Migration data for small areas - error



I’m pulling migration data, in and out, for a relatively small population (22-34 y/o, with BA or above) and a small area (Grand Rapids MSA). I’m going in knowing that this may be just too small to get anything out of.

I’ve pulled together all the statistics and can see some trends, but I’d like to put some confidence intervals around the data too see if they may not be statistically significant.

How can I find errors or anything that can help me do this?




I apologize for the slow reply. Unfortunately, we didn’t receive an automatic notification of your question.

If you are using the online data analysis tool then you’ll want to check the box for ‘Confidence Interval’ and/or ‘Standard error of each percent’ under ‘Table Options’.

If you are using a statistical software package, then the specific method will depend on the package and your preference about how you want the standard errors to be calculated. One option is to not adjust for cluster or strata. Otherwise, basic sample code for using cluster and strata can be found here. Finally, depending on that sample you are using, you may want to utilize replicate weights. More information on replicate weights can be found here.