Especially when using a grounded theory approach, seldom can this question be easily answered. In this research note, which reviews grounded theory studies that used interviewing as the primary data collection method, the author finds that typically 10 – 30 interviews are the normative range, although overall the range is from a low of 4 to more than 100 interviews.
The number of interviews depends on:
1) the research question: when the question is more focused fewer interviews are usually necessary
2) the researcher’s familiarity with the topic: increased knowledge may mean fewer interviews
3) single or multiple interviews: multiple interviews with a respondent may mean fewer interviews are necessary
But, at the end of the day, the primary decision factor is still theoretical saturation, which can only be assessed once data collection and analysis has begun.
3 thoughts on “How many interviews is enough?”
My questions is, if participants who fall into the inconsistent or contradictory categories are interviewed (Merriman, 1988, Why Triangulate?) should a control sample of participants who were in the consistent category also be interviewed?
Sorry, I meant Mathison 1988 Why Triangulate? NOT Merriman
Melinda… triangulation as described in my 1988 article doesn’t imply that more data will enhance validity, but rather that the researcher needs to exercise more thoughtfulness when the results obtained are consistent or contradictory (or for that matter when there is high degrees of agreement). Agreement, consistency, contradiction are not categories of people, but rather descriptions of how findings from various sources or data collection strategies relate to one another. So, it isn’t a sampling problem it is an interpretation problem. That said, one always wants to consider whether the data record is indeed adequate… a judgement call to be sure.