that there is more to ethnographic interview­ing than having an ordinary conversation such as you would have with a friend; it is also different in some way from the sort of TV interview in which both inter­viewer and celebrity subject are more or less following a predetermined script and have tailored their remarks to fit a limited time frame.

Ethnographic interviewing is indeed conversational in the sense that it takes place between people who have grown to be friends as the ethnographer has been a participant observer in the community in which his or her respondent lives. In that sense, it is different from the kind of interviewing that might be done by a news reporter prying information from a ‘source’. It is certainly not the same as a police officer grilling a suspect or a lawyer interrogating a witness or a health care professional taking a medical history from a patient. But on the other hand, it must necessarily go beyond the parameters of an ordinary friendly conversa­tion, since the researcher does need to find out certain things and must be vigi­lant in keeping the conversation on track – all without seeming to be coercive or impatient.

The ethnographic interview is therefore typically open-ended in nature – it 42 flows conversationally and accommodates digressions, which may well open up new avenues of inquiry that the researcher had not originally considered. In that sense it is a kind of partnership in which the informed insider helps the researcher develop the inquiry as it goes along.

The ethnographic interview is also conducted in depth. It is not merely an oral version of a shotgun survey questionnaire. Instead, it is intended to probe for meaning, to explore nuances, to capture the gray areas that might be missed in either/or questions that merely suggest the surface of an issue.

In order to make an interview work for maximal ethnographic results, the inter­viewer should prepare by reviewing everything he or she already knows about the topic at hand and coming up with some general questions that he or she wants to know more about. These questions should not harden into a checklist of survey items, but should serve as a guide for the main points of the conversation. Although the interview may be unstructured (in the sense of not being tethered to a formal set of survey questions), it is by no means haphazard. In addition to the open-ended questions with which the interviewer enters the encounter, there will be a variety of probe questions designed to keep the interview moving in pro­ductive directions. Some examples of useful probes include:

• neutral acknowledgements (‘Yes, I see … ’);
• repeating what the person has said as a question to make sure that you have understood correctly (‘So your family built the house on that side of the vil­lage in order to be nearer to the shrine?’);
• asking for more information (‘Why did your older brother think he needed to go to England for further study?’);
• asking for clarification of apparent contradictions (‘You told me you were born in 1925 but you described the arrival of the last indenture ship [which was in 1917] … ’);
• asking for an opinion (‘You described your teenage daughter going out on dates. What do you think of the way young people act nowadays?’);
• asking for clarification of a term (‘You talk about “liming” along the road. What, exactly, does that mean?’ [idling with a gang of friends, usually with alcohol involved]), or a complex process (‘Please take me one more time through the steps in refining sugar cane into molasses’);
• asking for lists of things in order to get a better sense of how insiders cate­gorize and organize the world around them (‘What kinds of beverages besides rum do they sell in a “rum shop”?’);
• requesting narratives of experience – concrete anecdotes that illustrate a gen­eral point (‘You speak of boys being “led astray” by drink. Can you tell me about a particular time when you felt you were “led astray”?’).

Complementing these positive steps that you can take to make an interview work, there are several things to avoid – things that might add up to interviewer bias. For example, do not:

• ask leading questions (‘Aren’t you ashamed of all the bad things you did when you were drinking heavily?’);
• ignore leads when the interviewee introduces new themes that seem impor­tant to him or her;
• redirect or interrupt a story;
• ignore the interviewee’s non-verbal cues (e.g. signs of boredom or anger);
• ask questions that seem to tell the interviewee the answer you want (‘Don’t you agree that AA has accomplished a great deal for the benefit of alcoholics in Trinidad?’);
• use non-verbal cues (e.g. vigorous nodding of the head, leaning over to shake the interviewee’s hand) to indicate when the interviewee has given you the ‘right’ answer.

In addition to these specific techniques designed to keep the interview flowing, there are several points that speak to the overall ‘etiquette’ of conducting an interview:

