There will be many references on the Internet.
That said, depends on what you want the sample for---and as posted above, other factors concerned with your research project.
Hello Arshan,
The usual formulae given for sample size determination (whether for statistical power in testing a given hypothesis or for attaining some desired level of precision of some population parameter) all assume that the sample taken is a probability sample. So, those methods do not apply in your case.
Inference to a population is far more risky and uncertain when you stray from probability sampling. (However, qualitative studies almost exclusively rely on non-probability sampling methods.)
Here's some simple generalizations for convenience sampling:
1. Larger sample sizes are preferable to smaller sample sizes. This is so because, all other things equal, larger samples are less volatile in their results than are smaller samples.
2. If you use sample size formulae or programs, treat those values as optimistic lower bounds for an "appropriate" sample size.
3. Do recall that survey forms generally have noteworthy rates of nonresponse among solicited cases, so you would likely want to oversample (increase the target sample size) by a factor large enough to compensate for whatever percent of cases contacted who would opt not to participate.
4. You can, of course, find many instances of guidelines as to how many cases one should minimally collect in order to perform some type of analytic method (e.g., in regression analysis, 10-20 cases per variable in the model as a lower bound).
5. If you collect information about personal characteristics from participants, you can make comparisons between such features for your study sample vs. a known population (for example: age, level of education, income, ethnicity, sex, political persuasion, etc.). To the extent that your sample has comparable composition on the set of variables, you could at least claim that your sample wasn't wildly different from the target population (on the features tested). Even if that's the case, there's no assurance that the similarity will apply to all variables salient to what it is you are studying.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping