Prep for Census 2010′s First Wave

In a few weeks, the first detailed results of the decennial U.S. Census will start pouring from Census headquarters in Suitland, Md., and a Panic Season will commence in unsuspecting newsrooms. What are these numbers? Where can I download them? Didn’t we just get new Census data? Can you tell me whether Census counts X or Y or Z?

On deadline, that’s a lot of potential headache. I know you want to avoid the pain, so take some advice from a guy who survived reporting on Census 2000: prep is everything.

Here are five steps you can take now:

1. Know your Census products: These days, “Census data” means more than it did a decade ago. The advent of the American Community Survey — a survey of about 3 million households each year that replaced the old Census long form — means we get annual estimates in between the full decennial counts. And the ACS comes in three flavors: single-year data plus three- and five-year aggregates, each providing different levels of geographic granularity.

The regular releases of ACS data make Census seem more routine these days, but the data coming out soon are different. TheseĀ aren’t estimates from a sample — they’re the complete counts taken in spring 2010 via a short questionnaire sent to every household in America.

This first wave of Census 2010 data, coming state-by-state in February and March, are the Redistricting Data (P.L. 94-171) summary files. They’ll contain the basic counts of population by race for every state, county and place in America, all the way down to the smallest geographies, called blocks. As its name implies, these data will be used to redraw the boundaries of legislative, electoral and other districts in states — a process journalists will want to keep tabs on.

Later, in the summer, Summary File 1 will offer more detailed data on age, sex, households, families, and housing units — again from complete counts. Then, in the fall, we’ll see the next release of ACS data. Got all that?

2. Know the questions Census asks: Because the ACS has added complexity to Census reporting, it’s crucial to know the differences between it and the decennial count. A good starting point is to look at the forms people fill out — the questionnaires for Census 2010 and the most recent (2010) American Community Survey. Reading those will help you know what data are available and help you answer questions that might come up in the newsroom. For example, “Does the Census ask about religion?” You’ll know the answer is no, because there are no such questions on those forms.

For this first wave of Census 2010 data, note that the form only asked a few questions related to age, sex, race and relationship to the householder.

3. Know your geographies: Beyond the nation, states and counties, decennial Census data are aggregated for many geographies — from school districts to voting districts to Census tracts. TheĀ technical documentation lists them, and this geographic area reference manual, though old, has a ton of info. It’s helpful to know, for example, how blocks build up into block groups and then into tracts and what each level represents. Reviewing the available areas also can spark story ideas. Which of your local school districts had the biggest population increase? How has diversity changed in your state’s congressional districts?

4. Know the 2000 data: A big part of reporting Census 2010 is comparing the new data to the 2000 count. Hunting down comparable 2000 P.L. 94 data will be a chore on deadline, so do yourself a favor and get it now for the geographies you plan to cover. You can find tables on the Census Factfinder, or if you’re super adventurous you can download the raw data by state. If you’re going that route, be sure to read the technical documentation.

5. Know where to learn more: Study the Census Bureau’s site, of course, but there’s more out there. Investigative Reporters and Editors has a Census resources page with information on training, webinars, tipsheets and links (assembled with the help of my Census mentor and colleague Paul Overberg). The Pew Research Center has an All Things Census blog where it’s tracking the progress of the count and the meaning of the results. And the upcoming IRE computer-assisted reporting conference in Raleigh will offer sessions on Census as well.

Good luck and have fun, Census hounds.

3 Responses to “Prep for Census 2010′s First Wave”

  1. I think it is also very important for reporters and editors to become familiar with the concept of Margin of Error that is built into the ACS data. These data are published with Margins of Error because they are sample-based estimates.

    There will be cases where what looks like a real change from 2000, or a real difference between two geographic areas, is not statistically significant and should not be reported as a difference. Also, different data sets from the Census Bureau are useful for different purposes and to mix them up may do a disservice to the public.

    I strongly recommend the “What the Media Need to Know” edition of the Bureau’s “Compass” series of handbooks for understanding the ACS.

    See http://www.census.gov/acs/www/guidance_for_data_users/handbooks/

  2. Great Job on being ready to go with this! I’m a data junky and look forward to seeing the numbers along with your analysis!

  3. Anthony says:

    Mike,

    That’s an excellent point. Many journalists had a difficult time deciding what, if anything, to report out of the recent five-year ACS data because the margins of error for the smaller geographies made comparisons between geos mostly meaningless.

Leave a Reply