Is this right?
Stratified Proportional Sampling:
For example there's 2 Groups: Grp 1 & 2:
(Grp 1's total population/total combined pop of both Grps) x total sample population needed
= sample population needed from Grp 1
Your formula is correct. However, stratified sampling and stratified proportional sampling are usually considered the same by many authorities. When we do stratified sampling we usually consider the percentages in the allocation of sample size.
You should use stratified proportional sampling when you have a good estimate of the number or proportion of cases in each strata. If you do know know the proportion in each strata, you should use simple stratified sampling (50/50). Otherwise, you may bias your results with inaccurate proportions.
Cady -
If you have an idea of the standard deviation in each stratum, and you'll need to know that to estimate sample size needs very well, then you should consider more optimal stratification. A good textbook such as
Cochran, W.G(1977), Sampling Techniques, 3rd ed., John Wiley & Sons, or many other old standards, such as Hansen, Hurwitz, and Madow, or Kish should be very helpful for you. Other books to consider might be the more introductory
Blair, E. and Blair, J(2015), Applied Survey Sampling, Sage Publications,
or the well-written opus by Sharon Lohr:
Lohr, S.L(2010), Sampling: Design and Analysis, 2nd ed., Brooks/Cole.
Other books, if you have auxiliary data available as well would include
Särndal, CE, Swensson, B. and Wretman, J. (1992), Model Assisted Survey Sampling, Springer-Verlang, and.
Brewer, KRW (2002), Combined survey sampling inference: Weighing Basu's elephants, Arnold: London and Oxford University Press.
You can also find good information online. The Pennsylvania State University puts a good deal of good statistical information online for its classes, available to all. Here is an example:
https://onlinecourses.science.psu.edu/stat506/node/28
There you will see the proportional allocation you mention shown first, followed by optimum allocation. Also, at the bottom are links to other "lessons."
A pilot study could help with this and other aspects of your survey.
Cochran has a chapter on sample size needs for simple random sampling, which includes suggestions for estimating standard deviations, such as a pilot study, and then a chapter on stratified random sampling, with more information on sample sizes, followed by other information on sample designs, etc.
Cheers - Jim
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis