Janice M. Morse
QUALITATIVE STRUCTURED TECHNIQUES
Analysis of any kind involves a way of thinking. It refers to the systematic examination of something to determine its parts, the relationship among parts and their relationship to the whole.
—Spradley (1979, p. 92)
Since the 1960s, as a part of ethnographic fieldwork, cognitive anthropologists have used structured methods of ethnoscience for eliciting lay concepts from participants. Their goal has been to try and find a way to access and describe their participants’ worldviews using culturally embedded implicit knowledge. But for our purposes of identifying lay concepts, these techniques provide an excellent way to access concepts used by participants in their everyday discourse and to explicate those lay concepts. These techniques also help us to determine the various forms of a particular lay concept, to identify and delineate reporting of lay concepts (such as symptom reporting), and to move and develop these lay concepts toward scientific use.
Anthropologists use methods of structured interviewing to elicit what they call cognitive domains of cultural groups, using linguistic techniques and logical relations between and within classes. These techniques allow for the emic identification of concepts—that is, to identify concepts from the person’s perspective, and not from the perspective of the researcher. These strategies form an important set of techniques for the use of researchers who are interested in concept development.
TARGETED STRATEGIES FOR CONCEPT AND ATTRIBUTE DEVELOPMENT
The key, when eliciting the meaning of terms, is not to ask for meaning per se, but to ask for use (Spradley, 1979, p. 97). Although we do need to develop definitions as labels in concept development, the definition in itself is inadequate for understanding and is of very limited use in research.
Each concept belongs to a meaning system, or a collection of allied concepts. The meaning of any concept is derived from its relationship to these other concepts. Further, we must differentiate between what a concept denotes (i.e., things the concept defines) and what it connotes (refers to over and above their immediate referential meaning). For example, a lemon denotes a fruit of a certain color and flavor: That is, the characteristics of the thing in itself. But in the connotative sense, it may refer to something else—that is, a referent that shares a particular characteristic of the lemon, but not all. For instance, it could refer to something sour, bitter; to a color; to something faulty (as, “that car is a lemon”); to the shape of the fruit; and so forth. The referent (the thing that the concept refers to or represents) may be anything, either concrete or symbolic, but what is important is the relationship between the referent and the concept, and the fact that the referent may be using only one or more of the attributes, so it is not an exact replica.
The final way we infer meaning is to examine the parts of the concept or the parts of the whole. Recall that a concept refers to something—and a behavioral concept refers to a collection of behaviors. Techniques of eliciting what the concept refers to, its components, and the relationships between them may be obtained from the use of particular types of structured interviewing techniques. These techniques, such as, “free listing,” “semantic relations,” “sentence frames,” and “20 questions,” enable the associations between the concept and its referents, between the parts of the whole, or by identifying and examining its components.
Fieldwork: Identifying a Lay Concept
The process invariably begins with fieldwork. By observing the group, the researcher becomes familiar with the local forms of speech, the participant’s “world,” and the things that are mutually important to the researcher and to the participants. During the course of these observations, the researcher may stumble across an unfamiliar term, one with which the researcher is unfamiliar, and that warrants further inquiry as the concept of interest. For instance, Spradley (1970), when studying tramps in the Pacific Northwest, became interested in the term “flop” or “making a flop.” He was told that “the most important thing to a tramp is something to eat and a place to flop” (pp. 98–99), so that a flop was a term that tramps used to refer to the place where they slept. But what exactly is a flop? And how do tramps rate or rank the quality of their flops? This became a significant part of Spradley’s (1979) ethnography, describing the lives of urban nomads.
Another example: While doing fieldwork in emergency rooms, Morse noticed that nurses reported patient’s condition when asked: “What’s his comfort level?” The responses to comfort level appeared to be concerning the patient’s emotional and behavioral states: Patients were reported to be: “resting comfortably,” “really scared,” “tolerating it well,” and so forth. Yet, although comfort level was clearly an important concept used widely for rapid assessment and reporting of stages of relative comfort and discomfort, and was almost universally understood in clinical practice, it was not a concept that was found in nursing texts. That is, comfort level assessments were not formally taught or even acknowledged in the classroom.
Free Lists: Identifying the Domain
At this point, if the concept appears obvious, meaning that it is a term used frequently by the group and is not a one-time specialist word, the researcher may move directly to free listing. Free listing is a technique in which the researcher asks participants to list as many examples as possible of a particular term (or concept). These examples provide the researchers with the domain (scope) and variation. For example, you may ask the participant to list freely the words used to describe types of pain; the researcher would record the words, attending to both the accrual word used and the order is which the participants listed it. Using free lists ensures that your domain is developed inductively. Thus, it provides “an extremely useful and powerful tool” (Bernard, 2012, p. 233) for concept identification.
