Manzoor - I don't view it as a misnomer myself. You appear to be forwarding a 'critique' of qualitative sampling based on the fact that it 'differs' from quantitative sampling - and it's general principles of larger pre-determined size and potential for generalisability. Instead, we could turn that on its head i.e. if we start with qualitative non-probability sampling, then quantitative approaches often lack the flexibility required of qualitative sampling. Less 'confrontational', we could identify that quantitative approaches can be more inductive and some qualitative sampling approaches more deductive i.e. theoretical, quota, maximum variation etc. Qualitative sampling techniques generally state a potential range for size - but so do quantitative approaches.

You state a few things that can be questioned - especially in your latter sentences. Firstly, the sample is not planned during data collection. It may change during data collection - but only when it is clearly stated, before data collection, that further data may be sought if 'saturation' is not met with the initial sample. By 'representative', I assume that you mean 'generalisable'? Generalisability (statistical) is not the intention of qualitative research. Social contexts, however, can be generalised i.e. those suffering loss will generally be sad regardless of the sample characteristics. Measurement scales are not used in qualitative research - nominal or otherwise. Initially, you also state about 'wide heterogeneity' in qualitative samples. This is often not the case. In studies with multiple and specific inclusion/exclusion criteria - the sample becomes 'contained'.

Manzoor - I don't view it as a misnomer myself. You appear to be forwarding a 'critique' of qualitative sampling based on the fact that it 'differs' from quantitative sampling - and it's general principles of larger pre-determined size and potential for generalisability. Instead, we could turn that on its head i.e. if we start with qualitative non-probability sampling, then quantitative approaches often lack the flexibility required of qualitative sampling. Less 'confrontational', we could identify that quantitative approaches can be more inductive and some qualitative sampling approaches more deductive i.e. theoretical, quota, maximum variation etc. Qualitative sampling techniques generally state a potential range for size - but so do quantitative approaches.

You state a few things that can be questioned - especially in your latter sentences. Firstly, the sample is not planned during data collection. It may change during data collection - but only when it is clearly stated, before data collection, that further data may be sought if 'saturation' is not met with the initial sample. By 'representative', I assume that you mean 'generalisable'? Generalisability (statistical) is not the intention of qualitative research. Social contexts, however, can be generalised i.e. those suffering loss will generally be sad regardless of the sample characteristics. Measurement scales are not used in qualitative research - nominal or otherwise. Initially, you also state about 'wide heterogeneity' in qualitative samples. This is often not the case. In studies with multiple and specific inclusion/exclusion criteria - the sample becomes 'contained'.

Definitely not.

Matt

Depends on how large the sample is.

I agree with Dean Whitehead that it appears to be forwarding a 'critique' of qualitative sampling based on the fact that it 'differs' from quantitative sampling - and it's general principles of larger pre-determined size and potential for generalizability but my point is that by misnomer in qualitative research it does not fulfill the essential characteristics of sampling and sample size. Here are some points for consideration:

1). There is no specific formula for calculating sample size in qualitative approach. Although sample size does matter in qualitative research and Sandelowski (1995) proposes this principle: "an adequate sample size in qualitative research is one that permits the deep, case-oriented analysis that is a hallmark of all qualitative inquiry, and that results in a new and richly textured understanding of experience" yet the problem lies in generalizing the results based on a fewer number of cases as compared to quantitative research.

2). There are no specific rules. Various methods used in qualitative research and the the number of cases/interviews to be conducted vary significantly and also there is principle of saturation and adding more participants to the study does not result in additional perspectives or information.

The small sample sizes generally associated with qualitative research will not produce statistically generalizable results, and the use of random sampling may lower the quality of the sample that is chosen by failing to represent groups that are relatively rare. I think some qualitative researchers may try to copy statistical approaches to sampling, when the real value of qualitative research is from the in-depth insights that can be produced, not whether the sample is representative..

Having said that, a well-drawn sample can be very helpful in providing multiple perspectives, so the drawing of a sample is very important even if it is purposive rather than a statistical sample. Designing the sample to provide an appropriate level of diversity can be quite valuable.

مرحبا عزيزي

بالطبع كلا

The only reason it is a “misnomer“ is because so many people limit their understanding of sampling to quantitative versions. Random selection of participants is just one of many options, and I agree with Bradford Cheney that it is often inappropriate for qualitative research.

'Sampling' is not a misnomer in qualitative research since both qualitative and quantitative research uses samples; essentially, quantitative explicates the parameters of phenomena by numbers, while qualitative seeks more in-depth understanding of each or any one of those dimensions. Of course, to an extent, quantitative can do this too-----but still more superficially than qualitative; which is why, depending on our research purpose, we sometimes need both flavours in the mix.

A well-designed sampling in qualitative research seeks to capture as many of the socio-demographic groups and other sub-groups in the study population as possible. This uses the purposive (purposeful) sampling technique and is based on the Principle of Maximum Variation in Qualitative Studies Methodology. The purposive sampling predicated on the Principle of Maximum Variation, however, differs from random sampling used in Quantitative Studies, and the two seek to achieve different aims. Both approaches are 'sampling.' Thus, sampling is not a misnomer in Qualitative Methods.

The problem with maximum variation sampling is that when you combine it with the typical small Ns in qualitative research, you end up with 1 or 2 people in each of the sub-samples. It should be obvious that if you end up with some "unusual" people in some of those subsamples, that will influences you overall interpretation of the phenomenon.

I'm not saying that maximum variation sampling is worthless, but we need to very careful in using it, and equally careful about the claims we make for it.

My own preference is to avoid the terms sample and sampling, just because they are so widely identified with the procedures in quantitative research. Instead, I prefer to concentrate on "selection" because that is what we are actually doing -- selecting the participants for our research. So the key issue is how we define the criteria for our selection.

But that is probably just a semantic distinction, and I expect the vast majority of qualitative researchers will continue to use the language of sampling, even if that does confuse some people who are not familiar with what we mean by it.