Hello, I need a clarification on the way stderr is calculated in the boottheta worksheet output. It seems to me that the values are actually SDs and not SEs. Could you explain why? Thank you, Dora
This is because if you were to divide by the number of bootstrapped samples, you’d end up with a near-zero SE with higher numbers of samples.
Hi, That is how a standard error is computed from a bootstrap distribution ( standard deviation). If you are interested in confidence intervals you can use the percentiles of the distribution. 90 % CI = 5th - 95th percentile ( leaving out 10 % of the distribution). Other methods to compute CI’s exist when distributions are skewed (bca,etc..) Bests, Samer