Qualitative research doesn't, by definition, aim to precisely estimate population parameters or test hypotheses. However, most qualitative projects do attempt some kind of generalisation, if not of a numerical nature. Qualitative researchers need to ensure a balance between variety and consistency of responses so they can interpret the results and be sure of their transferability. Qualitative studies commonly use convenience and purposive samples, although probability sampling might be viable in some cases.
Other People Are Reading
Sample sizes in qualitative research are invariably smaller than in quantitative projects. However, the main difference between sampling in qualitative and quantitative research is that qualitative projects rarely use random probability in which each respondent has the same chance of being included. Such samples are common in surveys whose results are generalised for the whole population. It's not usually practical to conduct a qualitative study on a probability sample if the population is large or diverse. When studying small and relatively homogenous populations, such as students of one university or patients suffering from a particular disease in one hospital, a probability sample is possible and can add credibility to the study. In most cases, though, a non-probability sample will be most appropriate.
The easiest way to select respondents for a qualitative project is to use a convenience sample, composed of respondents willing to take part and easy to access. In most cases, the researcher defines at least some essential characteristics of the respondents but the focus is on quick and easy contact. Convenience samples are frowned upon by the methodologists as they are neither theoretically supported nor statistically representative. Still, they are very common in qualitative studies, particularly when the qualitative phase is a prelude or a pilot for a quantitative one. Convenience sampling is particularly useful in commercial research when attempting to find respondents from niche segments.
Most serious qualitative studies use some form of purposive sampling, designed with a particular goal in mind. A common approach to purposive sampling is to maximise variation in an attempt to include as wide a selection of people, objects, communities or situations as possible. One way to achieve that is to stratify the population using other variables in addition to the main criterion. In many consumer studies, all respondents will be users of the products, but they will be stratified by age, sex and social group as well as pattern of use (heavy/light) in order to tap the possible variety of experiences. Minimising variation is sometimes appropriate for group interviews when a researcher plans to have relatively homogenous groups to put the participants at ease and to facilitate group process. In selecting respondents for individual interviews, researchers might attempt to include typical cases, extreme cases or concentrate on key informants who are particularly rich sources of data.
Although qualitative research doesn't aim to estimate quantities in the population, and thus doesn't strive for statistical significance and minimising error, sample size is not a trivial issue. Qualitative researchers need to ensure that they obtain results from a variety of respondents and that all or majority of possible perspectives are covered. A common strategy is to leave a study somewhat open, and continue until no significant new material is gathered. A study by Guest et al. in "Field Methods" proposed that 12 interviews in a homogenous population is often enough to reach saturation, although that depends on how important it is to discover relatively rare perspectives. Peter DePaulo suggested in an article in "QUIRK's Marketing Research Review" that 30 respondents is a good starting point if aiming for range of responses from a fairly varied population.
Sampling only just begins with the sample design: the actual respondents need to be now located and recruited to participate in the research. There are two ways of recruiting respondents for qualitative projects. One is to go out into the world (or send professional recruiters) armed with screening questionnaires and quotas of respondents to find. The other is to use an existing group of potential respondents, called a panel, who all in-principle agreed to participate in research. Sometimes, although there is no panel available, there is a ready sampling frame -- for example, a list of hospital patients with a particular diagnosis or members of an organisation. When the study aims to research a small subgroup of a population and there is no list or panel that contain them, recruiters need to find them using various methods, from classified ads and announcements in the media to screening at the point of purchase. Snowballing is a common technique where one successfully recruited respondent provides leads to others. This is particularly useful when recruiting hard-to-reach respondents, subcultures and socially deviant groups.
- 20 of the funniest online reviews ever
- 14 Biggest lies people tell in online dating sites
- Hilarious things Google thinks you're trying to search for
- "Field Methods"; How Many Interviews Are Enough?: An Experiment with Data Saturation and Variability; Guest et al.; February 2006
- "QUIRK'S Marketing Research Review"; Sample Size for Qualitative Research; Peter DePaulo; December 2000
- Sampling Strategies in Qualitative Research; Antonia D'Onofrio; 2003
- American University; Notes on Qualitative Research; Julie Mertus
- Faculty of Education, University of Plymouth; Research in Education; Qualitative Research; Peter Woods; 2006