From: mconely@arcticcatinc.com
Date: 5/11/2005
Time: 11:55:30 AM
I am not able to statistically derive how or why it is acceptable to say that the standard error of the mean for a confidence interval is(or can be estimated as) the sample std. dev divided by the square root of the sample. Large sample sizes result in a relatively small value range for high confidence intervals(ex. 99% confidence interval). I believe it is trying to adjust sample data so it more accurately predicts population parameters but I am leery of trusting it and can't follow methodology. Can anyone explain how the approximation was derived or why it should be considered statistically valid? My real life experiences have shown that population parameters are typically very close to the sample statistics originally calculated(meaning +/- 4 sigma based on sample statistics usually results in about 95.5% of the population).