Monday, June 22, 2009

It's Not You, It's Your Data

I was away at my niece's high school graduation in Charlotte, NC in the middle of June, so I didn't come upon Bill Dowd's What? Why me? Intrusive queries from the Census Bureau, published in the Albany Times Union on Sunday, June 14, until this past weekend.

Mr. Dowd writes what I guess is a cheeky dealing with the American Community Survey. But I'll reply to his points seriously anyway.

"In a few days your household will receive a questionnaire in the mail for a very important national survey, the American Community Survey. When the questionnaire arrives, please fill it out and mail it back promptly. The U.S. Census Bureau is conducting this survey and chose your address, not you personally, as part of a randomly selected sample."

Interesting that this very process nullifies any statistical viability of such a survey.

The address is randomly sampled; I do not understand how this "nullifies any statistical viability". It's not totally random in that, if your address is picked, safeguards are in place so that the address won't be selected again soon.

Not only have I now received another Census Bureau letter concerning "The American Community Survey," I also received a questionnaire I am informed I must fill out and return, under penalty of law.

From Census News Briefs #67: "A person who 'refuses' or 'willfully neglects' to answer any census questions is subject to a fine of not more than $100. A person who 'willfully' provides false information can be fined up to $500." But it isn't the intent of Census to get people fined; it is to get the information it's required to collect.

They again demand my cooperation, even though they swear they don't know who I am since I was "randomly selected" -- by them.

Same source: "The Census Bureau requests names of people residing in the household, to ensure full coverage and minimize the likelihood of duplication, and a telephone number, in case it must follow-up to obtain missing or unclear information..."

I suppose I'm being unnecessarily stubborn in asking what these survey questions have to do with those topics:

How many times have I been married?

How many times has my spouse been married?

How are the two people in the household related to each other?

What are the household residents' ethnic origins?

When did the residents move into the current residence...

What is my monthly condominium fee, if applicable?

Has anyone in the household ever served in the military?

What kind of work do the residents of the household do?

Relationships and residence characteristics have implications in everything from health care to road design. Other questions involve voting rights, government benefits and other issues.

Unless the Census Bureau has, while I was sleeping, been assigned the task of planning the nation's highways, running the health care system, handling public safety programs, and building new schools, that makes no sense. The information it demands obviously will be shared with someone outside the Census Bureau. Many someones, no doubt.

No, Census does not plan the nation's highways; it provides the data in the aggregate which allows transportation planners to plan highways, educators to build or raze schools, and the like. Bill Dowd's individual information isn't being shared. Roger Green's individual information isn't being shared. The data provided show how many children in a particular block group, not characteristics of a particular child. In fact, the Census Bureau is very quick - overly quick, some believe -to SUPPRESS the data when it seems possible likely that an individual person or household might be revealed.

I'm hoping that Mr. Dowd actually look at the data already created through the American Community Survey so he can see that no individual person's information is revealed, only totals for various geographies. If he -or anyone else - needs help deciphering this, they can leave a comment and I'll do my best to explain.
On another Census topic, I was please to see this: White House Press Secretary Robert Gibbs confirmed during a press briefing [last week] that the Obama Administration was "determining the best way to ensure that gay and lesbian couples are accurately counted" in the 2010 census. The Wall Street Journal reported... that the Administration had directed the Census Bureau to explore ways to tabulate responses to the census relationship question, to produce data showing responses from married couples of the same sex.

The Census Bureau decided during the Bush Administration that it would re-code same sex spouse responses as "unmarried partners," citing provisions of the federal Defense of Marriage Act (DOMA) that it believed prohibited the tabulation of data on marriages that are not between opposite sex couples, as well as concerns about data quality based on responses collected in the American Community Survey. Rep. Mike Quigley (D-IL) sent a letter earlier this week to President Obama, urging the Administration to recognize same-sex marriages in the 2010 census. Rep. Quigley and more than 50 other members of the LGBT Congressional Equality Caucus wrote to the President last month, suggesting that the Bush Administration had misinterpreted the DOMA and saying that reporting statistical data on same-sex couples "simply provides basic information about how Americans respond to the Census Bureau's questions" and "is not tantamount to federal recognition of same-sex marriage."

I never believed that DOMA forbade reporting what is happening in the states, and this is the type of data openness I suspect that Census would have preferred.

No comments: