Tag Archives: qualitative interview

Qualitative Interviewing: 4 Reminders about Question Construction

The goal of the interview is to collect whole answers from the interviewee. Question formulation and delivery are critical in this process.

Interviewees respond not only to the kind of question that is context dependent but also to the way the question is asked. The wording, ordering, and kind of language used affect the context (e.g. perception of the interviewee) of the question. Even your tone, enunciation, gestures, and facial expressions as interviewer affect the direction of the interviewee’s answers.

A question could be closed-ended or open. A closed-ended question could be answered by a yes or no, or a list of choices is provided where the interviewee may choose his/her answer. Otherwise the question is open-ended when answers depend on the participant’s own categories and opinions.

It is said that closed-ended questions are hard to construct but easy to use in data gathering while open-ended is the opposite. Actually, both types are hard to formulate since both need rapport and both are prone to errors when not properly prepared.

So when preparing your questionnaire, there are important things to consider in question construction and delivery.

4 Reminders About Question Construction

1. Start with the easiest question.

An interview schedule (whether for structured or semi-structured setting) should always start with the easiest, most comfortable question to establish rapport for a one-shot setting or maintain the rapport for a multilevel setting. Depending on the culture, the interviewer should be cognizant on what subtopic or question would the interviewee consider as the easiest and most comfortable. Usually, questions about basic personal information seem the least threatening and thus, the most uncomplicated.

2. There is proper ordering of questions.

The ordering of the questions affects the entire interview process. Whether you choose the deductive (from general to specific) or the inductive (specific to broad) arrangement, questions are always in a network. That is, the subsequent question should be linked to the one it follows. This is to help the interviewee in organizing the information that she/he is going to share.

Unless inevitably necessary, questions should not be isolated from one another. When you ask questions randomly, as if they are just popping out of nowhere, you will definitely confuse the interviewee.

3. Remember that questions are connected.

Since questions are linked together, a preceding question may significantly affect the interviewee’s answer to the subsequent question. Bradburn, Sudman, and Wansink (2004) cited a study that illustrates how the ordering of questions about advertising affects the women’s answers. Women’s attitudes (answers) when asked about their opinions toward advertising were more positive when questions about dress advertising preceded the general questions about advertising than when it was the other way around.

Given this tendency, you may opt to carefully arrange the questions in such a way that the unnecessary effect of the preceding question will be minimized and will not be carried over to the next. That is why, a pilot interview is always recommended to evaluate your prepared schedule.

lots of questions

4. Learn how to probe.

However efficiently formulated your questions are, always expect that the interviewee’s answers may not always be complete. Reasons could be that the questions are not clear, the interviewee is reluctant to answer or there is problem with the retrieval of the information needed. If this is the case, probe.

If questions are misunderstood or unclear, just rephrase the question and immediately ask again or you may just ask again later. The interviewee may hesitate due to uncertainty of what more information is needed.

If this is the case, ask for more by paraphrasing his answers followed by prompts e.g. “What else?”, “Why?”, “What do you mean?, “In what way?, or “How?”

Prompts also are important to aid recall when interviewee has difficulty in remembering. You may give examples to serve as retrieval cues or to clarify the needed information (Dawson, 2007).

References

Bradburn, N.M., Sudman, S., & Wansink, B. (2004). Asking questions: The definitive guide to questionnaire designs – for market research, political polls, and social and health questionnaires. CA: Josey-Bass: A Wiley Imprint.

Dawson, C. (2007). A practical guide to qualitative research: A user friendly manual in mastering research techniques and projects. Oxford: How To Content.

©2015 February 10 J. G. Pizarro

Qualitative Interviewing: 3 Mistakes to Avoid in Question Formulation

There are common mistakes that are often committed by an interviewer who is new in the field. Even the seasoned ones sometimes inadvertently overlook these errors. To avoid these mistakes, a carefully prepared and tested set of questions is the key.

One classic example given by Bradburn, Sudman, and Wansink (2004) to illustrate the criticality of proper construction of questions is the difference between the two questions namely, “Is it a sin to smoke while praying?”, and “Is it a sin to pray while smoking?”. The inquiry brought about by an argument between two priests is to find whether it is a sin to pray and smoke at the same time.

When both asked each other’s superior, the first question (of the first priest) got a ‘yes’ answer while the latter (question of the second priest) got a ‘no’. The difference in the answers is not due to conflicting opinions but due to the disparity in the context.

