1

I have a population of interest ($N = 5000$) for which I know some demographic information. I have a sample 1500 members of that population.

So I have a good sized sample, and I know exactly how it is biased (at least based on the demographic information I have).

Is it possible for me to make generalizations about the entire population based on the responses of my sample?

Can someone point me to a good resource on where to read about how to handle drawing inferences and calculating stats in this situation? Do I treat this as a stratified sample, but not all strata are equal (or equally representative of the whole).

cardinal
  • 24,973
  • 8
  • 94
  • 128
pocketfullofcheese
  • 111
  • 3
  • I have some vaguely related thoughts [here](http://stats.stackexchange.com/questions/31488/identifying-the-population-and-samples-in-a-study/32149#32149). They may serve as grist for the mill, although they won't really answer your question. – gung - Reinstate Monica Nov 22 '13 at 03:33
  • 5k is very specific. What is your population & your sample? What is the nature of the inference you want to make? – gung - Reinstate Monica Nov 22 '13 at 03:36
  • its not actually exactly 5000, but close. I do have an entire population (its a list of academic journals). I know journal country, discipline, etc for all journals. I know more details about my sample. I want to know, for example, the mean number of articles published per year of my entire population, based on my sample. I expect this to vary by discipline and country, for example. – pocketfullofcheese Nov 22 '13 at 17:41
  • I'm not sure what I can add beyond what's in my linked post. What matters is if the biased aspects are relevant & if so, how. If you knew the extent of the bias, you could correct for it, but then, of course, you also wouldn't have to b/c you'd already have the answer. You could look at information on finite sample corrections (eg, [this](http://stats.stackexchange.com/questions/5158/explanation-of-finite-correction-factor), or more [here](http://stats.stackexchange.com/search?tab=votes&q=finite%20population%20correction%20is%3aquestion)). Hopefully someone will come along who knows more. – gung - Reinstate Monica Nov 22 '13 at 19:13

0 Answers0