• Try to avoid interjecting yourself too much into the narrative. Some manuals advise against ever expressing your own opinions, but I would not go so far – you are, after all, a real person with your own perspective, and you will not likely impress the person you are talking to if you act like a blank wall. But neither should you use the interview as a forum for expounding your own ideas or criticizing or belittling the ideas of the person you are interviewing.
• Maintain good eye contact. This does not mean staring fixedly at the person you are interviewing – doing so would probably only convince the intervie­wee that you are a lunatic. ‘Normal’ eye contact involves occasionally glanc­ing away. But it certainly does not include prolonged periods of staring off into space, minutely examining your tape recorder, intently writing notes, or fiddling with your computer.
• Try to monitor and avoid undesirable non-verbal cues (e.g. facial gestures that indicate disgust or disapproval, moving your chair away from that of the person you are talking to).
• Spend some time in ice-breaking chit-chat. Plunging directly into the inter­view tends to give the session a police-grilling quality. Allow for some getting-to-know-you time (which may be shorter or longer depending on the mood of the person you are interviewing or the amount of time you have allotted for the session) even if it appears that the topic of such ‘small talk’ is somewhat off track. In fact, in participant observation research, nothing is ever really completely off track – important proxemic and kinesic cues, as well as clues about people’s values and attitudes, often come through in these unguarded conversational moments. So even if the conversation appears casual, you cannot be completely ‘off duty’.
• Accept hospitality when offered. Many ethnographic interviews are held in homes, restaurants, or other places where people normally meet to talk (i.e. not sterile labs, imposing offices, or hushed libraries) and it is only nat­ural to share some refreshment as long as it comes in the form of manageable snacks – if a large, elaborate meal is in the offing, it is better to postpone the interview.
• Be aware of the condition of the person you are interviewing; do not overtax those who are in frail health or otherwise distracted, no matter how much you would like to adhere to your own agenda.
• Do your homework! While you may not yet have developed your own in­depth understanding of the people and their way of life when you begin your interviews, you should at that point not be completely in the dark. There will be things you have observed that you want to ask about – events, behaviors, expressed points of view that you will want to pursue and clarify. By this point you should know something about the major social institutions in the community, as well as something about the history of the group. You should also have at least a rough idea of who is who in the community and how they relate to one another.
• Personalize the interview. Ask the person you are talking to to share photos, scrapbooks, and other memorabilia that lend a personal touch to the com­mentary. You may also want to ask the interviewee’s permission to borrow those materials to copy or study further. If so, the originals should always be returned promptly and in the same condition as you received them. (If these items are of particular historical or cultural value, you may want to discuss with the interviewee the possibility of donating them to a museum, library, or other appropriate public institution.)

Some specialized interview types

The general instructions for the ethnographic interview as outlined above will serve well in most cases, but there are some situations in which specialized vari­ants on the interview method are helpful.

The genealogical interview was a staple of traditional anthropologists (and other social scientists interested in the lives of people in non-urban settings) because kinship – the ties of family and marriage – were often central to the ways in which ‘pre-modern’ communities were organized. The systematic collection of genealogical data could be used to elicit information about the patterns of inter­personal relations in the community. It could also be applied to studies of rules of descent (including ownership of property), marriage, and residence, as well as to studies of migration patterns and religious practices.

Kinship is rarely as central to modern urban communities as it was in the small-scale folk societies of earlier times. But even with increased mobility, the ‘ties that bind’ are merely attenuated, not absent. ‘Blood’ and marriage may no longer define a person’s place in the world, but the ways in which people estab­lish and maintain relationships with one another are still governed by definable patterns and expectations – they are not random or disorganized. And so the tra­ditional genealogical method has evolved into the social network analysis, which traces the connections among people in extended situations (such as members of the geographically widely dispersed Indian ‘diaspora’), often relying on sophisti­cated computer models to sort out these widely ramified links. Although in such cases the analysis per se needs to be done by complex technology, the data are initially generated by the same old-fashioned ethnographic means – asking peo­ple questions about their relationships – that characterized the genealogical stud­ies of several decades ago.

Using genealogical interviewing methods, I was able to determine that the pattern of sponsorship in the Trinidad AA operated through kinship lines. A man’s drinking partners were likely to be close relatives (especially his cousins on his father’s side) and when any one of them decided to seek sobriety, he would sponsor other mem­bers of the group. It happened that many of the regional AA groups were in fact com­posed of the members of what had once been a kin-based ‘drinking crew’.

It was very difficult to elicit genealogical information from the adults with mental retardation, but from what I was able to glean I was able to see that those who rec­ognized a strong kin network were usually more successful in completing their train­ing than those who felt disconnected from even abandoned by their relatives. While not conclusive by any means, such an insight could form the basis for a more struc­tured survey that could at a later date either confirm or disprove the association between strength of family ties and successful completion of a habilitation program.

Oral history is a field of study dedicated to the reconstruction of the past through the experiences of those who have lived it. While those with political or economic power often write their memoirs of great events, the ordinary people have often not had the opportunity to tell their stories. Oral history therefore pro­vides a way for those previously marginalized and rendered voiceless (e.g. women, members of minority groups, the poor, people with disabilities or of alternate sexual orientation) to put their stories on the record. The oral history interviewer brings together as many of the surviving participants in a given event of some significance (be it local, regional, national, or international) and gives them the opportunity to tell their personal stories – all of which together form a mosaic representation of that event. That representation may give us a different picture from that enshrined in the official history books and thus help put that official picture into a larger perspective.

