Who estimated the 2010 Census supplies and printed materials contract? (RIP dead trees!)
The following article comes from a former 2010 Census manager:
A month ago, MyTwoCensus posted a picture of hundreds of boxes of materials that were being recycled. Although no one will argue that recycling is good for the environment, the truth is none of this should of been printed in the first place. The Census Bureau estimate of the census forms they needed were off completely wrong and here’s why:The company contracted to print the enumerator questionnaires (D1-E) printed over 200 million questionnaires. However the NRFU workload was only estimated to be 47 million households. Critics could argue that the NRFU workload was unknown before the printing contract was awarded however the American Community Survey estimate of 2008 only showed 128 million households in the United States. If America had a zero percent response rate they would still require only 128 million questionnaires, not the 200 million that was printed. There were 15 million enumerator supplemental questionnaires (D-1E SUPP) printed which field staff used when there were six or more household members. Information on how many households have six or more people was unavailable however only 23.4% of American households have 4 or more members. Any statistician can say with confidence that households with six or more members is negligible.The errors in estimate cascaded to other printed materials such as information sheets and notice of visits also being overprinted. For example even though census procedures specifically allowed only three personal visits; in some regions to increase accuracy and avoid going to a proxy some enumerators made more than three visits. However Stephen Morse’s picture clearly shows there are still hundreds of boxes of these forms unused. Another example were the forms for eligibility employment verification (I-9) forms. The contract for the I-9 forms was 20 million, yet the census only ended up hiring about 600,000 employees across the nation and estimates show only a couple of million applicants.None of this news should be surprising, Census models are completely inadequate. In 2009 during address canvassing they threw away millions of taxpayers’ dollars on training employees for which there was little or no work available. Each office returned palettes and palettes of office supplies such as pens, pencils, paper clips and rubber bands. The companies contracted to provide these were the ones who received the money.If Census managers are infuriated over this picture perhaps they should be funneling their energy towards ensuring that their money is spent on technology that actually works and the proper amount of printing. Surely printing 200 million questionnaires for just 48 million households assuming assignment prep error is a little overkill.