My Two Census

Yesterday, the New York Times ran the following editorial:

The 2010 census, in its final stages, has apparently been a success — something not thought possible just a couple years ago, when unsteady management, political interference and other problems threatened to derail the effort. The count was salvaged only after last-minute scrambling and major new spending — and after new leaders were put in place by the Obama administration.

For a time, it seemed as if Congress would learn the lessons from the near disaster of 2010. In March, a bipartisan group of House and Senate lawmakers introduced a bill to improve the census, mainly by giving the bureau director more power to run the agency without interference. In April, the Senate committee in charge of the census unanimously passed the bill. The bill has not gone anywhere since then.

Why does that matter, when the next count is a decade away? The best chance for passing a bill is now, when public awareness of the census is high. And the sooner reform is passed, the better, because census planning, done right, is a decade-long project.

The administration, which had to rescue the current census, should certainly know that. But it is the administration that appears to be standing in the way.

At a hearing this spring, one of the bill’s co-sponsors, Senator Thomas Carper, Democrat of Delaware, said that Commerce Secretary Gary Locke had complained about a provision giving the director greater independence to communicate directly with the commerce secretary and Congress about problems with the census. He said Mr. Locke also objected to giving the director greater influence over the bureau’s budget.

Mr. Carper suggested that independence to communicate was nonnegotiable, but a compromise on the budget could be found. There is no sign of progress.

In the next few weeks, Mr. Carper’s staff will issue a report on the bill to help other senators as they consider the legislation. The bill is a brief 11 pages and it is uncontested, at least on Capitol Hill. How much help do the other senators really need?

Mr. Carper should speed up the report. If the administration still has problems with the bill, it should make them public and allow the process to move forward openly. Basic reform of the census is needed, and the time to make those changes is slipping away.

MyTwoCensus analysis:

The first part of this editorial labels the 2010 Census a “success” but never states why it is considered as such. Perhaps this is based on the cursory observation of the participation/response rates that were similar to those of 2000. This may be a “success” when taking a quick glance at figures, but let us remember that the Census Bureau’s budget for 2010 was infinitely larger than it was in 2000. (And it took home an extra $1 billion in funding from the stimulus package.)

The second half of this disjointed editorial has a bit of validity, though it isn’t articulated well. Yes, it would be better for America for the Census Bureau Director to have a fixed term that ends in a year that is in between Presidential election years. But Gary Locke has legitimate concerns, and those must be addressed before rushing a bill through committee. The same Senate that can’t pass Climate legislation that’s been on the table forever shouldn’t be expected to jump on legislation related to the 2020 Census.

And here’s a little caveat/prediction for the New York Times: When the mainstream media learns just how much of a mess the 2010 Census was in some parts of the country, and in particular New York (where a dense concentration of media moguls and reporters utterly failed to cover the giant mess that is the New York regional census office) they will be begging for re-enumerations, recounts, and heads to be put on the chopping block. MyTwoCensus.com will elaborate more on this information in the coming days and weeks.

Note: An earlier version of this post questioned why President Obama hadn’t signed a bill seeking to reform the GOP’s “census” mailers. I referred to a blog post that I wrote on May 18, 2010. I subsequently learned from comments on this post that President Obama signed the bill on May 24. I was never made aware of this action by President Obama until today and I apologize for the confusion. Those people who refer to a bill from April should know that the GOP found a loophole in this legislation and continued to issue deceptive mailers. Furthermore, the comment about President Obama was just an aside from a post that focuses on many other important matters which I hope are not overshadowed by my simple error.

We are glad to see the New York Times and New York Daily News reporting about the Census Bureau’s latest scandal in Brooklyn. At this point, the big question is whether the individuals involved with this data fabrication effort will be formally charged with crimes. Hopefully by Monday we will know the answer…

Though the New York Times reported that New York now has 8.4 million people according to 2009 Census Bureau estimates, MyTwoCensus is hoping that Sam Roberts and his colleagues at the Times will be able to investigate the likely undercount of New York City stemming from the fact that 2010 Census enumerators are finding themselves unable to access buildings because of doormen and other security restrictions. This is a problem that has been reported to MyTwoCensus.com on many occasions. And it doesn’t seem to be getting fixed any time soon. Will NRFU operations end before this problem is solved? Will Census Bureau officials resort to the illegal/unconstitutional practice of guesstimating how many people live in each building and then make up false information about the ages and races of the occupants?

If you have answers, don’t hesitate to let us know!

Today, the New York Times published an editorial that praises Congress for initiating bi-partisan reforms of the Census Bureau as it initiated legislation that mandates the Census Bureau Director’s term to be fixed at five-years, a plan that makes it easier to work around the decennial census. However, Secretary of Commerce Gary Locke and the White House were at first keen on this idea, but have now stalled the plan, despite seven former Census Bureau directors asserting that this is the best way to reform the Census Bureau. Robert M. Groves, the current Census Bureau Director, also supports this plan — but apparently the egos of the others have got in the way of progress:

The Obama administration, which should be supporting the bill, is instead raising objections. It has objected to a provision that would allow the census director to report directly to the commerce secretary. It also has objected to a provision that would require the director to send Congress the bureau’s budget request at the same time it goes to the White House.