In the first question, there is the assumption that the individual is already praying when he/she opted to smoke along with praying. While the second implies the opposite, that is, the individual is already smoking when he opted to pray (maybe to ask for strength to resist the vice). This reminds us that a slight change in the wording of a question changes its meaning and context.

The following are six reminders on the common mistakes to be avoided when preparing and asking questions for interview.

3 Common Mistakes to Avoid in Question Formulation

1. Avoid underestimation, overestimation, and over-assumption.

Underestimation. Use simple words but not too simple. Do not underestimate your interviewee’s capacity to understand. Doing so might offend her. Unless necessary, slang language should be minimized not because it is informal but because some slang words cannot be easily understood. Jargons too should be simplified.

Overestimation. You should not overestimate as well. Do not ask questions beyond the interviewee’s ability to comprehend and also do not over-assume that the interviewee is knowledgeable as this situation is prone to social desirability bias. This bias refers to the interviewee’s motivation to appear “good” like being smart or morally good in front of the interviewer. Thus, even if he does not totally understand the question, or has no knowledge regarding the topic of the question, he might attempt to answer it to appear knowledgeable.

Over-assumption. One example of over-assumption is when you assume that an interviewee who has a driver’s license automatically has experience in driving an SUV. Without the preceding question “Do you happen to have driven an SUV?” the question “How does it feel driving an SUV?” is an over-assumption.

2. Avoid double-barreled questions.

Do not confuse the interviewee by asking double-barreled questions. This may result to vague answers since the interviewee gets confused to the question that contains two (or more) concepts (or objects) that are put together needing two (or more) different opinions but asks only one answer. There are actually two (or more) questions compounded together.

For example, “Are your teachers morally good and kind?” and “Do you support homosexuality and gender equality, or do you support heterosexuality and freedom of religion?”

In the first question, being morally good is different from being kind. The two concepts should be separated to formulate two questions asking for two different answers. The second question contains four different concepts that should ask four different opinions.

3. Avoid leading questions.

Social desirability bias may also be at work if leading questions are delivered. These are questions that influence the direction of the interviewee’s answers either to correspond with what the interviewee thinks as socially desirable answer or as the answer expected of her by the interviewer (Seidman, 2006). These could be in the form of predisposing questions, leading probes, or loaded questions.

Predisposing Questions. There are questions that predispose the interviewee to provide a socially desirable answer. The question “Do you jog?” for example, may seem neutral and not leading at first. But to some interviewees, this may become predisposing since jogging is considered as fashionable and good. The interviewee is prone to provide an answer that appears good. Researchers suggest the use of the word “happen” in “Do you happen to jog?” since it neutralizes the question implying that it is not expectant of a positive answer from the interviewee (Bradburn et al., 2004).

The question “What is the opinion of an honor student like you on cheating?” will likely elicit a socially desirable answer since including the phrase “honor student like you” is leading the interviewee to desire to be seen as good. The question should start with “What do you think is the general opinion of students on cheating?” which could then be followed by “How about your opinion?”.

Leading Probes. Asking leading probes is like subtly shoving the answers into the interviewee’s mouths. For example, the probe question “Are you saying that you are already in love with him?” when you want to clarify what she means by the statement “Well, I think I miss him now”, is leading because it creates an idea that may not be originally present in the interviewee’s mind. The probe should be open-ended like “Why?”, “How?” or “What do you mean by that?”

Loaded questions. These are worded in such a way that they will get answers expected or desired by the interviewer. These questions contain loaded words which could be emotional like “apathetic” and “problematic” or political like “trickery” and “defraud.”

The question “Do you think your teachers are too burdened and apathetic to help you in your academic concerns?” is loaded in such a way that it will likely elicit a biased answer. The use of the loaded words implies that the writer of the question have biases against teachers. It should be rephrased into “Do you think your teachers help you in your academic concerns?”

In his early surveys on workers, Karl Marx asked the question “Does your employer or his representative resort to trickery in order to defraud you of your part of your earnings?” (Bradburn et al., 2004). He was clearly an advocate of the working class given his loaded question leading the interviewees to provide a biased opinion against the capitalist employers. The question should be asked “Does your employer treat you fairly when it comes to your earnings?” to get the real answers of the interviewees.

References

Bradburn, N.M., Sudman, S., & Wansink, B. (2004). Asking questions: The definitive guide to questionnaire designs – for market research, political polls, and social and health questionnaires. CA: Josey-Bass: A Wiley Imprint.

