It will be useful to make it clear what we mean by interview by asking some basic questions. It is reasonable that an expert in interviewing would be able to provide sensible answers to the following questions:
Each of these questions will now be considered in a little detail.
First I will clarify the context of the words ‘type’ and ‘method’ used here. The word ‘type’ is intended to indicate that there are several different sorts of interview in respect of the way that interviews are conducted. I have used the word ‘functions’ below to consider interviews that are typically carried out. These functions would include interview for a job for instance. Within this context the word ‘type’ is used to highlight the fact that interviews can be conducted in several different ways. The main types that I will discuss are concerned with interview focus and interview structure. The word ‘methods’ is being used to indicate that an interview can call on several different strategies to help the interviewer collect information from the interviewee. Some of these will involve general questioning and some will involve more formal methods.
A good book covering interviews, even though it is now quite old (1989), is “Knowledge Acquisition for Expert Systems” by Anna Hart. I would still recommend it.
Focus within an interview is to make sure that the interview addresses a specific topic or concept or even a small part of a concept. Many interviews will have a specific focus such as what happened at a particular road accident or how do you change the wheel on a motorcar. Other interviews may not be so concerned with focus, particularly in the early stages of an investigation. Such interviews may be held in order to establish boundaries, what is to be included later and what is not. Non focused interviews are sometimes conducted as brainstorming sessions but here, there is not normally an interviewer and interviewee.
When a strict focus is required for an interview, part of the interviewers job will be to maintain that focus.
Structure in an interview refers to how much control the interviewer intends to have over how a particular interview is carried out. For instance, a completely unstructured interview may almost be like a conversation and the interviewee may end up deciding what to discuss and how things are discussed. A structured interview may mean that the interviewer has carefully prepared a set of questions and knows exactly how these are to be asked and what type of answer is expected for each. In such cases, a good interviewer would be able to conduct many interviews using the same structure or control and ensure that all interviews were similar. Such detailed structure may be required for a well controlled set of research interviews where it is important to remove interview bias as far as possible.
Although structure and focus can sound similar, it would be possible to design interviews that were for instance, focused and unstructured or structured and focused etc. If I wanted to interview someone about their experiences of lightning strikes at sea, I may decide on an unstructured interview because I may not be aware of what experience the interviewee actually has. However, I may want to strictly control the focus so that time is not wasted discussing which ports from around the world have the best bars (or maybe that particular example would not be a waste of time).
This topic is not a reference to the style of the interviewer. For instance whether the interviewer is forceful and authoritative or whether the mood is very relaxed and friendly. This of course will matter within an interview. The topic here however relates more closely to how certain formulations of questions lead to certain types of answers and consequently to the elicitation of different types of information. For instance one could ask an expert what type of computer would be best for a highly mobile executive that needs to regularly exchange information with head office staff. One could instead ask that same expert how can I decide which is the best computer for a highly mobile executive etc. Both questions concern the same thing but because they are actually different questions, the expert will probably answer each in a very different way.
It is also necessary to know what sorts of questions narrow a focus or delve deeper into a topic and which type broaden out a discussion. A question such as, can you give me a specific example of that is likely to focus more finely whilst a question like what sort of situation is this likely to occur in, is probably going to broaden the scope of the discussion.
Asking people to tell a story about a situation or describe specific cases is often likely to provide much more detail than asking the person to perform some operation on their memories such as generalising them etc. The police use a ‘cognitive interview’ for eye witness testimony, which asks people to take several different perspective views of a situation. It is argued that this helps a person to remember more detail than they would recall from a straightforward linear interview.
Human memory has many particular features and human memory is quite different to computer memory. It is interesting that quiz questions encourage people to remember isolated facts whilst research has shown that people are much better at remembering scenes and highly integrated information. The more senses that are involved in a task the better. Human memory is thought to have components for episodes and for things that affect the individual directly etc. A good interviewer should know something about this and be prepared to make use of it. (Alan Baddeley: Your Memory: A User Guide)
There are more formal interview methods that can be used in certain cases. Many of these have been proven within the psychology environment and are aimed at eliciting particular types of information or estimates of values of parameters such as which thing is most useful etc. A technique called ‘card sort’ can be used to help an interviewee to assign values to things or place things in order of something. The ‘repertory grid’ can be used to help assign and identify order where there is uncertainty. In general, decomposition methods can help an interviewee to break information down in order to get at the details required or identify tasks in the correct degree of precision.
