Samples....and confidence levels
Message posted by Andrea on October 8, 1999 at 12:00 AM (ET)

If I work out that for survey I need a sample size of 5000 for 95% confidence interval; If I take this 5000 and divide it into 5 subsections for analysis (e.g. 5x1000....probability in population for each is 1/5)...is my analysis for these 5 subsections still 95%. What if I take each of these 5 and divide them into 10 again in proportion to the complete population, do I still get 95% confidence? (Or do I have to work out a sample size for each of my subsections in turn?)

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Samples....and confidence levels
Message posted by JG on October 8, 1999 at 12:00 AM (ET)