Seidman, I. (2006). Interviewing as qualitative research: A guide for researchers in Education and Social sciences (3rd ed.) New York: Teachers College Press.

© 2015 February 8 J. G. Pizarro

Qualitative Interview Designs

The critical part of a research process is the data collection procedure. Even if you have an apt and interesting topic with an appropriate framework, if the corresponding data collection method is a slapdash, the results would be unreliable and weak. Research is scientific, that is, it should follow a carefully planned methodology.

Interview, for one, is not just going into the field and starting the conversation after obtaining the consent of the participant, and preparing a set of questions and recording tools. It should be designed.

How many levels should it take? How will the interaction flow? What should be the medium of dialogue? The answers should always correspond to the objectives of the research.

Setting Style

One-shot. For baseline exploratory design, this one setting interview would be enough. If the goal is just to provide an initial reference for a specific inquiry, then a follow-up may not be necessary. All the inquiries could be compressed in one short time. Rapport could be built within the first part of the interview. This is ideal for less sensitive and/or simple topics that do not require extensive data.

Multi-tiered. For emergent design and/or in-depth inquiry, a multilevel setting interview is recommended. When topics are more sensitive, more complicated, and more prone to biases and prejudice, a strong rapport is needed to encourage the full disclosure of the participant. This implies longer time for rapport building. You may set the first level for establishing the needed relationship and for preparing the participant for the next level/s. For example, you may start with the easiest and most comfortable topics and may ask the general questions in this level. This deductive arrangement will guide and help the participant retrieve, organize and share the information needed. If the approach is emergent, the next levels will provide opportunity for clarifications and confirmation as a result of preliminary analysis.

Interaction Design

Structured (Narrow Setting). This design is formal and is strictly on the script which could be a highly organized questionnaire set or a standardized interview schedule (Best & Kahn, 1995). The interviewer follows the protocol in uniformity so as to have maximum control of the setting to minimize extraneous variables such as the possibility of researcher influence like gender, age, biases, emotions, preconceptions, assumptions, and the like. Extraneous variables are those which may contaminate the data from the participants. For example, an interview with women participants, if the interviewer is a woman, the interviewee might be more accommodating and open. In another interview of the same topic and participant, if the interviewer is a man, the interviewer’s gender may influence the way the participant provides the answer.

Semi-structured. This design requires a prepared guide question (schedule) that is flexible but retains continuity with spontaneity and fluidity of the conversation process. If the conversation slightly deviates from the topic, the interviewer allows it but tactically returns to the topic to maximize resources. This is usually recommended if the inquiry may allow other important information that may emerge and thus may enrich the data (Dawson, 2007).

Unstructured (Free-flowing Conversation). In Filipino indigenous method, this is known as pakikipagkwentuhan. There is no need for a guide question or list of topics here. The interviewer has only the topic and allows wherever the conversation flows with little directional influence from him/her. The interviewer subtly emphasizes the topic to serve as cue (for the participant) to put into the surface all the elements of the data that the researcher wants to extract from the participant. Interviewer may ask questions sometimes to clarify some information and to retain the conversation process. This design is ideal for in-depth studies like life history and phenomenological research (Marvasti, 2004).

Medium of Dialogue

Face-to-face. This is the most common medium since this allows collection of other details like behavioural and context observation. This is ideal for in-depth study since in this kind of inquiry, trust is critical for rapport.

Phone Interview. This is suggested if distance will not allow a face-to-face interaction and/or the goal is for the participant to answer few simple questions for structured or semi-structured design. This is an alternative to face-to-face interview if the verbal data are the only source for analysis and behavioural and context observation like facial expression and physical setting are no longer necessary for the research.

Online. This could be a video call or internet chat. Chatting is ideal for interviews that allow anonymity of the participants and/or interviewer. This could also be useful in minimizing extraneous variables to reduce bias (Jupp, 2006). While a video call could be a substitute to face-to-face interview.

References:

Best, J.W. & Kahn, J.V. (1995). Research in education. New Delhi: Prentice-Hall of India Private Limited

Dawson, C. (2007). A practical guide to qualitative research: A user friendly manual in mastering research techniques and projects. Oxford: How To Content.

Jupp, V. (2006). The Sage dictionary of Social Science research methods. Thousand Oaks: Sage Publications

Marvasti, A.B. (2004).Qualitative research in Sociology: An introduction. Thousand Oaks: Sage Publications.

© 2015 February 1 J. G. Pizarro