The conceptual domain (or the scope) may be elicited from participants’ words, lay concepts, and sentences used in everyday language, that are all in the same level of contrast, using free lists and pile sorts (Bernard, 2012, pp. 233–235). By asking participants to quickly and out loud to list terms enables the concept to be explicated by the participants, and from these lists, examples of the concept, what it is not, and where the boundary lies, may be developed.
The terms selected by the participant are usually at the same level (i.e., they are not components of parts; Weller & Romney, 1988). In other words, asking this question of 10, 20, 30, or more people will give you lists of equivalent terms. Some terms will be offered by many of the participants. Count them and sort them according to the frequency that they are used. Those that occur most frequently are the more common terms, and may be more important, or salient. Some may be mentioned by only one participant, and therefore occur only once. As your sample size increases, so will your list become more stable, and new terms will be less likely to appear (Weller & Romney, 1988).
When compiling a free list, the researcher asks the participants: “Name all of the types of_________ that you know?” or “What types of _____________ are there?” For instance, Spradley (1970), used free lists, supplemented with interviews, and found that drunks listed almost 100 types of flops or places to sleep (pp. 99–104). If we continue our example of comfort levels, we would ask: “Name all of the types of comfort levels that you can think of_____?” and “What types of comfort levels are there?”
These terms, representing examples of your comfort level, provide you with your concept, or what anthropologists call your conceptual domain. The descriptors of terms for types of comfort level as listed by 15 nurses as examples of emergency department patients’ comfort level are listed by frequency in Table 15.1. Note there are no restrictions on the number of terms a participants may provide—in this hypothetical study, the most terms listed by any one participant were 20.
All of the nurse participants thought that patients’ comfort level may be scared, anxious; most thought of frightened patients. The states that are listed less frequently may be patients who are not seen so often, or lesser known terms.
What sorts of concepts are best suited to free listing techniques? Bernard (2012) suggests rich domains, subjects that are not concrete, not taught, not structured. Questions must be carefully considered: For instance, if you were asking about ingredients used to bake a cake, most would respond with flour, sugar, baking powder, eggs, salt, and flavoring (chocolate, lemon, vanilla), and there would be little variation in the responses; but if you were asking flavors of cakes or types of cakes, the variation would increase. We would, for instance, ask for types of cakes and get answers such as Christmas cakes, birthday cakes, sponge cakes, chocolate cakes, pavlova, cream cakes, fruit cakes, and so forth. As most researchers use the technique in areas or topics with which they are not familiar, the answers are always interesting. Hopefully your topic, obtained from 20 or 30 participants, will produce at least 30 items from which you will be able to produce a frequency list.
The format of the interviews for identifying folk concepts are somewhat structured, with the researcher planning to ask a few questions, to which the participant may respond freely. For instance, to get more contextual information about comfort levels, the interviewer may ask:
“Tell me about _____________” (a particular type of patient).
Then ask, “What is his comfort level?”
“What are the signs that a patient has a _____________ comfort level?”
“Describe other patients who have had a ____________ comfort level?”
And you may ask: “Is ‘comfort level’ a term used by nurses, or do you use it when reporting to non-nurses, such as physicians or physical therapists?”
The goal of these guided interviews is to get as much contextual information as possible that will later help you understand, define, and describe the term (in this case, comfort level), to determine appropriate usage, where, when and why it is used, and to what it refers. These interviews are transcribed and analyzed. Categories are developed showing the use of the term, the scope, and the possible attributes. Once this information is obtained, the investigator then moves onto conducting structured interviews, such as semantic relations.
Semantic relations are the identification of terms within a language, and the terms with which they are related or linked (Spradley, 1979, pp. 100–101). Research is conducted as follows:
• They are identified by listening carefully when participants talk and when reading the interview text. They may be related to the names of things and occur in a set of two or more in relationship to each other. The concept may be of one of three types of relationships:
Inclusion: An X is a kind of Y.
Attribution: An X has Y.
Sequence: From X comes Y.
• Prepare to interview about the particular concept. Prepare question prompts from an extended list of relationships. The types of semantic relationships and type of questions to be asked to elicit this are as follows:
1. Strict inclusion is asked as: X is a kind of Y?
2. Spatial is asked as: X is a place of Y, Y is a part of X?
3. Cause and effect is asked as: X is a result of Y, X is a cause of Y?
4. Rationale is asked as: X is a reason for doing Y?
5. Location for action is asked as: X is a place for doing Y?
6. Function is asked as: X is used for Y?
7. Means-end is asked as: X is a way to do Y?
8. Sequence is asked as: X is a step (stage) in Y?
9. Attribution is asked as: X is an attribute (characteristic) of Y?
(Spradley, 1979, p. 111)