The trouble with getting counted

Posted by Dave on March 23, 2010 | 6 Comments

The U.S. is ramping up its decennial census, and a few days ago, I received my census form in the mail. Or, perhaps I should say, a census form. It wasn’t really mine at all, because it wasn’t addressed to me. It was addressed to a nearby house that doesn’t get mail delivery. You see, our street—just two blocks long—has a “north” and “south” portion. My, northern, portion was developed about a decade and a half ago, and for some reason they decided to use the same house numbers as they had on the homes down the street, just with a “north” added to the street name. So there are two homes on the street with the same number: mine, and another one about 50 yards away.

The homes on the south portion of the street all get their mail delivered at the town Post Office, in P.O. boxes. I wondered if my northern neighbors had a similar problem, so I emailed my friend across the street. Sure enough, his census form had no “north” in the address either.

I decided to call the census bureau’s toll-free number to see if they could resolve the problem. After listening through two menus worth of options, none of which answered my question, I was finally transferred to a representative, who told me I should throw away my form and wait for a census taker to knock on my door.

But as U.S. Secretary of Commerce Gary Locke said on the Daily Show the other night, that’s not ideal:

The Daily Show With Jon Stewart	Mon – Thurs 11p / 10c
Gary Locke Pt. 2
www.thedailyshow.com
Daily Show Full Episodes Political Humor Health Care Reform

For every one percent of census forms that aren’t mailed in, Locke says, it costs the government 85 million dollars. Wouldn’t it be simpler for the Census Bureau to fix whatever problem it has with its database and re-send the forms to our street?

The representative on the phone told me that they received the addresses from the Postal Service, so if the database was going to get fixed, the Post Office was going to have to do the fixing.

Fine. I decided to call the Post Office. As you might guess, the Post Office was no more willing to accept the blame than the Census Bureau. The official I talked with told me that they painstakingly audited their delivery addresses on a monthly basis. Just last month, she had followed my carrier along her route and verified all the addresses on the list.

But she did give me the number of a Census Bureau official responsible for my region. So I called him. He was shocked to hear that the phone representative had told me not to send in my form—this guy must have heard Locke’s cost lecture! But once I explained the confusion with addresses, he realized that this could be a serious problem.

It would be possible, for example, for people on my end of the street to send in their misaddressed forms, which would then get counted as the other end by the census takers. Then the door-to-door census takers would return to my end of the street, instead of the other end. If conscientious census respondents like me don’t send in their misaddressed forms, the puzzle only gets more complex.

In the end, the census official told me to pick up a blank form at my town hall and send that in. That way their computers wouldn’t choke on the incorrect bar-coded version they sent me in the mail. He also said he’d instruct his subordinate in charge of my town to take special care, since many residents get their mail delivered to P.O. Boxes, and there’s lots of new development with new addresses that might not be in the Post Office database (I’m still not clear on why the Census database doesn’t agree with the Post Office).

But if it’s that difficult to count people in an upper-middle-class suburb like where I live, imagine how hard it could be in a dangerous inner-city neighborhood, or in a barrio with thousands of immigrants with questionable citizenship.

In fact, it’s probably easier, cheaper, and more accurate to count people using statistical methods that account for likely errors. But that’s illegal: The U.S. Constitution requires every individual to be counted. And given the potential for abuse of statistics, that might not be such a bad idea.

On a global basis, when trying to craft population policy, use of statistical methods is probably the only way to get a reasonably accurate count. Since politics are less likely to come into play, it makes sense to use a cheaper and better counting method. What to do with those numbers after they are collected is a much more difficult question.

Comments

2 Tweets