Follow Up to Answer to "Release of 2009-2013 5yr PUMS"



I appreciate the clarity of your answer describing how to combine single year PUMS release to generate multiyear estimates. I have a couple follow up questions.

  1. Like some PUMS users, I do not have ready access to access to statistical packages and rely completely on the on-line tool for analysis. In the case of recreating multiyear estimates produced by the Census Bureau, is it possible to recreate these by analyzing individual years using the on-line tool then averaging the results? In theory, one might expect each year to contribute 20% of the five year average value, but I realize that this is by no means assured. I suppose I could try this then compare the results to the PUMS Estimates for User Verification, but am hoping you could way in before I head down a potentially long, fruitless road.

  2. Could one use this method to create a valid three year average for 2012 to 2014? I realize that there would be no user verifcation estimate that could be used as a benchmark.

Thanks for your assistance.

Cliff Cook



Please see this previous answer on approximating the multi-year samples using the Online Data Analysis System. You can follow a similar method to approximate the 2014 multi-year samples; however, you are correct that PUMS estimates for verification purposes are not available if you do choose to create a 3-year estimate for 2014.

Hope this helps.