How can non-proportional sampling be used to "correct" for a problem in the accessible population?
How can non-proportional sampling be used to "correct" for a problem in the accessible population?
Answer: Assume an accessible population that doesn't match the target population. For example, there are way too many 18-21 year-olds in the Elementary Psych Pool for it to be representative of Iowans (or Americans or humans) in general. You can try to correct for this by purposefully over-sampling people older than 21 from the Elem Psych Pool, to get the sample percents closer to the target values. This is, in a way, a case of two wrongs making a right. Your accessible population was wrong and your method of sampling from it was wrong, but the combination turns out to be closer to right.
Comments
Post a Comment