Hi everyone. I'm new to ACS data and am a bit confused. Below is a screenshot from page 64 of ACS General Handbook where the formula for determining MOE of the proportion and a worked example.
However, it appears to me that the worked example swaps the MOE(X-hat) and MOE(Y-hat). Am I totally wrong or missing something?
Any help totally appreciated. Thanks so much!
Dear David--
This is a huge problem in the ACS. In the staqndard publicshed numbers, they use the normal approximation to compute the putative standard errro, which they turn into a margin of error. However, for situations where the percent of given category is small (usually thought of as less than 30%) or large usually though of more than 70%, they still use the normal approximation for the estimate. This results in some estimates taking on absurd values. For instance, if they find one person of a given group in the sample they will put out a margin of error that goes below zero, which is obviously absurd if one case was found in a given distribution. To remedy this they now also report errors based upon random replicates, which are also possible to compute using the PUMS data. The other approach, which seems to only have been used for so-call Language Determination files is to use Bayesian estimation techniques. Shortly after they began releasing the ACS I brought this issue to their attention and they responded, eventually saying that it was an important issue and they were trying to figure out how to deal with it in a producation environment. In a way the Random Replicate approach is the answer. For your edification, here is the correspondece I had with the Bureau https://www.dropbox.com/s/lfjo11wm5ma1axx/Memo_Regarding_ACS-With_Response.pdf?dl=0 ;