Notes on Spradley's
Ethnographic Interview

Steve Borgatti, Boston College

 

Doing Ethnography

Discussion issue:  Spradley seems to think that it is ok and indeed desirable to do research that actively helps your informants. Oppression of jails. Plight of the gypsies. Doesn't this remove objectivity? Cancian: we must empower our respondents and challenge the power structures that oppress them. Liberating commercial sex workers and pornography stars.

Language

Informants

 

HOW TO DO RESEARCH

Locating informants

Characteristics:

  1. Choose informants with thorough enculturation. They know their own culture well. Culturally competent.
  2. Choose informants with current involvement. Not retired from that lifestyle or culture
  3. Choose cultural scene unfamiliar to researcher. Three reasons: 1, prevents taking things for granted, makes them sensitive. 2. analysis of familiar is difficult. becomes superficial. can't find anything of interest. see no point in writing down the obvious. Or worse, you deliberately "complexify" and "problematize" the mundane. 3. can't sincerely ask dumb questions. people think you are putting them on.
  4. Choose informants with adequate time to be an informant. Must be available!
  5. Choose informants who are non-analytic. i.e., not versed in researcher culture. just repeat back researcher's culture. great ex. on pg 53.

Discussion issue: Spradley seems to think the ideal situation is to study a completely alien culture. This differs somewhat from Bartunek's view, does it not? Spradley's view is interesting because his orientation is totally emic -- to understand things from the insider's point of view, but he thinks this is best done by an outsider.

Interviewing

An interview is a specific speech event, separate from conversation, lecture, job interview, etc.

Characteristics of an ethnographic interview:

  1. There is an explicit purpose that comes from the ethnographer.
  2. Interviewer provides framing explanations. 1. what project is about. 2. what i'm doing right now -- writing things down. 3. asking for native language. 4. what kinds of questions i'm asking now.
  3. Three types of ethnographic questions. 1. descriptive (what do you do all day at the office?). 2. structured (domains, paradigms, attributes, relations, items). 3. contrast (what's the difference between bass and pike?)
  4. Asymmetrical turn taking rather than sharing.
  5. Repeating and restating (also known as active listening)
  6. Expressing interest
  7. Expressing ignorance
  8. Encouragement to expand rather than abbreviate responses
  9. Incorporating informants terms without mocking
  10. Use of hypotheticals. E.g., "if we were standing in the bar now, what would it be like?"

Need to be able to deal with embarrassment. Be able to say things like "A moment ago, you described your acquaintance Charlie as "an asshole". Are there different kinds of assholes?"

Making an ethnographic record

Field notes, tape recordings, pictures, artifacts. (remember Goodwin?)

When writing notes, remember to identify which are your terms and which are informant terms. Whenever possible obtain verbatims and don't unconsciously translate. many errors here. later you find that you didn't get it at all.

Field notes are notes that you write up at the end of each day of interviewing. It helps give you some distance and also begin synthesizing and assimilating. Makes you a better observer and interviewer the next day.

Once you start the analysis, will also have coding notes and theoretical notes.

Descriptive questions

1. Repeated explanations and framings of what you are doing and asking about

2. Restating embodies non-judgmental attitude which contributes to rapport.

3. Where possible don't ask for meaning (why ... , what do you mean by ... ), ask for usage. It's not that you don't want meaning, it is just that sometimes explicitly asking for meaning implies justify, explain better, be better at being informant. So it can be better to ask what would you say if x was different, or what other kinds of situations would be described that way?

Basic types are:

  1. Grand tour. a. would you describe the inside of a jail for me? b. what's a typical night at the bar like? c. What happened last night? d. show me around the office. e. can you play a game of  backgammon and tell what you are doing as you do it?
  2. Mini tour. smaller unit of experience. Tell me what happens in a phone call.
  3. Example questions. What's an example of "blowing it" ?
  4. Experience. What are some of the experiences you've had working with professors?
  5. Native language questions. How would people normally refer to getting drunk?

Analyzing Interviews - what to look for

  1. Learn the cultural referents of words and concepts. E.g., what every person in the culture "knows" about mice.  [Note that these are stylized cultural "facts" -- not objective truths.] Children like them. Adult females and elephants scared of them. Mouse traps used to get rid of them. Theme park and cartoons use mice as characters. Mice are subjects of scientific experiments in medicine. Get rich building a better mousetrap. There's only a few things, but they are not given in the dictionary definition of mouse. Yet anyone that is culturally literate knows them and assumes that others know them.
  2. Categories/domains. In cultural worlds, things are grouped. There are categories within categories, and relationships between categories. Boys is a category, so is girls, and they contrast when seen as elements of larger domain of children.
  3. Special category of relations among categories are sequential relations like months of the year, steps in a wedding, stages in a life.

