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 ‘reports’

The Smoking Gun Report from the Inspector General’s Office

Tuesday, May 12th, 2009

We urge all of our readers  to take a close look at the Inspector General’s most recent reports about the 2010 Census (located here: Observations and Address Listers’ Reports Provide Serious Indications That Important Address Canvassing Procedures Are Not Being Follow—OIG-19636-01 [PDF] Report). This report details many significant failures of the Census Bureau’s recent address canvassing operations that were brushed aside by Acting Census Director Tom Mesenbourg at today’s 2010 Census hearing in Philadelphia. Here are the major problems discussed in the report:

During address canvassing field observations, we found that some Census listers were not
consistently following the procedures in their instruction manual. In several cases we observed
listers skipping the procedure for knocking on doors. In at least one case a crew leader ignored
portions of the verbatim training and instead instructed listers to omit this procedure. We
received several additional reports from listers who were specifically told by their crew leader to
omit this procedure. Further, we observed listers map-spotting addresses from their cars when
they were instructed to collect a map spot at or near the main entrance of a structure—usually the front door.

Despite instructions to traverse every road in an assignment area, some listers we observed
completely skipped roads in rural areas when they assumed no houses existed on the road.
Address canvassing in rural areas can be difficult as tree cover and other conditions can visually
obscure structures. Road conditions also can pose significant challenges: for example, rough
terrain may necessitate four-wheel-drive vehicles, and some roads may only lead to fields or
barns, or may dead-end at a physical feature such as a river. Nonetheless, canvassing these areas is essential to accurately locate rural living quarters.

OIG staff observed address canvassing in 15 different locales in 5 of the 12 Census regions. We
identified the failure of listers to conform to address listing and map-spotting procedures in 7
different locales representing all 5 regions. We also received independent information on the
same problems for 2 locales not associated with our sample. Although our observations were not conducted on a statistically drawn sample and therefore cannot be considered representative of the entire operation, the widespread nature of the problem is noteworthy.

A number of factors may be contributing to this breakdown in procedures. Skipping procedures
reduces the time it takes to conduct address canvassing. We have received reports from Census
field staff that they are under intense pressure to complete their assignments within a limited
time frame and to minimize or avoid overtime. Some are concerned they may face termination if they miss deadlines or work unauthorized overtime. Production pressure may therefore be one cause for this breakdown, but Census needs to determine why these problems are occurring.

Failure to follow procedures negatively impacts the quality of the address list, map spots, and the subsequent enumeration. Living quarters that are not included on the address list have a greater probability of not receiving a decennial questionnaire and thus not having their residents counted. Address canvassing is the primary means for identifying “hidden” dwellings, such as sheds and makeshift garage apartments, but the likelihood of missing such living quarters increases if the lister does not attempt the required personal contact. Because of smaller populations, missing a single living quarters in a rural area has a greater impact on the quality of final census population counts.

Failure of listers to correctly use the handheld’s GPS capability—a key component of Census’s
nearly $800 million field data collection automation contract—jeopardizes Census’s ability to
ensure that living quarters are recorded within the correct census block. This accuracy is
particularly important for redrawing congressional and state legislative districts.

The Census is depending on its address canvassing quality control operation to identify and correct errors resulting from listers’ not following procedures. We are therefore expanding the number and breadth of our field observations to focus on this quality control operation, particularly in rural areas. Given the problems we have identified, we are concerned that Census has not completed its contingency plan for improving list quality in the event that the results of address canvassing are found to be deficient.

These shortcuts have cost impacts as well. Quality control operations may take longer to
complete and cost more than anticipated since improperly listed addresses that are identified or
deleted must be recanvassed. Inaccurate map spots can increase the time it takes for enumerators to find their assignments during enumeration and nonresponse follow-up operations and add to their chances of getting lost and enumerating the wrong housing unit or group quarters.

Inaccurately located rural living quarters may have a greater cost impact on subsequent census
operations, as locating and driving to these potentially remote units requires greater effort than
doing so in urban or suburban areas.

Note: We have added a new permanent link on the right side of this site that will take you to the Inspector General’s most recent reviews of Census Bureau activities.