A variant of oral history interview is the form of research known as the life- history. Rather than aim at a composite reconstruction of a particular event as in oral history, life history attempts to see the past through the microcosm of the life of one particular individual. Depending on the theoretical predisposition of the researcher, that individual could either be a ‘typical’ or ‘representative’ member of his or her community (such that his or her life story stands in for all those whose stories are not recorded) or an ‘extraordinary’ person (who represents the values and aspirations of the group).

The analysis of the extended narratives generated by oral and life-history research has been aided considerably by the development of computer software designed to pull out themes and patterns. But as in social network studies, no mat­ter how sophisticated the technology of analysis has become, the generation of data remains at heart the product of a traditional ethnographic interview.

My understanding of how and why the Trinidad Indians had become alcoholic despite a heritage of anti-alcohol culture was shaped by the oral history I col­lected from men who were in their forties and fifties at the time of my original research. They remembered back to the days of the second World War when Trinidad was used as a base by the U.S. air force. Trinidad was not in the thick of the war, and the young airmen had a lot of time on their hands – time they used by succumbing to the sensual pleasures of a tropical island. It was they who intro­duced the strongly American-influenced ‘rum and Coca Cola’ culture of conspic­uous consumption and hedonism. The young Indian men of that generation saw that the old colonial plantation system was a dead end, and they eagerly sought the jobs provided at the air base. But along with the jobs came the lifestyle they saw among the Americans. Drinking was no longer a taboo – it became part and parcel of the young Indians’ embrace of new economic potentialities.

Life histories form the basis of my research into the experiences of deinstitu­tionalized adults with mental retardation. Since my aim was to understand what it feels like to be mentally challenged in a complex, high-tech world, I could do no better than to see how people diagnosed with that condition had confronted the challenges of life. Unlike a clinical interview, which would focus on the specifics of the handicapping condition, a life-history interview gave the respondents the opportunity to talk about what was important to them in the course of their lived experiences. It was thus that I was able to discover the very strong concern about sexuality and the development of truly adult relationships.

While the classic ethnographic interview is open-ended in nature, as described above, it is also possible to conduct semi-structured interviews, which use prede­termined questions related to ‘domains of interest’ (e.g. ‘What are the ways peo­ple earn a living in this village?’, ‘What types of community-based programs are available to deinstitutionalized adults with mental retardation?’). Unlike the open-ended interview, which can roam rather freely around the area delineated by the general research questions, the semi-structured interview sticks closely to the prearranged topic and features questions designed to elicit information specifically about that topic. Digressions and new directions, so important in the open-ended interview, are not part of the semi-structured interview plan. The semi-structured interview should naturally develop out of an open-ended inter­view, following up and clarifying issues that came up in the course of the earlier, more conversational format.

The semi-structured interview may also be used to operationalize general factors into measurable variables, which can then be developed into working hypotheses, which in turn form the basis for a formal ethnographic survey (a closed-ended instrument designed to collect quantitative data from a relatively large number of informants). The mechanics of quantitative research are treated in the books by Flick (2007a, 2007b) in this series. The important point to keep in mind here is that in ethnographic research the large-scale survey with hypothe­ses testable through quantified data is an outgrowth of prior open-ended observa­tions and interviews; it is not a stand-alone method. Its strength is dependent on the value of the qualitative data that inform it (see Kvale, 2007, for more details of doing interviews).

A note on sampling

While there are recognized canons for determining the size of a population to be sampled in a purely quantitative study, the question of ‘How many [people should I interview, events should I observe]?’ can become something of a problem in ethnographic research. The best answer – although not necessarily the neatest or most definitive one – is that

As general as that rule may be, there are a few specific points you may wish to consider. Your sample should reflect the heterogeneity of the group you are studying. If it is a very diverse population, then you will need to interview and observe more in order to be sure that you have a good overview of all the differ­ent elements within the group. In a purely homogeneous group, a single person case study would be a legitimate ‘sample’. But as most study communities are in fact diverse to one degree or another, you should be aware of the range of varia­tion and include interviews and observations that reflect that range.

A note on recording interview data

Interview data are typically recorded on audiotape. Taping is a way of assuring the accuracy of what is said and, in the case of oral/life histories, it is essential to have the actual spoken voice available for replay. It should be noted, however, that audiotaping requires a fair amount of equipment (a recorder, possibly an external microphone, blank tapes, working batteries or an available electric 48 outlet) that may not always be feasible to acquire and tote around. While it is possible nowadays to buy reasonably inexpensive, more or less unobtrusive, but good-quality audiotape equipment, the cost of equipment goes up when one needs higher quality (e.g. for recording those voices that in and of themselves need to be preserved for posterity). Moreover, recorded tapes are only the begin­ning of a process; tapes need to be indexed and, in most cases, transcribed so that information can be efficiently retrieved from them. At best transcription is a slow, tedious process and the average researcher will have neither the time nor the skill to do it properly. On the other hand, the services of a professional transcriber can price a project out of the ballpark.

