We can divide qualitative research into a series of steps; (1) gaining entry; (2) category definition and observation; (3) data recording; and (4) analysis. The following brief discussion will address each of these steps.
For detailed guides to participant observation, see the treatments by Bernard (1988), Lofland (1971), Patton (1990), Schwartz and Jacobs (1979, Part II), and Werner and Schoepfle (1987a; 1987b). You can find examples of the problems and procedures of qualitative research in Glazer(1972) and Golden (1976).
Gaining Entry and Finding a Key Informant. The field researcher must attend the setting during the interactions of interest and occupy a role that does not cause the actors to change their natural behavior. Of course, the researcher can attempt to gather data by interview only.
Sometimes, one or more key informants will provide information to the interviewer.
However, data drawn only from informants leave the risk that they will mislead the researcher (as alleged in Freeman’s critique of Mead noted earlier). Even if the informants speak in good faith, one might doubt their accuracy.
For example, what kind of person would offer to serve as a key informant to a stranger come to study his or her group? Some key informants may be deviant members of their own community and, therefore, not well placed to describe it to others (Bernard, 1988).
Although helpful later in suggesting important areas of observation and interpretations of data, informants initially serve as contacts linking the researcher to the informant’s network.
Thus the researcher typically begins the entry process by persuading one or more members of the setting to accept him or her. If an observer can define his or her role in a satisfactory way, other members of the setting will permit entry as well.
For example, Lofland (1966) and the Festingerteam (1956) both gained acceptance by approval from central leadership of each cult under study.
The method of gaining acceptance by key members of a setting varies with the nature of the setting. Sometimes, the key informants may know the observer’s role (for example. Tally in Tally’s Corner). In others, acceptance may require that the observers conceal their research interest (for example when Prophecy fails).
In both cases, the researcher’s role must permit him or her to be curious and in need of instruction by other members. Sometimes such student rote will seem natural, as in the case of prospective recruits to a cult’s belief system.
In other cases, the observer may ‘”purchase” his or her entry and continued presence in some way. For example, the researcher may offer aid, as Liebow offered to one of Tally’s friends in a legal matter or by serving as watch queen in Humphrey’s study.
After a while, the observer may enjoy the bond of friendship with one or more of the setting’s members. Possibly one or more actors will become interested in the researcher’s task and support the activity out of curiosity or a desire to produce the best possible report about his or her own setting.
Although entry into a new setting may seem an enjoyable process of making new friends, some researchers find it harrowing. Some settings carry risks (for example, Humphrey’s arrest for loitering near a public restroom). First contact with a truly different subculture may inspire either excitement or “the desire to bolt and run” (Bernard, 1988, p. 163).
Even if the initial contact goes well, the qualitative researcher may feel “some form of depression and shock thereafter (within a week or two).
One kind of shock comes as the novelty of the field site wears off and there is this nasty feeling that anthropology has to get done. … Another kind of shock is to the culture itself. Culture shock is an uncomfortable stress response, and must be taken very seriously”.
Category Definition and Observation:
In the typical quantitative study, category or variable definition precedes data collection. In contrast, participant observers have only general questions as their guides and cannot translate them into standardized measures.
As a result, the qualitative researcher both observes and chooses what to observe at the same time. For example, the Festinger team (1956) knew at the beginning that they wanted to study proselytizing.
But they only discovered how proselytizing would appear in behavioral terms in the process of observing. Thus, such acts as being more or less responsive to callers from the media came to be included in the category of behaviors representing proselytizing.
The alternating process of observing and defining categories will repeat throughout the study and helps to avoid blinding preconceptions and rigid measurement procedures. The flexibility and sensitivity of this approach may allow profound penetration of the subject matter.
As counterparts to these advantages, this method faces the threats of unreliable measures, instrumentation shift, and observer bias.
What is to be observed will vary from study to study, depending on the topics, and overtime within a study, evolving with the definitions of the categories and the researcher’s interests? Lofland has noted that different levels of analysis may interest the observer, ranging from brief acts up to entire settings, and that recorded observations provide either static cross-sectional or longitudinal descriptions.
