Ross Woods, 2020, rev. '21, '22.
This approach to questioning is called free informal interviews or semi-structured interviews. It is an interviewing method that lets new ideas emerge and gives the interviewer freedom to explore them. The purpose of this kind of questioning is to explore subjects' perceptions. It is also very appliable to program evaluations.
The interviewer has a list of themes to be explored or open-ended questions, and creates a discussion with the interviewee. The interviewer does not simply follow a rigid list of set questions to get answers to each question. He/she may add follow-up questions to explore a topic, re-phrase questions, omit questions that appear irrelevant (or already answered in the discussion), and ask questions in an order that best suits the conversation.
This is is not the only kind of questionnaire. Some questionnaires have a list of closed questions, often “Strongly disagree, Disagree, Don’t know, Agree, Strongly agree.” These get a small amount of data, but are easy to use for large numbers of respondents. The kind of research is quantitative and data is expressed in statistics.
Consider these examples:
The goal is to have a natural, flowing discussion where people feel that questions are interesting and exploratory but not too difficult to answer. This comprises at least these elements:
Check that your questions relate directly to your main research question and will generate all the information you need to answer it. This is often easier if you first develop a set of sub-questions based on the main question, making it easier to write another set of questions for your questionnaire. This is part of construct validity
, the idea that data collection methods must result in a data set that suits the research problem and will provide an answer or solution. (Some instructors also call this alignment.)
The number of questions is not all that important; the criterion is whether interviewees find them easy to answer and they get all the information that you need. A small number of very complex questions might be too difficult for people to answer, so more questions would be better. If you have a large number of questions, you might find that interviewees will cover similar questions together as one discussion item.
There are practical limits on how many people you can interview. For example, students doing doctoral research of this kind often interview about fifteen people, although one institution has a minimum of thirty. Some topics are well-suited to less than fifteen.
Here are five kinds of questions that will help improve your interviews:
| Introductions | Commence the meeting with simple getting to know youquestions as an introduction. This normally includes some demographic information. |
|---|---|
| Questions to open new topics | The meeting will normally have a series of topics. Commence each new topic with a question that people feel is easy to answer. These questions have the following characteristics: • They are short and simple. • They ask for factual information. • There are more closed questions than open questions. |
| Main questions | The next kind of question on each topic are the main questions. They are specifically designed to elicit information that will lead to the solution of the research problem. They gather more specific information, personal reflections, analyses, and interpretations. These questions can be more thought-provoking questions but they still need to be clear and as simple as possible. |
| Follow-up questions | You can then ask follow-up questions to explore details. In an oral interview (and to some extent in a focus group) you can generate these questions spontaneously to explore interesting but unexpected issues that come up, or answers that are unclear. You do not normally need to put follow-up questions in your list of questions. It is more important to learn the skill of creating them on the run. Don’t ask questions that are only “What …?” Ask questions like “Why …?” “How …?” “Can you give me an example?” |
| Re-directing questions | Questions for steering the discussion back to the topic if it digresses or becomes unproductive. |
Always edit your questions before you use them in research. Check that the questions are easy to understand, preferably if only heard once in natural conversation. It is usually necessary to re-focus and simplify questions to make them useful. Don't start research before the questions are properly developed; it is better to get the questions right first, so that the first interviews or focus group meetings are not wasted with poor questions. With good editing, you might see faults that were invisible when you first wrote them, for example, questions that are ambiguous, confusing, incline respondents toward one particular answer, or incline them away from other answers. Here are some suggestions:
In you use free informal interviews, give yourself some freedoms so that the questions are not rigid and limited. Put this in the methodology section of your proposal, so that you are quite free to adapt to any unexpected situations. First, allow yourself to use questions in a different order. Interviewees sometimes refer back to previous topics of previous questions, so you can add information to that answer. (“That reminds me. You asked about ….”) Interviewees sometimes refer to things you haven't asked about yet, that come up in a later question. This most often happens with quick thinkers who see an implication to your question.
Second, it is essential to use follow-up questions to explore important issues that arise. Some questions might not need exploration, but if respondents answer all your key questions and there is nothing left to explore, then something has gone seriously wrong and your interview has probably failed. Perhaps your topic and research question might have been wrong, because it was easy to answer with simple information. Alternatively, your questions might have been wrong.
It’s not a problem if respondents give very different answers. Your task is then to find out why they did so.
Later on, as you do interviews and find out more about your topic, your list of questions will probably change. You might improve the questions or add more questions. You will need to report this in your methodology, and give the reasons for any changes.
You can do this kind of phenomenological research by itself. You can also do observations of behavior and compare people’s actions with their perceptions. For example, ten students in a group thought, “That book is very good and I want to buy it.” However, only one of them actually bought the book.