Electronic Textbook: "Ask and Ye Shall Receive:" An Introduction to Interviewing

"Ask and Ye Shall Receive:"
An Introduction to Interviewing

Last time out, friends, we talked about a popular channel of qualitative data collection: participant observation. In the next series of lesson packets, we'll zero in on perhaps one of the most widely used procedures: interviewing.

This time around, we'll briefly take a look at some relative strengths of interviewing, as compared with participant observation. We'll also take a look at the dimensions along which interviewing itself may vary. That is: as with Michael Quinn Patton's "five dimensions of participant observation," you may be a bit surprised at the "leeway" you have with regard to structuring interviewing. Finally, we'll end with Michael Quinn Patton's excellent "matrix of interview question possibilities."

To get you feet wet, open the link below and read this overview of the interview process in research.

Interviews - Bill Trochim, Cornell

I. Relative Strengths of Interviewing

As we saw in the preceding topic, participant observation does have a number of things going for it, in terms of being a desirable data collection procedure. For one thing, an in-depth immersion is possible. This in turn implies a larger number of sources of information - many things being observed, all of which can serve to "inform" the research question being addressed. The researcher's perspective, in addition to those of others, can serve as a cross-check on the 'reality' of that which is being observed.

On the other hand, as Michael Quinn Patton has pointed out, other key sources of information "elude outside observation!" How are we to directly "see" and "observe" important variables such as the following:

* Subjects' feelings, thoughts and intentions;
* Behaviors from an earlier point in time (i.e., have already taken place) which influence those feelings, thoughts and intentions;
* Subjects' general thinking styles, frames of reference, how they organize their worlds, and so forth

*** These CANNOT BE DIRECTLY OBSERVED! The best way to gather information on them is to ASK!

*** As Michael Quinn Patton has stated, "The purpose of interviewing...is to allow us to enter the other person's perspective."

Some may argue that "perceptions do not necessarily match reality." True, but when you think about it, the same "threat to validity" is also present to a degree in participant observation! Certainly a situation can be 'rigged,' however subtly, when it is known that there is an observer!(Ah, the bane of 'unannounced evaluation observations' to every classroom teacher ..!) Whether conscious or subconscious, subjects may present themselves differently to an outside observer than they would "be" in "everyday life" in that setting. This includes halo effects, Hawthorne effects, and reactive effects in general.

We would confront such potential incongruities or biases the same way we do in the more general sense of validity and reliability cross-checks: e.g., multiple observers, asking the same interview questions several different ways, or better yet, building a multimethod design that incorporates multiple qualitative data collection procedures (e.g., participant observation balanced with interviewing) or quantitative and qualitative procedures (e.g., adding in surveys, archival numerical data, etc.).

To remind our friends from Intro to Research and Research Design, it is recognized that no single method of design and analysis is without potential biases or flaws. Thus, the ideal in multimethod design and analysis procedures is to "counterbalance" those biases, by using combinations of two or more methods, to see if the results "converge" or point in the same direction regarding the findings. This gives us a greater degree of confidence that those findings are "valid," i.e., "we're picking up something real," as opposed to a bias such as a please-the-researcher halo effect!

II. Dimensions of Interviewing

Michael Quinn Patton has also proposed the following excellent dimensions of interviewing. My own 'take' on these is that in general, the tradeoff is as follows. I've shown Patton's four major types of interviewing as "starred elements" along the continuum. Below these, I show the "relative pluses and minuses" of each extreme.

Figure 1.
The Basic Continuum of
Interviewing Procedures

 

 

 

 

In the preceding diagram, then, we can see that, of the four methods identified by Patton, the major element of variation is structure. Please note, too, that there are definitely tradeoffs to be had at either extreme.

Such structure provides greater assurance that 'each interviewer will ask the same questions in the same way(s).' Thus, the possibility for interviewer bias, error, inexperience, etc., is reduced, which in turn helps improve the reliability of the results.

Stop a minute and open the link below to get more familiar with the problem of bias in the research interview.

Qualitative Methods Workbook - Scroll down to Chapter 15

On the other hand, with greater predictability (in process, number and types of questions, etc.) comes a concomitant loss of spontaneity. We've learned that, particularly when it comes to very preliminary, exploratory, grounded-theory work, such unanticipated (by the 'external' interviewer) willingness to 'go with the flow' can lead to very fruitful directions regarding "lived experience" as the subjects themselves perceive them.