A good interviewer should be aware of such methods, know where they may prove beneficial and be able to use them if required to do so.
An impartial observer would probably notice the greatest distinction between interviews in which formal methods are being used and interviews where they were not. However, the distinction between degrees of focus and degrees of structure can lead to quite different results. The type of interview then will have a strong bearing on the information that would be gathered from the interview. Since information gathering is an important feature of interviews it is clear that interview types are really different. This difference means that an interviewer should actively select a type of interview after considering what information is to be gathered from it.
Any sort of interview will have some common features although some interviews may look quite different from each other. Interviews will have a common objective (primary objective), which is to elicit information. The whole point of an interview is the information that is elicited. Such an obvious feature can often be forgotten when people focus on themselves or on how others perform etc. If the objective is how a person performs then why not simply get them to act or carry out a task rather then conduct an interview.
The primary objective is a feature that is common to all interviews and should always be of chief concern. Other factors within an interview are to enable this primary objective to be achieved more effectively or more efficiently. Other features that are common are that there will be at least one interviewer and at least one interviewee. Most interviews will have a topic or theme. This topic may be more or less tightly focused but it will generally be there. So the common features are:
Most businesses are interested in interviews as a way of identifying the best candidate for a new job. Most interviews carried out in business will be for this purpose. In general, candidate selection is possibly the most common function of an interview. Universities also use interviews to select from a large number of candidates for places on particular courses of study.
In more recent years, interviews to capture knowledge have become quite common place. Such knowledge capture may be carried out to support the construction of a computer based Knowledge Based System (KBS). Interviews for knowledge capture or knowledge elicitation will be different from those for candidate selection because a secondary objective is different. The primary objective in each cases is still to elicit information, either about a candidate or about some knowledge area. The secondary objective however will be to either lead to candidate selection, that is, discrimination between candidates, or lead to the documentation of knowledge in a formal way. There may be no discrimination involved here.
The motivation behind this work sprang from a requirement to improve interview performance for the elicitation of Knowledge Structure Maps (KSMs). The details of a knowledge structure map will not be discussed here. The requirement of KSM interviews are to elicit a map or visualisation that represents the structure of a knowledge area based on how experts may acquire that knowledge. Such knowledge acquisition is known as learning dependency because experts are expected to already know some things before they can fully understand others. That is some knowledge is dependent on a prior understanding of other knowledge. The elicitation of this sort of structure and this visualisation is quite demanding but can be done by someone that knows how to interview. However, knowing how to interview is not the same as knowing what questions to ask someone. Interviews for KSM may be similar to those for knowledge capture but are much more precisely structured.
There are other functions of interviews such as eye witness testimony or the interrogation of a spy. There would still be a common theme within such interviews however and that would be the capture of information. Quite how these interviews would be conducted would depend on many factors but someone conducting them would be more effective if they knew how to interview.
The outputs from an interview have already been stated but some clarification may be useful. In section 2.3, the ideas of a primary and secondary objective were raised. These would probably lead to outputs. The primary objective of an interview is to elicit information so the main output from an interview is the information that the interview was designed to elicit. Secondary objectives could also lead to outputs from an interview although these outputs may not be direct; they are more likely to be contributions to outputs. For candidate selection, the output of the process may be the selection of one of several candidates for a job. The interview would provide information that would make a contribution to this process but would probably not be the only thing that was considered.
Even for the elicitation of the information required to construct a Knowledge Structure map, the primary objective and output would be the information required to build the map. There is likely to be a secondary objective here also and therefore a secondary output. This would probably be the answer to or solution to knowledge related issues or problems that lead to the study in the first place.