Census Forms May be Short but Not Sweet

Lock your doors and protect your sons and daughters, the American Association of Geographers are in town this weekend for their annual convention. What does this have to do with strategic communications? A lot of geography these days is more about statistical analysis of population data than it is about drawing maps.

Last night, I was having beers with my cousin who is an economic geographer and in town for the weekend’s festivity. After the idle chitchat about the pros and cons of New York’s grid system dies down, the main topic of conversation this year will be the 2010 census data.

The U.S. Census Bureau used to send out a long form and a short form. They switched to just the short form for 2010. They also rewrote the questions but promised the data would be just as good. It’s not.

For many communities the data is basically meaningless. Just one example my cousin gave was a community that had a reported African American population of 10 percent plus or minus 30 percent. There are serious issues with your data if your margin of error is larger then your statistic.

Researchers and statisticians are trying to salvage what they can by cross-referencing census data with third-party public data but it’s slow going. In the mean time, we have to view studies based solely on census data with a critical eye.

If you see a geographer at a bar this weekend, buy her a drink and be thankful it’s not your job to sort this mess out.