All surveys that attempt to obtain high response rates in the United States have been experiencing declining participation rates over the past 20 years or so. In the best scientific surveys in the country, interviewers are asked to make repeated calls to first contact the sample households, and then repeated contacts to address concerns of the sample persons, sometimes completing the interview in several settings.

It is no surprise, therefore, to any one in survey research that the costs of the US decennial census have been increasing over time. In my personal judgment, much of the cost inflation is from increased reluctance to participate.

In contrast to a sample survey, however, everyone is targeted in a census. Therefore, solutions to cost reduction and increasing participation are vastly different than for surveys. Mass and targeted advertising is an option for a census.

The decennial census begins by mailing a questionnaire to most persons; this is the cheapest mode of data collection from households. If everyone filled out that form and mailed it back, vast sums of money required for followup activities to the nonrespondents would be saved. For this reason, much attention at the Census Bureau is focused on how to increase the mailback rate.

For three decades after the 1970 census, mailback rates fell sharply until the 2000 census. That was the first time the Congress authorized the Census to launch a paid advertising and public relations effort to help slow this rate of decline. We exceeded expectations in 2000, and we not only halted the decline but the American public increased their response rate. As a result the Census Bureau later in 2000 returned to the federal Treasury some $305 million in savings, partly because of this renewed civic engagement. The Congress and our oversight agencies generally applauded this effort in 2000, and encouraged us to do more for 2010 to help get a better count, improve accuracy, and hopefully again save the government some important funds.

We’re advertising again. The management equation on this is pretty simple. For every one percentage point we increase the mailback rate, we save about $85 million dollars of followup costs. This is a business proposition. We seek ways that get the message out, even if it is unusual for a Federal agency to do so. We spent about $85,000 on the Alaska trip to kick off the census, but garnered an audience from publicity (85 million) that would have cost 300 times as much in paid media (see earlier blog entries on the Noorvik trip). We will advertise on the 2010 Super Bowl, as we did in 2000. The Super Bowl is the top-rated and most highly anticipated television event in the U.S. An ad running once in the Super Bowl has the potential to reach 45% of adults over age 18. A thirty second spot on the top-rated regularly-scheduled show in America, American Idol costs $450,000 and is viewed by just about 9% or 10% of all households watching TV. The Super Bowl reaches 100 million viewers at a very efficient price compared to other shows.

The Super Bowl is also rare, in that viewers are just as tuned in to see the commercials as the program itself. Commercials that air on the Super Bowl have a multiplier effect. Advertisers are mentioned in multiple news media outlets and viewers will typically look to view them online almost immediately after airing. Therefore, airing once in the Super Bowl creates significant buzz leading to additional viewing potential. If just one percent of the folks watching the Super Bowl had their minds changed to mail back a census form they would have otherwise ignored, it helps save the taxpayers between $25-30 million in expensive follow up costs to collect these forms later.

We hope all these efforts help encourage every household to mail back their Census form. Each one of us who does can take a bit of pride in knowing that when we do, we help save our government substantial sums.

Director Robert Groves