How do you feel most comfortable speaking in front of an audience? Do you have the confidence to let your ideas flow freely, or do you prefer sticking to a plan? The Colorado Association of Libraries Conference (CALCON 2022) begins today, and there happens to be quite a connection between the styles in which we choose to present our ideas, such as at CALCON and how we conduct the next data collection method we are discussing–interviews. It’s a fitting time to take a closer look at interviews as another viable data collection method for evaluation in libraries since interviews fall into three categories (unstructured, semi-structured or structured), and much like understanding the most effective way to present information, you must understand how to structure, or not structure, your interview for it to produce quality data. Today, we will discuss the pros and cons of these three interview types. If you’re looking for some background reading, be sure to check out this post for an overview of interviews and when to use them. 

Unstructured and Ad-libbed

The first type of interview we’re going to cover, an unstructured interview, is the most conversational. An unstructured interview may be the most casual interview type, but the openness of what is discussed can lead to deep conversations and highly valuable qualitative data. In an unstructured interview you do not have set questions that you plan to ask your interviewee and instead, head into the conversation bringing only a topic that you would like to explore. This is similar to a presentation structured as an open discussion with the audience where no script exists. 

In order for an unstructured interview to provide you with quality data, generally two conditions need to be met. First, you must understand and work within the limitations unstructured interviews have. Data collected from unstructured interviews will not be easily replicable, so this is not the interview type to use if you are hoping to compare data across populations or through time. Unstructured interviews still produce valuable insights, but because each interview might cover slightly different material, the information is not consistent enough to generalize to a large population. This interview type is more effective for localized research and learning the interviewee’s story. 

The second condition is that it’s best to have had some previous experience as an interviewer. We don’t all jump into projects as seasoned researchers and of course, that is OK. However, how you word questions can have a large impact on the quality of the data you collect. As we’ve discussed previously, asking leading questions is going to give you skewed answers. You aren’t equipped with a full interview script during an unstructured interview, and you may find it surprisingly difficult to construct quality follow-up questions on the spot. The same applies to presentations: if you are not an experienced presenter, improvising on stage can be risky, but creating an outline and practicing your notes will set you up for success.  

The Semi-Structured Style

This leads us to the next way to conduct an interview, which is the semi-structured approach. This is the middle-ground, so conducting this type of interview can give you the best of both worlds. I would venture to say that the comparable presentation method, speaking freely while also following your notes, is the method preferred by many. An interview where you can stray from the script but do not scrap it all together allows you to ask follow up questions of your interviewee and explore interesting points further while also maintaining some consistency between interviews. Like unstructured interviews, semi-structured interviews should be conducted if you are still exploring your topic because your interviewee may be able to provide critical information you did not initially know to ask for directly. 

However, there are definitely cases where unstructured or structured interviews will be better options for you. The qualitative analysis of data from a semi-structured interview will still be more complex than that of a structured interview. On the other hand, if you are just beginning to explore a possible topic for evaluation and are interviewing an expert on the topic, allowing your interviewee to speak freely in a completely unstructured interview will likely be most beneficial.  

Structured or Scripted

Before a structured interview you plan every question you will ask and the specific order you will ask them. This is akin to writing a script and following it word for word during a presentation. By doing this during an interview, you will gather more consistent data that is easier to analyze. Because you are trying to work through a set number of questions, this interview type often contains less complex questions and therefore, receives shorter answers. Analyzing the data produced by a set of structured interviews is easier than a set of unstructured interviews. The study will also be more easily replicable. 

If you are conducting structured interviews and hoping to generalize your findings to a larger population, it is important to have a significant sample size. Luckily this is easier to complete with structured interviews than unstructured interviews because you are not asking follow up questions and will have a better idea of how long each interview will take. Consistency is key to avoiding bias, so you should not only try to keep your questions and question order the same but also the manner in which you ask each question.

One important factor to consider before conducting structured interviews is that your interviewees will not have as much opportunity to share the breadth of their experience. You must carefully develop your questions ahead of time to ensure that what you are asking will produce the data you are looking for. Unlike unstructured interviews, which usually take place to learn more about a topic, you should already have a solid understanding of the topic in order to know which questions you should ask.

Time for Questions

We hope this post has expanded your understanding of the possibilities and limitations for conducting interviews. Remember, if you still have questions, please reach out to LRS@LRS.ORG. In the meantime, here are five key takeaways from this post:

  1. You should carefully consider which interview type will best meet the needs of your evaluation as they all come with different benefits and challenges.
  2. It is best to have an experienced interviewer conduct unstructured interviews since they must come up with quality questions on the spot. 
  3. Unstructured or semi-structured interviews should be conducted if you are still exploring your topic and want the ability to ask follow up questions in response to what your interviewee shares. 
  4. Unstructured interviews are the least replicable and structured interviews will be the most replicable for future studies.
  5. You should stay consistent during structured interviews by sticking to your predetermined questions as well as the predetermined order of those questions.

LRS’s Between a Graph and a Hard Place blog series provides instruction on how to evaluate in a library context. Each post covers an aspect of evaluating. To receive posts via email, please complete this form.