Dear Eddie, if your population is uniform, you can take random samples simply (Regardless of spatial or temporal variations). But if you see that there can be groups or strata, you'd better use stratified or cluster sampling.
For example, you want to take samples from a country with 16 regions. If the variation between regions are lower than within region, use cluster sampling. If within variation are lower than between, use stratified sampling. In complex situation, multiphase sampling recommended.
Hope it helps
Thanks, Ehsan. But what if I do not know the size of my population (infinite)?
In Statistics When we say population, we imply indirectly an infinite population. Besides, infinity of the population in statistics is more acceptable in most cases.
Hi Eddie, I concur. Generally speaking the size of the population is typically irrelevant and if it is reasonably large, we can consider it to be infinite. In that sense we don't worry about or correct for the size of the sample in comparison to the population. In rare cases when the sample approaches the population we correct our statistics, but generally this is not a concern for reasons I am not going to get into. Generally speaking, a sampling frame is some kind of list or set of lists. So, if the population is university students, the sampling frame might be the lists of students from all universities in the country you are studying. This is where sampling strategies, such as cluster and stratification become important, because in a large country like the US, the prospect of obtaining and sampling from the lists of all university students in the country is pretty poor, so you would create a complex sample, stratified on region, university type (such as public/private), and other things, To get some idea of what this looks like, you could look at the "Final Report of the Women's Experiences in College Engineering (WECE) Project" in my profile or one of the papers on medical faculty, such as "Why are a quarter of faculty considering leaving academic medicine? A study of their perceptions of institutional culture and intentions to leave at 26 representative U.S. medical schools" to get an idea of how this looks in practice. I had more to do with the sample in the WECE study, and I wouldn't claim either is perfect. Bob
There is some confusion in your question but in my view want to ask about the sample selection when some one have infinite population.
Dear every population is infinite but to make finite sample technique is use in the research.
other is that the sample frame is different thing which draw for the sample. which we generally called list of the respondent.
in my view if doing a your research region wise, state or province wise then you need to Quota sampling under non-probability sampling technique where you divided your population in proportional or ratio wise then select sample from the that proportion or ratio. hope you will get the point what i am saying.
Hi...I am from Pakistan, I am doing my PHD dissertation. My population is university of Twin cities approx 6 renowned universities. List of population is unavailable that is why I opt Cluster Sampling. Please guide me that how I will select my sample? and number of invites to get exact sample size?
what if population of study is unknown?
what type of sampling technique we use?
@
- Badges
- Science topic
- Similar topics
- Psychology
- Emotion