Generating a 95% CI on Small Data Set

[Paige Kingston]

Hello,

I am trying to generate a 95% confidence interval for a small data set that is not normally distributed, by using the 95% confidence interval of a similar data set I have as a proxy for the distribution of the data in question. I have not been able to figure out how to code for this in STATA. Is there a way to do this?

Thanks for the help.

[Nick Cox]

I don't follow what you're asking here. You already know, or will calculate, the interval you are going to transfer from one context to another. What else is there for Stata to do?

Not the question, but I can't encourage this if I understand it correctly The whole point of a confidence interval is to be up-front about uncertainty. Also, a variable not being normal is neither here nor there, as either you use some appropriate specific or general method, or you fall back on the key result that means are usually close to normally distributed even if the underlying data are not.