And the tradeoff is in exactly the opposite direction for the other (right-hand side) extreme!

Too much spontaneity, particularly in the hands of relatively unskilled, inexperienced interviewers, leaves open the risk of interviewer error, subjects 'rambling, not covering areas which are "ex ante" important to address in an (albeit open-ended) questioning fashion. Thus, the reliability of the study may suffer gravely in the event of a "too loose and open" approach. This is particularly the case with less exploratory, more 'confirmatory' types of investigations. Some of the key variables,determinants, factors, outcomes, etc., are already 'known' to the researcher. This could be from prior work in the area. Thus, it is important that those "known variables" be "covered" in as objective and scientific a manner as possible. Their omission cannot be left to chance, as it might in a too-spontaneous interview session.

Finally, before proceeding into Patton's excellent in-depth chart of these interview possibilities, I do want to commend him for even including that 4th one - on the left-hand side of the continuum above, called "closed quant." This can cause confusion on the part of researcher and subject alike!

Say you are an interview subject in such a situation. The researcher is asking you a series of questions to which you are certainly replying "in words." They may range from demographic; i.e., "What is your primary ethnicity?" to seemingly even more open-ended ones: "What is the primary reason that you chose to enter the teaching profession?"

Now, you may indeed be responding qualitatively - giving verbal responses. For the second example, above, in fact, you may choose to give a lengthy verbal explanation of your primary motivation in choosing teaching as your profession - complete with richness and depth of feelings and emotions!

However, if the interviewer is recording your answers by checking off a choice on his/her corresponding interview form - such as a category for "Ethnicity" or even a category that in the interviewer's expert opinion "most closely matches" the lengthy verbal answer you've just given for "teaching motivation" - then this so-called "interview" is really a quantitative data collection procedure! For it is possible that all the 'words' will be tallied up and quantified only: e.g., "Total number and percent of respondents who are Navajo;" "Total number and percent of respondents who chose teaching due to 'wanting to make a difference in students' lives (if this was a choice on the interviewer's response sheet as he/she listened to and classified your answer)," etc.

That, then, is the nature of the "closed" in the label, "closed quantitative interview." Regardless of how you choose to verbalize your responses, they will all eventually be forced into such predetermined, 'closed' categories and tallied up for the reporting and analysis. This is what makes this "interview" in actuality a quantitative, not qualitative, procedure!

Now -- let's revisit those four types (3 qualitative, 1 quantitative) of interviewing in Michael Quinn Patton's own words! How about it?

 

 

Table 1.
Variations in Interviewing Procedures
adapted from: Michael Quinn Patton,
How to Use Qualitative Methods in Evaluation
(1987: Sage Publications, Inc.)

Type of Interview	Characteristics	Relative Strengths	Relative Weaknesses
Informal conversational interview	Questions emerge from the immediate context & are asked in the natural order of things; there is no predetermination of question topics or wording.	Increases salience & relevance of questions; interviews are built on & emerge from observations; the interview can be matched to individuals & circumstances.	Different information collected from different subjects with different questions.Less systematic & comprehensive if certain questions do not arise "naturally." Data organization & analysis can be quite difficult.
Interview guide approach	Topics & issues to be covered are specified in advance, in outline form; interviewer decides sequence & wording of questions in the course of the interview.	Outline increases comprehensiveness of data & makes data collection somewhat sys-tematic for each respondent. Logical gaps in data can be anticipated & closed. Interviews remain fairly conversational & situational.	Important & salient topics may still be inadvertently omitted by the interviewer. Interviewer's flexibility in sequencing & wording questions can result in sub- stantially different responses, thus reducing the comparability of those responses across different subjects.
Standardized open-ended interview	The exact wording & sequence of questions are determined in advance. All interviewees are asked the same questions in the same order.	Respondents answer the same questions thus increasing compara- bility of responses; data are complete for each person on the topics ad-dressed in the interview. Reduces interviewer effects & bias in the case of multiple interviewers. Permits decision makers to see & review the "instru- mentation" (i.e., the set of common questions) in advance of the interviews. Standard format also facilitates subsequent organization & analysis of the interview data.	Little flexibility in adjusting the interview to individuals & unique circumstances; standardized wording of questions may inhibit, constrain & limit the "naturalness" of interview questions & answers (i.e., interaction between interviewer & subject).
Closed quantitative interview	Questions & response categories are determined in advance. Res- ponses are fixed; they are 'slotted' into these predetermined fixed categories for compilation & analysis.	Data compilation & analysis is relatively straight-forward; responses can be readily aggregated & also easily compared across individuals & sub- groups; many questions can be asked in a relatively shorter time with the fixed format.	Respondents' verbalized perceptions, actual experiences, feelings, etc. must be made to 'fit' into those predetermined categories; may be perceived as im- personal, irrelevant, & somewhat mechanistic. Can distort what respondents "really mean" or have actually experienced by having the limited, fixed response choices.

