Log in / Register
Home arrow Environment arrow Research Methods in Anthropology: Qualitative and Quantitative Approaches


Another common technique in cultural domain analysis is called the sentence frame or frame elicitation method. Linda Garro (1986) used the frame elicitation method to compare the knowledge of curers and noncurers in Pichataro, Mexico. She used a list of 18 illness terms and 22 causes, based on prior research in Pichataro (Young 1978). The

frames were questions, like ‘‘can_come from_?’’ Garro substituted

names of illnesses in the first blank, and things like ‘‘anger,’’ ‘‘cold,’’ ‘‘overeating,’’ and so on in the second blank. (ANTHROPAC has a routine for building questionnaires of this type.) This produced an 18 X 22 yes-no matrix for each of the informants. The matrices could then be added together and submitted to analysis by multidimensional scaling (see chapter 16).

James Boster and Jeffrey Johnson (1989) used the frame-substitution method in their study of how recreational fishermen in the United States categorize ocean fish. They asked 120 fishermen to consider 62 belief frames, scan down a list of 43 fish (tarpon, silver perch, Spanish mackerel, etc.), and pick out the fish that fit each frame. Here are a few of the belief frames:

The meat from _is oily tasting.

It is hard to clean__

I prefer to catch__

That’s 43 X 62 = 2,666 judgments by each of 120 informants, but informants were usually able to do the task in about half an hour (Johnson, personal communication). The 62 frames, by the way, came straight out of ethnographic interviews where informants were asked to list fish and to talk about the characteristics of those fish.

Gillian Sankoff (1971) studied land tenure and kinship among the Buang, a mountain people of northeastern New Guinea. The most important unit of social organization among the Buang is the dgwa, a kind of descent group, like a clan. Sankoff wanted to figure out the very complicated system by which men in the village of Mambump identified with various dgwa and with various named garden plots.

The Buang system was apparently too complex for bureaucrats to fathom, so, to save administrators a lot of trouble, the men of Mambump had years earlier devised a simplified system that they presented to outsiders. Instead of claiming that they had ties with one or more of five different dgwa, they each decided which of the two largest dgwa they would belong to, and that was as much as the New Guinea administration knew.

To unravel the complex system of land tenure and descent, Sankoff made a list of all 47 men in the village and all 140 yam plots that they had used over the recent past. Sankoff asked each man to go through the list of men and identify which dgwa each man belonged to. If a man belonged to more than one, then Sankoff got that information, too. She also asked her informants to identify which dgwa each of the 140 garden plots belonged to.

As you might imagine, there was considerable variability in the data. Only a few men were uniformly placed into one of the five dgwa by their peers. But by analyzing the matrices of dgwa membership and land use, Sankoff was able to determine the core members and peripheral members of the various dgwa.

She was also able to ask important questions about intracultural variability. She looked at the variation in cognitive models among the Buang for how land use and membership in descent groups were related. Sankoff’s analysis was an important milestone in our understanding of the measurable differences between individual culture versus shared culture. It supported Goodenough’s notion (1965) that cognitive models are based on shared assumptions, but that ultimately they are best construed as properties of individuals.

Techniques like true-false and yes-no tests that generate nominal data are easy to construct, especially with ANTHROPAC, and can be administered to a large number of informants. Frame elicitation in general, however, can be boring, both to the informant and to the researcher alike. Imagine, for example, a list of 25 animals (mice, dogs, antelopes . . .), and 25 attributes (ferocious, edible, nocturnal . . .).

The structured interview that results from such a test involves a total of 625 (25 X 25) questions to which an informant must respond—questions like ‘‘Is an antelope edible?’’ ‘‘Is a dog nocturnal?’’ ‘‘Is a mouse ferocious?’’ People can get pretty exasperated with this kind of foolishness, so be careful to choose domains, items, and attributes that make sense to people when you do frame elicitations and true-false tests (Further Reading: sentence frames).

Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >
Business & Finance
Computer Science
Language & Literature
Political science