Although more and more ethnographers are using videotape to record a vari­ety of social interactions, it has not become a standard way of recording interviews, except among those who plan to use their interviews as part of filmed documentaries or other visual reports, or those who are particularly interested in capturing and analyzing the non-verbal aspects of the conversation. Although videotape equipment is readily available and not necessarily very expensive, it makes the transcription process even more difficult than is the case with audio­tape. Moreover, videotaped interviews present serious problems when maintain­ing the confidentiality of participants is at issue.

Unless one is an expert stenographer – and such people are becoming rare to the point of extinction – it is usually impossible to keep an accurate written record of an interview. Even if one were such an expert, it would be inadvisable to rely on such a technique as the researcher would then be spending an inordinate amount of time looking at his or her notepad, and thus losing valuable eye con­tact with the person being interviewed. An occasional jotted note is fine, but a complete written record is neither feasible nor desirable for most ethnographic interviews.

So, for better or worse, the audiotape remains the most valuable adjunct to the conduct of interviews and to the subsequent retrieval and analysis of the inter­view data. (See Schensul et al., 1999, pp. 121-200, for a thorough exposition of the theory and method of ethnographic interviewing.)

6. Archival research

Individuals and groups tend to collect stuff relevant to their histories, achieve­ments, and future plans. Sometimes the stuff is highly organized (e.g. minutes of meetings of a board of directors, family photo albums lovingly maintained by an ardent genealogist, back issues of newspapers). But more often than not it is sim­ply stored rather haphazardly and is thus often in a poor state of preservation. The challenge to the ethnographer is to find such sources of information, to make sense of them (in the likely event that they are not already organized), and to assist in their preservation for future researchers.

Some archived materials were originally collected for bureaucratic or admin­istrative purposes. These are known as primary sources and may include:

• maps
• records of births, deaths, marriages, real estate transactions
• census, tax, and voting rolls
• specialized surveys
• service system records from human service organizations
• court proceedings
• minutes of meetings.

It should be noted that even if these materials are highly organized and in a good state of preservation, they were not likely collected for the same purposes that animate the researcher. So the latter must still sort through them to get them to tell the story that he or she needs to hear.

Another potentially important form of archival data are the secondary data resulting from another researcher’s study. For example, a colleague who did fieldwork in Trinidad the year before I arrived had collected a great deal of genealogical information in support of her study of the transmission of certain genetic illnesses. I was not interested in genetics, but I was able to use the data she graciously loaned me to support my growing suspicion about the link between kinship ties and AA sponsorship. The fruits of many research projects are now available in excerpted and catalogued form on computerized databases. The Human Relations Area File is perhaps the best known of these sources of cross-cultural information.

Archival research rarely stands alone as an ethnographic skill, although it can certainly be the basis of a respectable stand-alone study if first-hand fieldwork is not feasible. But accessing and interpreting archived materials is almost always facilitated when the researcher does have first-hand experience in the community under study, and when he or she can check inferences made from the archived data in interviews with living members of the community under study.

There are several advantages to archival research:

• It is generally non-reactive. The researcher does not influence people’s responses, since he or she is not interacting directly with the people who pro­vided the information.
• It is usually relatively inexpensive.
• It is particularly important when one is interested in studying changing events or behaviors through time.
• It is also valuable when studying topics that might be considered too sensi­tive or volatile to observe or ask questions about directly.

On the other hand, the ethnographer using archived material should be aware of some potential problems.

• Archived data are not always unbiased: Who collected it? For what purposes? What might have been left out (intentionally or otherwise) in the collection process? Even haphazard collection results from a process of editorial selec­tion; the researcher who comes along later is therefore not dealing with ‘pure’ information.
• Even modern computerized databases are not always free of error: just because the information has been carefully transcribed does not mean that it was accurate to begin with.
• There can be physical or logistical problems in working with these data, which may be stored in inconvenient or physically unattractive (dusty, dirty, rat- or roach-infested) places.

Despite these caveats, however, archived data are simply too rich a resource to be ignored. (Berg, 2004, pp. 209-32, provides an excellent overview of the use of archival materials in ethnographic research.)

Source: Angrosino Michael (2008), Doing Ethnographic and Observational Research, SAGE Publications Ltd; 1st edition.