2010 Census: The Millions of Details

Originally published on The Director’s Blog

I am continually impressed with the number of simultaneously moving parts in the big machine that is the 2010 Census. There are few organizations preparing to have a once-in-a-decade contact within a very short period of time with over 130 million households in the country.

We have a massive list of addresses that was collected last year and now updated with new information (we have to make sure that the new information doesn’t duplicate the old). We have to translate our addresses to a form that the US Postal Service can deliver; we need to make sure that the millions of printed forms of multiple types (the English form, the bilingual form, the census of Puerto Rico) are mated with the correct return envelope and cover note and stuffed into the correct mailing envelope. Then we need to make sure that addresses are printed by machines onto the correct envelope.

With all of these moving parts, little glitches happen. Just last week, we discovered that one of our contractor printers printed addresses on the wrong forms, on a subset of questionnaires within 2 pallets of questionnaires, among the thousands of pallets of questionnaires. The error occurred during image addressing of the form through the open window. There were 10 address imaging machines working at the time, with nine machines dedicated to English “Update Leave” forms and one dedicated to “Puerto Rico Spanish Update Leave” forms. (Update Leave forms are dropped off by our own staff to areas where the Postal Service does not deliver universally.) An operator inadvertently put Puerto Rico forms into the forms feeder that was setup for labeling English stateside Update Leave forms.

The result was that a few houses in Arkansas received a questionnaire for the Puerto Rican census! We’ve tried to correct the problem and get the folks in the affected area the correct form. At least one kind person in Arkansas filled out the form and noted in the margin that they did the best they could deciphering the language even though they don’t understand Spanish. What an act of civic participation!

The glitch is embarrassing, but I suspect it will not be the last of such processing problems. We have in place quality control checks and review steps, but at the volume of work we perform, such mistakes cannot be completely eliminated. Our pledge is to admit mistakes when we find them and try to correct them as quickly as possible. We appreciate the commitment of the American public to make this a good census. With their help we’ll get everyone the form they should have, receive their answers, and process their data efficiently.

Please submit any questions pertaining to this post to ask.census.gov