In light of this complexity, how does the observer make sure to cover and record the full range of categories?
Comprehensive observation requires the observer’s presence at the times and places of the actions of interest. Unless the observer knows in advance when these activities are going to occur, he or she must take up virtual residence in the setting.
If information comes from interviews, the interviewer has more control of when and where the interview will take place. Still, how can the interviewer ask about each point of interest? In the absence of standardized questionnaires, the interview at least can follow some preset guidelines.
A written or memorized interview guide provides a checklist of topics that the interviewer wants to cover. These checklists include reminders about the categories of interest to the researcher in an order that seems likely to promote rapport (usually holding more threatening questions until later).
The guide provides only a general approach. The actual questions are composed on the spot to fit the natural rhythm of the dialogue and to promote maximum, unbiased disclosure by the interviewee. The art of such interviewing requires a sensitive ear and the ability to gently probe without suggesting any desired answers.
The researcher must record the observations and interview material either by audiotape or by handwritten notes. Although taping one’s notes and interviews may seem faster initially, transcribing them from tape may take six or more hours for each hour of interview.
In either case, the resulting transcripts should have labels with basic information such as the interviewee’s identification (name or code number), demographic characteristics, and location in the social network under study along with the time and place of the interview or observation. Bernard reminds us that “the difference between field work and field experience is field notes”
1. Make many shorter notes instead of a few long ones,
2. Separate notes into four types:
(a) field jottings done on the spot, (b) Field notes based on the jottings, (c) field diary about personal reactions that can later help interpret the notes and reveal observer bias, (d) field log of how time was spent.
3. Take field jottings all the time.
4. Do not fear that you can offend people by jotting notes.
5. Set aside time before the end of each day for writing up nil notes.
In the case of unrecorded interviews and of observations of behaviors in the setting, the observer must make the notes as extensive and accurate as possible.
Make field notes as soon as possible after observation. One exception to Bernard’s fourth rule about not offending people with note taking involves disguised research. Since conspicuous note taking could spoil the observer’s accepted role, the unobtrusive researcher may have to make brief notes of key words and phrases in fleeting moments of privacy (for example, going to the bathroom or stepping out for a walk).
As soon as possible thereafter, and certainly before going to sleep, the researcher must construct fully detailed notes from the jottings and recollections.
Whether originally taped or handwritten, the notes should then be typed with multiple copies. One copy can serve as a permanent record of the raw data.
The analyst can rearrange or annotate other copies as part of the interpretive process. One type of annotation involves coding each observation with a number that allows you or others to quickly look up all instances of any behavior type. For example, if you were studying cult recruitment, you might devise a code number to put in the margin of your notes next to every instance of recruiting behavior.
For such a note-coding system, some anthropologists recommend using the Outline for Cultural Materials (Murdock, 1971). Because qualitative research may extend for months and the researcher does not want to omit any potentially useful details, the resulting documents can easily run to hundreds or thousands of pages.
Given such massive data sets, qualitative researchers increasingly record their field notes and codes in a computer database to enable fast relational file searches.
The analysis of qualitative data begins with the first observation. As the observation phase winds down, analysis becomes more intense.
Analysis organizes the hundreds of pages of raw notes into a meaningful pattern. It interconnects discrete observations and locates these connected events within a small number of conceptual categories.
As with a jigsaw puzzle, the researcher fits and refits the pieces according to a variety of tentative models until few unconnected pieces remain and the fit seems subjectively and logically satisfying.
A final report-gives the resulting “jigsaw” picture as clearly and convincingly as possible. A common reporting method combines quotations from interviews with anecdotes from the field observations to illustrate and support the analyst’s general argument.
In support of a causal model, the analyst may report the approximate frequency and distribution of the different categories of observation (for example, high versus low proselytizing) as evidence.
Such event counts may even support basic statistical analysis, but the qualitative researcher seldom relies as heavily on statistics as does the quantitative researcher.