Hello, Can someone provide a simple explanation for the 4x multiplier in the VRE variance estimate question? I'm not familiar with survey design and analysis, so the references in the documentation are hard to follow. I assume this arises from the survey design and weighting process.
That formula (4/80 * sum of squared differences) seems to also imply that the values in the VRE tables are not directly usable as replacements for the published estimate. If the VRE table values are directly usable, then the variance formula would be the usual sample variance.
the question is coming from a need to create a custom measure from the sum of ACS values. I'd like to bootstrap re-sample the data to create robust measures of the custom value's variance.