Making a Domain Analysis

Steps:

  1. Select a semantic relationship such as "causes" or "is a kind of"
  2. Select a sample of informant statements. I.e., choose a corpus to analyze.
  3. Search out words or phrases that could be parts of a category (e.g., marked linguistically as a list, as in "first of all there, there are Xs. Then there are Ys and Zs..."), and search out possible cover terms. these are domains.
  4. Formulate structural questions for each domain. Testing hypotheses about relations between domains and domains and items. Like: "Are there different kinds of parties?" If yes, it is a domain. Then ask "what kinds of parties are there?"

Structural Questions

Structural questions don't occur as often as descriptive questions in normal conversation, so they require more framing. They also tend to elicit lists, so you need to ask many times (repetition) to get all the items in the list.

Remember to ask for a cultural perspective in addition to a personal perspective. Fundamentally, you want to know what flavors of ice cream are recognized in the culture, not just the ones that your informant likes -- you are not really writing a book about the informant but rather her culture. 

Types:

  1. Verification & elicitation questions
    1. verification of hypotheses. Is a hotel lobby a kind of flop?
    2. domain verification. Are there different kinds of assholes? What are the different kinds?
    3. included term verification. is X an illegal activity?
    4. semantic relationship verification. would people here ever say something like "dancing IS life?" or "dancing is like life?"
  2. Frame substitution. start with real sentence like "you get a lot of assholes in finance" . Then ask, can you think of any other terms that go in that sentence instead of assholes? you get a lot of _____ in finance. (can do this systematically by giving them list of terms to choose from).
  3. Card sorts. write phrases or words on cards. then lay them out and ask the questions above. can ask which ones are similar.

Taxonomic Analysis

Folk taxonomy is set of categories (domains) organized on a single semantic relationshisp, such as "a kind of".

Great example of organization at police dept. on page 140.

Need to test your taxonomies against informants.

Contrast Questions

Find out how a given symbol (word, event, construct) is different from others. This is how to discover dimensions of meaning.

What does "man" contrast with? (god, dogs), woman,  boy, wimp,

Types:

  1. dyadic comparisons: what differences are there between these?
  2. triadic. which is the most different?  shark  seal dog
  3. sorting cards.

Componential Analysis

Componential analysis systematic search fro the attributes or features of cultural symbols that distinguish them from others and give them meaning. The basic idea in componential analysis is that all items in a domain can be decomposed into combinations of semantic features which combine to give the item meaning. This is a perspective that was first developed in linguistics. [go see handout on this]

Trusty ...

pg 178 compares with drunks, lockups, kickouts, etc.

Steps:

  1. select contrast set (set of items)
  2. inventory all contrast features; select binary ones
  3. collapse closely related dimensions to create multichotomies
  4. conduct new interviews to elicit missing data and suggest new dimensions

The componential approach, together with the assumption of analogical processes is the basis for Levi-Straussian analysis of stories and culture in general.

Cultural Themes

Theme is a postulate or position, declared or implied, which is tacitly approved and promoted in a society. It's an organizing principle, such as, for Apache culture, Men are morally superior to Women. Or for our culture, Men are sexual predators, from  whom women and children must be protected.  Ruth benedict: cultures are not trashbins of memes but rather coherent: complex patterns whose components are interrelated.

Theme is a cognitive principle on which other things rest. Can be TACIT or EXPLICIT. Most are tacit, so one objective of ethnography is to reveal these hidden themes. They also connect up different elements of a culture.

ME: You can also think of them as theories that explain (make coherent) large sections of a culture).

Strategies:

  1. total immersion in culture
  2. make a cultural inventory
  3. make a componential analysis of all domains
  4. search for elements in common across all domains (e.g., sex and age distinctions)
  5. identify organizing domains -- ones that strongly pattern behavior, such as stages in 911 calls, or making a bucket.
  6. make schematic of cultural scene: matching rooms to activities in the booking of tramp sequence, or how central theme relates to aspects of tramp's life.
  7. use generic (etic) keys (often functional): social conflict, inequality, cultural contradictions, techniques of social control, managing interpersonal relations, acquiring status, solving problems, subsistence technology,