In the article, "How Many Interviews Are Enough? An Experiment with Data Saturation," (Field Methods, Vol. 18, No. 1, February 2006) the authors Greg Guest, Arwen Bunce and Laura Johnson investigate this question.
Specifically, these authors were interested in the minimum number of in-depth interviews does it take to get a reliable sense of themes and issues and variability. That is, does it take 6 interviews, 18 interviews, 100 interviews to render a useful understanding of most of the issues; or another way to ask the question is, when does adding more interviews not make a difference in rendering substantial more information? When is enough?
To answer these questions, they conducted a study among a group of women (sex workers) in two African countries. The in-depth interview guide consisted of six structured demographically oriented questions, sixteen open-ended main questions, and fourteen open-ended sub-questions. To determine the degree of data saturation (useful understanding of most themes/issues), the authors used the point in data collection and analysis when new information produced little or no change to the codebook.
After collecting and analyzing their data, data saturation occurred at when they had analyzed 12 interviews.That is, 92% of the total number of codes they developed for the entire study were developed by the 12th interview.
1. Maximum data saturation obtained with minimum number of interviews.
2. Time and cost-effective
1. Must be done with a purposive sample: people specifically interviewed because of their knowledge or experience related to the specific topic.
2. The individuals should be relatively similar (homogeneous), for example female sex workers, street children, etc.