As you scan the preceding comparative rows of the matrix of possibilities, you've probably already realized that it is possible to have a single interview session which is a "hybrid," or a mixture of the different methods, for different parts of the interview.

The structure can vary in either direction. For instance, the interviewer may choose to start the session with 'neutral, structured, warm-up' demographic questions. These responses may be recorded as closed quantitative items. Then, as the interview progresses, the interviewer may "loosen up" with gradually more open-ended questions, and perhaps close by totally following the direction of the subject's responses -- e.g., spontaneous follow-up prompts as opposed to prepared questions at that point.

The opposite direction is also possible. The interviewer may opt to "set a blank stage" in the beginning, then gradually focus the questions, and finally bring closure to the session with a standardized set of (quantifiable, closed-ended) demographic or other types of questions.

As can be seen from the preceding Part III discussion, the "happy medium" might be captured by the following two procedures:

1. Standardized open-ended interview; and
2. Interview guide approach.

There is an excellent source on standardized open-ended interviewing that I'd like to recommend to you. It is by Floyd J. Fowler, Jr. (a noted survey design expert, by the way, who's based at the University of Massachusetts, Boston) and Thomas W. Mangione. It is entitled Standardized Survey Interviewing: Minimizing Interviewer-Related Error and was published in 1990 by Sage Publications, Inc. in Newbury Park, California.

We'll be spending more time in future modules on the second alternative, the interview guide approach. As we have seen in Table 1, above, this one attempts to balance the open-ended, 'go with the flow' if warranted nature of qualitative research with some preplanned directions (e.g., basic question topics) to enhance reliability and comparability. For now, I'd like to mention that other terms for "interview guide" that are popularly used include "questioning route" and "interview protocol." We'll be using these interchangeably.

Now -- as we head into this area, let's close out our introduction to interviewing by taking a look at Michael Quinn Patton's excellent "matrix of question options." This will get you started thinking about the basic categories or types of questions that you might ask in an interview. These will comprise the rows of this matrix. The columns, on the other hand, represent the time dimensions within which you can plan and structure your questions. Not all questions need be about the present - nor should they necessarily be! Remember, qualitative is about 'rich, thick description!' Not seemingly precise Type I error rates, p-values, etc.! Thus, we might need for subjects to "guess-timate" about the future - and likewise benefit from their "best recall" of the past!

Let's begin by sketching out this matrix and then also briefly discuss each of the types (rows). This matrix appears on the following page. Many researchers have found it helpful to "fill in" such a matrix when they are initially drafting and revising their tentative interview questions.

 

Table 2.
Matrix of Question Options
adapted from: Michael Quinn Patton,
How to Use Qualitative Methods in Evaluation
(1987: Sage Publications, Inc.)

Category/Type of Question (Rows)/ Time Dimension for Framing the Question (Columns)	Past	Present	Future
Behavior/ experience questions	Fill in your actual planned questions	Fill in your actual planned questions	Fill in your actual planned questions
Opinion/ value questions	Fill in your actual planned questions	Fill in your actual planned questions	Fill in your actual planned questions
Feeling questions	Fill in your actual planned questions	Fill in your actual planned questions	Fill in your actual planned questions
Knowledge questions	Fill in your actual planned questions	Fill in your actual planned questions	Fill in your actual planned questions
Sensory questions	Fill in your actual planned questions	Fill in your actual planned questions	Fill in your actual planned questions
Demographic/ background questions	Fill in your actual planned questions	Fill in your actual planned questions	Fill in your actual planned questions

 

Now, let's step through the different types/categories (rows) to give you a feel for what sorts of questions you might be filling in for each of these!

