Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 2 down vote favorite

Following the question here, Probability Proportional to Size (PPS) Sampling Method seems appropriate in constructing a sample when the primary units are of different sizes in the target population.

As I understand, this is a self-weighting technique, so does not require any special consideration in the data analysis stage. It is also a probability sampling technique so meets the assumption of correlation and regression models.

I am looking for some good resources that can help me better understand this technique. Perhaps a flowchart that illustrates this method would be great to show to my 'non-statistical' target audience.

share|improve this question

2 Answers

up vote 3 down vote accepted

To me, the ultimate resource on PPS is Brewer and Hanif (1982) Sampling with Unequal Probabilities. Unfortunately, it is nearly impossible to lay one's hands on. It is also highly technical and assumes a knowledge somewhere between Lohr (2009) "Sampling: Design and Analysis" and Thompson (1997) Theory of Sample Surveys. The latter lists and explains about half a dozen PPS methods (Brewer gives about 50).

share|improve this answer
Thanks Stask. I am not very 'strong' in statistics, so I am after a simple explanation and a guide on PPS (though I will access the books you have recommended). I found a paper that explains it quite well. I don't have the points to link it here but it is in the post that is linked in the question above. What do you think of the method outlined in that paper? – Adhesh Josh Jan 30 '12 at 22:59
It's a fair explanation of the method, but it's a different design: a systematic PPS sample. You cannot have unbiased variance estimates with systematic samples unless you utilize some special tricks (e.g., running two threads with two independent seeds and twice the sampling interval). – StasK Feb 2 '12 at 19:39
up vote 0 down vote

Because we are required to make selection probability for each element included in the study to be proportional due to its size measure.

share|improve this answer
2  
Welcome to the site, @user20698. I don't see how this is an answer to the OP's question. The OP is looking for books or other resources that illustrate this method & could be used with clients. Could you edit your answer to address that? If not this answer should probably be deleted. Since you're new here, you may want to read our about page & our FAQ (which contains info on answering questions on CV). – gung Feb 10 at 16:04
2  
... and please use your profile to customise username and publish your affiliation. – mbq♦ Feb 10 at 20:38

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.