Content created: 2012-04-29
File last modified:
For generations, "participant observation" has been the prime source of data for cultural anthropologists. This normally involves residence among one's informants for a long time, typically a year or two, sometimes ten or more. For class purposes, this is not feasible, and data collection is more dependent upon formal interviews.
Informants are usually delighted to talk to you, especially to tell you about things that they are expert about and interested in, such as themselves. However never forget that they are you doing you a favor, however much they may enjoy it. Be courteous, obliging, sympathetic, encouraging, and above all obviously interested. Smile more than you usually do. Bringing them cookies is absolutely in order. So is thanking them profusely.
Informants' Comfort. Formal interviews require the full knowledge and cooperation of the informant. To conform to modern "human subjects research" standards, informants should always be told that there is no obligation to answer any particular question, and that the interview can be stopped at any time. (This is different from journalism or police interrogation. Journalists and police interrogators have standards too, but they are not considered to be dealing with "human subjects" and therefore have less obligation to make informants feel comfortable.)
Informants' Privacy. The informant should be told what level of confidentiality to expect. For example, "I don't plan to discuss what you have told me in a way that can be traced to you. Or, "Barring the unforeseen, only my professor and I will be able to link your name to the material you provide to me." Or, "I expect to discuss specific cases only with pseudonyms." (Click me.)
Time Commitment. You should tell your informant in advance roughly how much time you think an interview will require, so the informant knows how to plan and how extensively to answer your questions. (In contrast, be sure that you yourself have enough time for it to go much longer if the informant turns out to be providing good information and enjoying doing so. Some informants love talking and can be longwinded.)
It is possible to conduct an interview over Skype or on the telephone. It is not as fecund as a face-to-face interview, but it can still be useful. Interaction through a keyboard, however, should not be considered an interview. It is simply not realistic to imagine that your informant will type more than s/he will talk, and pretending that an Email exchange is all your informant understands about something is unfair both to your informant and to your reader. Conquer your inner wimp and face the world of living humans.
Preview of the Topic. Tell the informant ahead of time what the broad topic is that you are interested in. That way you will probably avoid being turned down after you start if the discussion goes in directions your informant didn't anticipate. (Click me.)
Sometimes you are interviewing an acknowledged expert about something (a famed potter, for example), and most of the interview will probably focus on that person's expertise. Sometimes an expert is in the position of speaking for an institution (a person such as the founder of a company or the head of an office). That is more difficult, since such a person may try to make the institution look good, or anyway, not to reveal anything that might make it look bad. In these examples, the potter may be happy to tell you how many pots break during firing. The head of an office will probably not tell you how many people get fired for various reasons. (That is one of the reasons why journalists often like to talk to former politicians and military officers rather than the ones still in office.) Whether this complicates your project depends, obviously, on the goals of your interviewing.
At other times you may interview a person who does not consider himself to have any special knowledge of what you say you want to know about, and who tells you to go find an expert. ("I don't know much about history. Go look it up in the library.") To avoid that reaction, keep the topic broad, but also stress that you are interested in the informant's particular perspective or in what the informant may have heard or learned from others. (For example: "We have been studying about the goddess Mazu, and I am really interested in what individual Taiwanese people know about her or think about her and how they learned about her.")
Non-Directive Interviewing. In general, you hope to learn from informants something you don't already know. This means it is important to ask questions in such a way that you don't accidentally "lead" informants into saying what you already think rather than what they mean. In other words, you need to conduct "non-directive" interviewing.
For example, imagine that you are interviewing an elderly Chinese informant about Chinese and Japanese attitudes toward each other. If you say "Tell me about why Chinese so hate Japanese," you have told the informant what to think, and thrown into relief an attitude that your informant may not hold or may hold but only very weakly. Your informant's attitudes towards Japan or Japanese will be far more reliably visible if they emerge on their own (although naturally a follow-up question can pursue a subject once an informant raises it). Even if the informant denies your lead ("We don't hate the Japanese!"), you have tainted the discussion by putting him or her in the position of having to defend or attack a point of view that you have introduced. This is only rarely very productive.
A non-directive approach would perhaps ask, "Tell me a little bit about relations between Chinese and Japanese at that time." This doesn't presume that relations were good or bad, or even that good-bad is the relevant axis on which to describe them. (You may have surprises. For example: "I thought the Japanese soldiers were incredibly cool, and I felt really inadequate not being able to speak Japanese.) Once the informant starts on a discussion, you can then ask for examples or clarifications that gradually elicit what it is that you are trying to learn.
Directive Interviewing. Despite its advantages, non-directive interviewing is not usually very efficient at getting around to what you most want to know. You nearly always need some specific questions to set the topic or to follow up or clarify. For example imagine that you are eliciting a life history and have reached the school years. You might say, "I'd like to start by learning a little about what your school was like when you lived there." Or: "Can you tell me a little more about some of the other students? Did you have any special friends among them?" But in some cases a straightforward question makes the most sense: "Was she his sister or his aunt?" "Do you remember what year that was?"
Sustaining Flow. Most informants prove pretty talkative if they are on a topic they both know and care about. HOWEVER:
Pauses. Some people have higher tolerance for pauses in a conversation than other people do. The fact that your informant stops talking for a time doesn't necessarily mean you should move on right away. It may be no more than a normal pause to organize what to say next. Then again, it may be that the infrormant thinks the subject is finished and needs a little nudge if you want more detail. Being aware that there are personal and cultural differences is the first step to treating the situation productively.
Actual lags in the conversation are opportunities. Sometimes an informant has said something provocative but somewhat beside the point at the time. If so, you may want to shift to that topic when the conversation seems to run down later. ("You mentioned earlier that you didn't go there any oftener than you had to. Why was that?") Other times the conversation may have moved off of a subject you wanted to know about into something else that seemed interesting and relevant, at least to the informant, but left you hanging a bit. As an attentive interviewer you may wish to return when opportunity presents itself. ("We never really got back to the issue of your violin lessons." "I don't think I really understand why you didn't think he should know about it.")
Focus Groups & Zoom Rooms. Journalists, advertisers, and sociologists make use of focus groups. The typical focus group consists of a half dozen or so diverse community members who are asked open-ended questions about public policy. Focus groups are far more formal than casual discussions, but can discover or illustrate sometimes unexpected reactions to a marketing slogan, a policy proposal, or a current event.
Participant observation often involves discussions among groups of people, obviously, even if nobody identifies the participants as a focus group as such. However some anthropologists have made good use of directly stimulated conversations on topics that they want to learn about. One UCSD anthropologist, living in a Muslim community in the southern Philippines, for example, recorded (with permission) groups of female friends discussing their religious lives and practices, and how they differed from the perspectives of men they knew. Because they were talking with each other, there was no barrier of language or formality that could have hindered their communication one-on-one with the anthropologist, who largely just listened and occasionally asked a question to keep the discussions going.
With the Covid-19 pandemic of 2020, corresponding with the flowering of Skype, Zoom, and other web-based interviewing tools (often with built-in recording functions), "virtual" interviews and focus groups became inevitable. Though hardly participant observation in the classic sense, they have great potential to be very productive. Caution: It is usually very difficult and often impossible to sort out overlapping voices on an audio recording. See below.
Biased Questions. We all know that how a question is worded can affect the answer. That is why the opposite of "right to life" is "right to choose" when both phrases refer to the legalization of abortion, or why "freedom of religion" is sometimes a code word for opposition to "gay rights." However even very minor shadings of difference can affect the outcome of a question.
If you do an opinion poll —which is not ethnographic interviewing— this is a particularly severe problem because there is not usually enough redundancy in answers for the bias to be evident. With opinion polls, intended for hundreds or thousands of respondents, it is possible to pretest alternate wordings of questions. In ethngraphic, one-on one, unique interviews, that is not possible. However the interviewer still needs to be very alert to exactly how a question is being posed. (Click me.)
Linguistic Interviewing. Technical analysis of language features requires quite different interview techniques from the ethnographic ones discussed here. But less technical sociolinguistic topics arise in anthropology courses, and they necessarily involve observation of speech patterns. Often this does not involve formal interviewing, and actually may require unobtrusive observation of public behavior. For example, you may hang out in a coffee house and notice how many times women, as against men, address each other as "you guys" because you want to know whether the usage is gender-tagged, and if so, how much. Or you may wonder what proportion of the time the word "dude" is accompanied by a warning or word of advice.
Sometimes what you need is a sample of sustained speech from a single informant or a series of single informants. For example, some people do and some do not change the pronunciation of "the" depending upon whether the following word starts with a vowel ("thuh" goose as opposed to "thee" elephant), and you wonder making the change correlates with age, dialect, social class, or gender. What is needed here is a speech sample, not particular content; but it is usually critical that the speech be unselfconscious; the interviewee must not be focusing on speech and speaking, but on the context and content.
Even if you tell the informant you are interested in the way people express themselves, you can't say exactly what you are looking at until afterward, since that would produce a self-consciousness that would almost certainly distort the data you are trying to collect. You need to find a way in which the informant is happy to give you as long a speech sample as you need, while not knowing exactly what you are looking for in it. This is usually not difficult, but it is important to be attentive to keeping the speech spontaneous. (It is also important that the interviewee understand that you are interested in speech and are trying to collect a completely natural sample.)
There is no way you are going to remember everything you are told in an interview, any more than you remember everything that happens in a class. (And we all know how that turns out.) Obviously, you need to take notes, but there are some risks to be aware of:
Given these difficulties with taking and using written notes, audio recordings seem as though they should be the obvious solution to all these problems. In fact, audio recordings can help a lot, BUT:
When the time is up or the questions have run out or the informant has dried up, you should thank him or her profusely and ask if it would be okay if you make contact again in event there is something you don't think you got quite right. Most informants will answer in the affirmative. (That also means you need to keep adequate contact information.)
Some informants will ask if they can see your final write up. A copy of your interview notes is usually fine (from which you should probably redact comments like, "she begins to weep" or "he suddenly becomes very uncomfortable as though hiding something"). The informant will usually find them boring.
A copy of your analysis is a bit chancier, since it may treat issues that your informant finds of little or no interest, regards as clear evidence of your stupidity, or considers to be insufficiently flattering. The informant is not in your course, doesn't understand the context of your assignment, and has not seen interviews with other informants. Don't make promises you are not prepared to keep, but you may find it a good idea to be rather vague about just what you promise.
Once you have interviews from more than one informant, your interviews are your claim to fame because they are a data source available to nobody else. They are subject matter about which you are the unique world authority. Whatever their limitations, they are what justify your claim to the attention of your readers or listeners. They are critical supports for your best analyses and they are the first-order disconfirmations of your wrong hunches. They should be as organized and helpful to you as you can make them. You don't have to put them on an altar and sacrifice chocolate-chip cookies to them, but treat them with the respect they deserve.
The first step in that is making sure they are well organized and adequately indexed so that you can find information when you look for it. Beyond simply backing stuff up, three "best practices" should be observed almost fanatically:
If you make those three assumptions, you will soon find yourself organizing things adequately. Well, actually, you should probably also assume that if you do not make those three assumptions, or if you procrastinate, you will be eaten by a big, bad wolf.