However, the editorial strays from its initial goals later on and says this:

The census was in dire straits when President Obama took office, and it took a while for the administration to get organized. The 2010 count is now on track, thanks to the efforts of Mr. Locke and Robert Groves, the bureau director — both Obama appointees.

The New York Times has it wrong. The Census Bureau and the 2010 Census are not “on track” at this point. The myriad technical failures and other problems have already hampered the accuracy of this count and will continue to do so in the immediate future mean that the 2010 Census is NOT on track.

The following New York Times editorial concerns the class action lawsuit that we reported on last week. For many months now, MyTwoCensus.com has criticized 2010 Census hiring practices. Here’s the editorial:

The Census Bureau is hiring a million or more people to assist with the 2010 count. It is temporary work, but it pays well. With national unemployment at nearly 10 percent, it looks like an excellent opportunity. That is unless you are one of the nearly 50 million Americans with any arrest or conviction on record.

A new class-action lawsuit has been filed on behalf of applicants who say they were unfairly turned down for census jobs based on an opaque screening policy that relies on F.B.I. checks for any criminal histories. Those checks are notoriously unreliable. A 2006 federal report found that half of them were inaccurate or out of date.

The Census Bureau is vague about what makes someone ineligible. In Congressional testimony, it suggested that it is excluding people who have been convicted of crimes involving violence and dishonesty. The bureau’s Web site seems to say that applicants whose background checks turn up any arrest — no matter how trivial, distant in time, irrelevant to the job — receive a letter advising them that they can remain eligible only if they produce “official court documentation” bearing on the case within 30 days. Incredibly, the letter does not identify the alleged criminal activity. Applicants must prove eligibility, even if they don’t know why they were flagged.

Official court records are often unobtainable for the millions of people whose convictions have been sealed or expunged or for people who have been arrested and released because of lack of evidence or mistaken arrest. This problem falls heaviest on black and Hispanic communities where stop-and-frisk policies and indiscriminate arrests are common.

The hiring problem is not limited to the Census Bureau. After 9/11, Congress required port workers to undergo F.B.I. background checks to keep their jobs. Last year, a study by the National Employment Law Project, an advocacy group for workers, found that the government had mistakenly denied credentials to tens of thousands of those workers.

States and cities are wisely revising employment policies. The federal government needs to develop a fair and transparent screening system for job applicants and a more effective appeals process. Congress must also require the F.B.I. to verify the criminal records — and find missing data before issuing background checks.

Below is Stuart Elliot’s commentary as he answers readers’ questions about 2010 Census ads. Will the Census Bureau and Draftfcb threaten to take away the New York Times’ contract because Stuart was somewhat critical of the ad campaign?

Q: (Reader)

Just this morning, as I was reading NYTimes.com, I was struck by the short films that the U.S. Census is running to help persuade people to return their census forms.

Usually I ignore any and all advertisements online, but I found these fascinating because they present people who have reasons (good or less good) for mistrusting the government as encouraging citizens to participate. Who is responsible for the films?

A: (Stuart Elliott)

The films are part of a Web series, called Portrait of America, which features “real people expressing their reservations about participating in the Census and then overcoming them once they examine the form,” says Wally Petersen, a spokesman for DraftFCB in Chicago, part of the Interpublic Group of Companies.

The DraftFCB New York office created the Web series in its role as the lead agency for the Census campaign. “More than a dozen agencies produced more than 400 pieces of marketing communications” to encourage participating in the Census, he writes in an e-mail message, adding that the work is “targeting multiple audiences” in terms of races and ethnicities and appears in 28 languages.

Q: (Reader)

Something about those ubiquitous U.S. Census television ads has me scratching my head: the closing call to action. On screen invariably is the phrase “Census 2010” while the voice-over announcer invariably says “2010 Census.”

This strikes me as a weird inconsistency, not what one usually sees in a presumably well-considered, well-heeled campaign. In other words, “Huh?” What can you find out?

A: (Stuart Elliott)

Back we go, dear reader, to Mr. Petersen, who offers this reply in another e-mail message: “The formal name, ‘United States Census 2010,’ is a mouthful and sounds too bureaucratic. Lots of brands have nicknames. Look at Mickey D’s, for McDonald’s; B.K., for Burger King; and the Shack, for Radio Shack.”

“Saying ‘2010 Census’ simply functions as a short and memorable nickname,” Mr. Petersen concludes.

A New York Times article about Haitian families “doubling up” in the wake up the January earthquake highlights yet another issue that census-takers will have to deal with…primarily in Florida, New York, and other areas with large Haitian communities.

In other areas, such as Cleveland, Ohio — a city that some institutions suspect was undercounted in 2000 — the financial crisis and subsequent loss of jobs has resulted in extended families living temporarily…or long term…with one another. Nobody ever said enumeration would be easy…

The following comes from Justin Wolfers, a professor at my alma mater, the University of Pennsylvania, who writes for The New York Times’ Freakonomics blog and answers the age old question… “Can you trust the census?”:

No. At least that’s the conclusion of an important new paper (ungated versionhere) by Trent Alexander, Michael Davern and Betsey Stevenson, who find enormous errors in some critically important economic datasets.

