HOW MANY INTERVIEWS ARE NEEDED IN A QUALITATIVE RESEARCH? IS THERE ANY RULE OR POPULAR PRACTICE?
There are a number of influences on how many interviews a researcher may conduct:
1) The type of sampling techniques that are employed (cf Patton 2002, pp. 232)
2) Resourcing of the study, can place limitations on what sampling is feasible (Kvale 1996; Seidman 2006). If the interviewer can travel only within a limited area, then there is a geographical limitation on the sample. For example, Stoodley (2009), limited his travel to Southeast Queensland due to resourcing (p. 76). Such a geographical limit imposes a cultural limit on a sample.
3) Sampling continues until the researcher senses she has reached saturation. Saturation is a problematic term (Guest et al. 2006; Mason 2010; Morse 1995). Since the first use of the term “theoretical saturation” by Glaser and Strauss (1967, p. 61), the meaning of saturation has become blurred. Glaser and Strauss intertwined data collection and analysis for one category until saturation, before moving on to collect and analyse data for another category. (In grounded theory, a category is a conceptual element of the grounded theory being discovered (p. 36).) The type of saturation the researcher is aiming for may not be theoretical saturation. “Saturation of knowledge” (Bertaux 1981, p. 37) is a better term. Bertaux describes how the researcher is surprised or learns a great deal from the first few interviews. By (say) the fifteenth interview, the researcher recognises patterns in the interviewees’ experiences. More interviews confirm what the researcher has already sensed.
How saturation of knowledge is reached or passed during sampling is uncertain. According to Mason (2010), it is more likely PhD students using qualitative interviews will stop sampling when the number of samples is a multiple of ten rather than when saturation has occurred. Guest et al. (2006) found that 12 interviews of a homogenous group is all that is needed to reach saturation. Conceptually, saturation may be the desired end point of data collection. Operationally, the decision to stop interviewing is a function of a combination of all or some of the following factors:
• interview structure and content (Guest et al. 2006); the more unstructured and variable the content, the more interviews are required
• heterogeneity of the group (Guest et al. 2006); the more heterogeneous, the more interviews are required
• the number of interviews done already (Ryan and Bernard 2006); the weaker the sense there are enough interviews, the more interviews are required
• the complexity of the interviews (Ryan and Bernard 2006); the greater the complexity the more interviews are required
• the researcher’s experience, fatigue (Ryan and Bernard 2006), and confidence (Mason 2010)
• the number of researchers in the research team (Ryan and Bernard 2006)
• the more interviews, the more defensible the researcher believes the research will be (Mason 2010)
• doing what was stated in a research proposal (Mason 2010)
• the nature of the sample being limited by the sampling technique (Browne and Russell 2003)
• resourcing (Kvale 1996; Seidman 2006)
• the orthodoxy of the method: the number of interviews is expected to fall within a certain range
• meeting all of the purposeful sampling criteria the researcher has determined that are necessary for the study.