Experience/ Behavior Questions:	These deal with subjects' actions - be they past, present or future. Through such questions, you'd be getting the subjects to describe activities, decisions, behaviors, etc., that would actually be observable. In evaluation research, for example, these might relate to a program. "If I followed you through a typical day [of Program/Activity X], what would I have seen you doing?" These types of questions, by the way, are sometimes called "simulation" questions. The idea is to simulate the types, nature, sequence, etc. of actual behaviors or activities.
Opinion/ Belief Questions	These questions are aimed at understanding subjects' 'world-views' of things, as alluded to on pg. 1 of this lesson packet. How do they cognitively structure their reality? Anytime you see "keywords" such as the following, you can be sure you have an opinion/belief question: "What is your opinion of [...]?" "What do you think about [...]?" "What do you believe about [...]?" As a cautionary note, these are sometimes confused with the next two categories: feeling and knowledge questions.
Feeling Questions:	Unlike beliefs, which deal with "cognitive subjectivity," feeling questions deal with "affective subjectivity." Here you are tapping into subjects' emotional responses - i.e., feelings of happiness, fear, anxiety, confidence, and the like.
Knowledge Questions:	While feelings and beliefs are subjective, knowledge questions deal with subjects' factual information. What things does the respondent know about [...]? Patton illustrates this for an evaluation example. Knowledge of a social program may consist of: what services are available, who is eligible, how long people spend in the program, rules and regulations, enrollment procedures, and the like.
Sensory Questions:	These are just what they sound like! They deal with what is heard, touched, seen, tasted, or smelled. Example: "When you walk through the doors of your mother's house, what do you see?" "What do you hear the counselor saying to you at the beginning of each session?"
Background/ Demographic Questions:	These are the standard items that describe subjects' identifying characteristics. They may include age, educational level, annual income, time enrolled in the current program, and place of residence. At first glance, they may appear to also be knowledge (factual) questions. In a sense, this is true: but they relate to a particular subtype of "facts:" those pertaining directly to the subject. Patton characterizes them as "more routine in nature" as a result.

The time dimension (columns of Patton's Table 2, above) may seem comparatively self-explanatory. For instance, for a behavior question, the interviewer may ask the subject: what he/she did yesterday; is doing today; or plans to be doing tomorrow.

However, I would like to introduce to you a special label for "past" questions, because it is so popularly used in focus group interviewing, in particular (the topic of our next lesson packet). This is called the "think back" question. By asking the subject to "think back" to some past experience or event, the interviewer is mentally taking the subject back into that past reality. He/she could presumably 'flesh out the details' of that past by asking virtually any type of question in that context.

"Think back to your first day as a participant in [Program X.] What did you do on that first morning?" (Experience/Behavior Question)

"... and what did you think of the announced objectives of the program?" (Opinion/Value Question)

"... and what were your feelings as you were asked to reflect upon the conclusion of the warmup activity?" (Feeling Question)

" ... and what was the first thing the group facilitator said to you?" (Sensory Question)

"...and what were the stated objectives of that program as announced by the keynote speaker?" (Knowledge Question)

"...and how long had you been employed as an educator in that school district at the time that you enrolled in [Program X]?" (Background/Demographic Question)

 

Not all of these categories of questions, nor the time period (past, present, or future), in Patton's grid, would necessarily apply equally, or even be relevant, to a particular interview context. Rather, this grid is meant as an "imagination/possibility-stretching guide," to get you to thinking about as many different types of potentially useful interview questions as possible.

Next time, we'll take an in-depth look at a particularly popular and "information-rich" type of interview setting: the focus group.

Till then, please remember - it's OK to have a little fun with the process ... !

Go back to Introduction to Interviewing

E-mail M. Dereshiwsky at statcatmd@aol.com
Call M. Dereshiwsky at (520) 523-1892

Copyright © 1999 Northern Arizona University
ALL RIGHTS RESERVED