Let’s start with the 2000 Decennial Census. Your responses to the Census were used for two purposes. First, the Census Bureau tallied up every response to produce its official population counts. And second, it produced a 1-in-20 sub-sample of these responses, which it made available for analysis by researchers. Just about every economist I know has used this Census sub-sample, as do a fair number of demographers, sociologists, political scientists, and private-sector market researchers.

The errors are documented in a stunningly straightforward manner. The authors compare the official census count (based on the tallying up of all Census forms) with their own calculations, based on the sub-sample released for researchers (the “public use micro sample,” available through IPUMS). If all is well, then the authors’ estimates should be very close to 100% of the official population count. But they aren’t:

Source: Inaccurate Age and Sex Data in the Census PUMS Files: Evidence and Implications
Trent Alexander, Michael Davern and Betsey Stevenson

The two estimates are pretty similar for those younger than 65. But then things go haywire, with the alternative estimates disagreeing by as much as 15%. In fact, the microdata suggest that there are more very old men than very old women — I know some senior women who wish this were true! The Census Bureau has confirmed that the problem isn’t with the authors’ calculations. Rather, the problem is in the public-use microdata sample.

What’s the source of the problem? The Census Bureau purposely messes with the microdata a little, to protect the identity of each individual. For instance, if they recode a 37-year-old expat Aussie living in Philadelphia as a 36-year-old, then it’s harder for you to look me up in the microdata, which protects my privacy. In order to make sure the data still give accurate estimates, it is important that they also recode a 36-year-old with similar characteristics as being 37. This gives you the gist of some of their “disclosure avoidance procedures.” While it may all sound a bit odd, if these procedures are done properly, the data will yield accurate estimates, while also protecting my identity. So far, so good.

But the problem arose because of a programming error in how the Census Bureau ran these procedures. The right response is obvious: fix the programs, and publish corrected data. Unfortunately, the Census Bureau has refused to correct the data.

The problem also runs a bit deeper. If the mistake were just the one shown in the above graph, it would be easy to simply re-scale the estimates so that there are no longer too many, say, 85-year-old men — just weight them down a bit. But it turns out that the same coding error also messes up the correlation between age and employment, or age and marital status (and, the authors suspect, possibly other correlations as well). When you break several correlations like this, there’s no easy statistical fix.

The following comes from the venerable New York Times:

Published: October 19, 2009

With the start of the 2010 census just a few months away, Senator David Vitter, a Republican of Louisiana, wants to cut off financing for the count unless the survey includes a question asking if the respondent is a United States citizen. Aides say he plans to submit an amendment to the census appropriation bill soon.

As required by law, the Census Bureau gave Congress the exact wording of the survey’s 10 questions in early April 2008 — more than 18 months ago. Changing it now to meet Mr. Vitter’s demand would delay the count, could skew the results and would certainly make it even harder to persuade minorities to participate.

It would also be hugely expensive. The Commerce Department says that redoing the survey would cost hundreds of millions of dollars: to rewrite and reprint hundreds of millions of census forms, to revise instructional and promotional material and to reprogram software and scanners.

During debates in the Senate, Mr. Vitter said that his aim is to exclude noncitizens from population totals that are used to determine the number of Congressional representatives from each state. He is ignoring the fact that it is a settled matter of law that the Constitution requires the census to count everyone in the country, without regard to citizenship, and that those totals are used to determine the number of representatives.

(The Census Bureau already tracks the number of citizens and noncitizens through a separate survey.)

Adding a new question about citizenship would further ratchet up suspicions that the census is being used to target undocumented immigrants. That would discourage participation not only among people who are here illegally but also their families and friends who may be citizens and legal residents. That leads to an inaccurate count.

And since census numbers are also used to allocate federal aid, undercounting minorities shortchanges the cities and states where they live.

Advocates for the census say that Senator Orrin Hatch, a Republican of Utah, has also raised the idea of another bad, last-minute change. Under current practice, the only people living abroad included in the census are military personnel and federal civilian employees, and the families of both, stationed overseas. Mr. Hatch, these officials say, wants to include certain other Americans living abroad temporarily, a definition that would be tailored to include — you guessed it — Mormon missionaries.

There seems little doubt that the goal would be to increase population numbers for Utah — to try to garner another Congressional seat. As of Monday, Senator Hatch’s office would not say whether he plans to pursue the idea. He shouldn’t.

Both of these changes would be discriminatory and ridiculously expensive. If Mr. Vitter and Mr. Hatch wanted to argue their cases, they should have done it 18 months ago — or wait until after this count.

Changing the survey now would be a disaster for the census and for American taxpayers. The Senate should defeat any and all attempts to alter or delay the 2010 count.

At this moment, the #1 most e-mailed New York Times article is called “For Today’s Graduate, Just One Word: Statistics.” To all of the statisticians out there, this one’s for you.