0
$\begingroup$

I have a sample of count data (N = 226) representing a parameter of a population. The sample contains many zero values and a few non-zero values. How can I best estimate the level of confidence that the population is being truly represented by this sample?

The 226 0.1 acre plots are distributed across a 3000 acre study area. On these plots, snags over a certain size were counted. Eight plots had one snag, and one plot had two. I need to estimate the total number of snags on the study area.

$\endgroup$
1
  • $\begingroup$ "Level of confidence that the population is truly represented by the sample" -- you can't really know that, unless you know something more about the population than you do about your sample. You can look at the bootstrapped distributions of statistics of your sample, but they only generate confidence intervals for their population counterparts in the case of random sampling. And with few non-zero data points, it may be that none fall in your sample. $\endgroup$ Jan 8 '16 at 19:43
0
$\begingroup$

You won't be able to find the answer analytically, but I'd suggest that you look at how you sampled numbers. What was the mechanism by which you got the data? Is it reasonable that the population yields these values? If not, then I think it's very likely that there was an error somewhere with your sampling process and you might want to find another way to sample.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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