List

I was recently asked about interviews and how many should someone do. This is a tricky, tricky question because qualitative research is messy and the beauty (and the curse) of qualitative research is in its ability to adapt it’s methodological and methods based on the research study design.

In other words, there are no guidelines that I can point anyone too about the “correct”number of interviews. More is not necessarily better, but I could also make the argument you need to do more than 1-3.

So let’s try to parse this question through.

Interviews are a common qualitative method and can best be defined as a one-on-one conversation where the researcher poses questions (either in person or asynchronously) and the participants provide answers. Interviews are usually structured (as set of pre-set questions that are used for all interviews with no variations or follow-ups) or semi-structured (some pre-set questions but open ended for follow-up questions based on the initial answers) or open ended (a free flowing conversation that starts from a conversation starter with no pre-set questions). In technical and professional communication (and the rhetoric of health and medicine), most interviews are semi-structured.

Using the interview as your method enables you to gain insights into the participants experiences, thoughts, feelings and perceptions. This is why it’s a great qualitative method because often researchers are wanting to know more about the experiences of participants.

The problem with interviews are that they can be time consuming to do from setting and completing the interview, but more so because of the need to transcribe and then work with and code the interview data. Interviews results in rich data sets (rich defined here as depth of information) but that richness requires a great deal of time to work it.

This leads to the perennial question of how many interviews are enough. In technical and professional communication and the rhetoric of health and medicine, the number of interviews is not normally justified or discussed in the published literature. Indeed, its typical to throw out the number done and move one, which is highly problematic. My guess this is because most folks don’t fully understand that the number of interviews is a key part of the methodology and/or the number chosen was for practical reasons (e.g., time factors) rather than research reasons.

Some of my recent work specifically examined the number of participants in empirical research where the interview was the primary method. The range was from 1-147. (yes, 147 but they seemed to be structured interviews and it was definitely an outlier.) The average was 15. But as we point out in this piece, no one adequately discussed why and how they chose this number and no one discussed data saturation.

As with all things related to research there are lots of discussions around even the term saturation, but I
tend to fall in the camp of liking the term because it helps readers of reward understand how the data became results, which is so important to the trustworthiness and validity of qualitative research.

So let’s agree for the moment that saturation is an ok term that we need to try and achieve. There are different types of saturation:

  • theoretical saturation
  • thematic saturation
  • data saturation

all of these are determined directly through the data and analysis (Ala the interactive nature of grounded theory, which many claim introduced the term saturation) or they are determined more directly by the researcher’s perception of the data. While some could argue that this is the same thing, there are discussions in the literature across fields about how these things can be different. The bottom line is there are different types of saturation based on your methodological orientation to the data and your research.

The number of interviews is going to typically be tied to your methodological orientation and your research study design. So for example, an ethnographic study will likely require more interviews than say a phenomenological study.

In some ways, you can never reach saturation because no two person’s experiences are exactly the same. But saturation does help define boundaries so that you’re not researching forever and once you decided (as part of the larger research study design and methodological orientation) about the type of saturation it becomes easier to get a sense of the number of participants that you’ll need to talk to.

For much of the qualitative work that I have done as an academic and practitioner, I’ve never needed to do more than around 20 or so interviews. More than that often results in the same big ideas (or what many folks call themes or codes) being reinforced and then a myriad of smaller ideas coming out. The latter of which can confuse and slow down research (particularly in the workplace where you need quick turnaround keyed to a specific problem).

It is important with interview data to work with the data as you go so you have an idea of what you’re finding. You can’t leave it to one big analysis session. This getting to know your data as you go helps with the question being asked here about how many is enough.

I’ll leave you with just one thought: trust yourself. If you’ve done the hard work of setting up a strong study through methodology, methods, and practices, then you’ll likely know when you’ve done enough.

 

Sources to look at for saturation

Bowen, G. a., 2008. Naturalistic inquiry and the saturation concept: a research note. Qualitative Research, 8(1), pp.137–152.

Damschroder, L.J., Pritts, J.L., Neblo, M.A., Kalarickal, R.J., Creswell, J.W., Hayward, R.A.: Patients, privacy and trust: patients’ willingness to allow researchers to access their medical records. Soc. Sci. Med. 64(1), 223–235 (2007)

Francis, J.J., Johnston, M., Robertson, C., Glidewell, L., Entwhistle, V., Eccles, M.P., Grimshaw, J.M.: What is an adequate sample size? Operationalising data saturation for theory-driven interview studies. Psychol. Health 25(10), 1229–1245 (2010)

Fusch, P.I., Ness, L.R.: Are we there yet? Data saturation in qualitative research. Qual. Rep. 20(9), 1408–1416 (2015)

Guest, G., 2006. How Many Interviews Are Enough?: An Experiment with Data Saturation and Variability. Field Methods, 18(1), pp.59–82.

Hancock, M.E., Amankwaa, L., Revell, M.A., Mueller, D.: Focus group data saturation: a new approach to data analysis. Qual. Rep. 21(11), 2124–2130 (2016)

Hennink, M.M., Kaiser, B.N., Marconi, V.C.: Code saturation versus meaning saturation: how many interviews are enough? Qual. Health Res. 27(4), 591–608 (2017)

Morse, J.M.: Data were saturated…. Qual. Health Res. 25(5), 587–588 (2015)

Morse, J.M., Barrett, M., Mayan, M., Olson, K., Spiers, J.: Verification strategies for establishing reliability and validity in qualitative research. Int. J. Qual. Methods 1(2), 1–19 (2002)

O’Reilly, M. and Parker, N., 2012. ‘Unsatisfactory Saturation’: a critical exploration of the notion of saturated sample sizes in qualitative research. Qualitative Research, [online] 13(2), pp.190–197. Available at: <http://qrj.sagepub.com/cgi/doi/10.1177/1468794112446106> [Accessed 21 Jul. 2014].

Wray, N., Markovic, M. and Manderson, L., 2007. ‘Researcher saturation’: the impact of data triangulation and intensive-research practices on the researcher and qualitative research process. Qualitative health research, 17(10), pp.1392–1402.

Source for interviews

Nigel King, Christine Horrocks, Joanna Brooks, interviews in qualitative research, Sage.

Print Friendly, PDF & Email