My Two Census

Formerly the non-partisan watchdog of the 2010 US Census, and currently an opinion blog that covers all things political, media, foreign policy, globalization, and culture…but sometimes returning to its census/demographics roots.

Posts Tagged ‘non-response’

MyTwoCensus Exclusive: Census Bureau Official Debunks 2010 Census Field Operation Myths

Monday, March 29th, 2010

The following piece does not reflect the views of MyTwoCensus.com and was written by a senior Census Bureau official who has requested anonymity (but whose identity has been verified by MyTwoCensus.com):

With the Census starting to hire for the largest census operation, Non-Response Followup next month, Mr. Stephen Buckner neither denies nor confirms the length of census jobs but says they will last up to several weeks. However, he is defending an agency that has a long history of mismanagement such as the Harris handheld computer debacle, address canvassing cost overruns and group quarters workload overestimation to name a few. His response does hint at many of the assumptions the modeling and analysis branch at headquarters makes:

Assumption: The average employee will work part time hours (for example 25 hours a week for Group Quarters Advance Visit).

The decennial census is being conducted in the midst of the worst unemployment rate since the Great Depression. Many people are desperate for any sort of work even if it is short term. These people will work more than 25 hours a week because they have been out of work for months. They jump on the opportunity with enthusiasm because and this is any sort of a glimmer of hope they’ve seen in months. You work more than 25 hours a week, chances are you’ll finish quicker.

Assumption: The average Census employee has eighth grade reading, math and map reading skills.

Due to the high unemployment rate the Census Bureau has attracted very highly educated overqualified employees. Most employees have college degrees and some have advanced degrees. There is no way to measure how quickly the highly educated workforce will work. But they will complete training and understand it much better than those of an eighth grader making for quicker completion of work.

Assumption: The average Census employee will complete about one non-response interview an hour.


The Census Bureau has the same cookie cutter production quotas for the local census offices however it does not account for:
* hard to enumerate areas in urban cities with a large immigrant and illegal population
* high density housing where travel time between assignments is negligible
* rural areas where enumerators and listers can get lost, and their vehicles stuck

Assumption: About 29% of employees won’t qualify for work because of an unfavorable background check. And up to 50% of applicants will either quit or be released from employment.

This leaves about 71% of employees who are available to be hired. If you factor in job refusal, not showing up for training, not completing training and resignations; the bureau thinks they need to recruit five applicants for every one position and front load each operation with 1.5 times the required staff. When you take millions of highly educated, unemployed Americans who were previously were working they are desperate for work then you have favorable background checks, and people who won’t refuse work, will show up and complete training. And although it is a tough job it pays much better than food service and retail jobs so they probably may not quit as easily as the Census model has them.

Assumption: The performance evaluation system for Census Bureau managers is purely based on meeting production goals or exceeding them.

Mr. Buckner says that “we want nothing more than to hire less than what is required to do the job.” This is untrue especially when you consider the following. If manager A finishes ahead of manager B then manager A is the better manager. That is why managers overhire, overstaff because the quicker they finish the better their performance is perceived. So managers maximize their chances for success by hiring the maximum number of people need for the operation.

Assumption: The more quickly you finish the more work you are given.

The Census Bureau rewards those who finish quicker with more work. If another area is working slowly they will assign the work to someone else or take the food out of people’s mouths. Managers assume those who finish quicker are better workers and the quicker they finish the better management looks.

Assumption: Like the military, the census bureau relies on the chain of command and military group think.


The Census Bureau works just like the military chain of command.  If you question authority then it will result in a dead end for career employees and termination for temporary employees. If you don’t like the group you are deemed not a team player or insubordinate. The managers knew that training all these people during address canvassing was unnecessary but they did so anyways because no matter how stupid the idea sounds the agency teaches you not to question authority and do as you are told. So during address canvassing many managers trained more people than needed and replaced people who turned over although work was ahead of schedule.

Any business school model (and it doesn’t need to be a Harvard, Wharton one) can tell this is a managerial economic problem. A professor once told me you can solve any managerial economics problem by asking these three questions. One. Who made the mistake? Two. Did they have enough information? Three. Are their incentives in lines with the goals of the company? The answers are respectively: Census Bureau managers, no and no. The Bureau prides itself on collecting high quality data about the nation’s people and economy. However their managers are evaluated not on producing quality work at the lowest cost but on how quickly they can produce it. There is no incentive for coming in at the lowest cost. That is why the Bureau is getting so much bad publicity lately for over hiring, overspending and false pretenses of good short term employment for the millions of Americans desperate to find any sort of work out there.

In an economy such as this the Census Bureau should do the American people a favor: hire the right amount of people, give them some solid work that pays well and keep them working longer, providing them some hope and inspiration that their federal government cares about them and treat them with the respect they deserve in this time of hardship. Non Response Followup is the Census Bureau’s final chance to make good on their promise to count every person accurately, let 600,000 employees make a paycheck who desperately need money.

It hurts no one if the Bureau hires less people than what they require because the enumerators are smarter in this census so they will be more productive. If the Bureau under recruits then they save money on advertising, the fewer employees work longer to put food on the table. If they over recruit and over hire then they waste money on recruiting and training, there is a false pretense of work and employees lose hope in an agency already fraught with problems. Under recruiting, hiring less people to work longer and making the most effective use of their employees might actually benefit the census. With all the President is  dealing with such as health care reform and two wars in Afghanistan and Iraq I think the last thing he is worried about is getting the count of Americans delivered to him on December 31st, 2010.