The ACS survey is “Random” rather than “Scientific”

More coverage on the ACS. Most of the press coverage for the ACS is positive.

GOP Rep Daniel Webster Bashes Census Survey As “Random” Rather Than “Scientific”
Matthew Iglesias | Slate
May 20, 2012

The title says it all.

[Selected Excerpts]
[Paragraph 2] In reality, of course, what would not be cost effective would be to try to survey every single American on such a frequent basis. By making the sample random the Census Bureau is able to make scientifically valid inferences about the state of the overall population. It’s a little bit sad that any member of congress doesn’t understand this, but effective representative democracy (thankfully) doesn’t require members of congress to be well-informed about every subject under the sun. It does, however, require members of congress to try to be well-informed about the sub-set of issues they choose to become active on. Webster is badly, badly failing that test here.

The Beginning of the End of the Census?
Catherine Rampell | New York Times
May 20, 2012

Editorial also includes a nice graphic that shows the history of questions in the Census/ACS since 1850.

The upshot of this editorial is that the likely outcome of a Senate/House appropriations bill is that the ACS will be voluntary rather than mandatory. While that is better than a de-funded ACS and Economic Census, the quality of the data are likely to suffer and the ACS will be more expensive.

[Selected Excerpts]
Paragraph 1]THE American Community Survey may be the most important government function you’ve never heard of, and it’s in trouble.

[Paragraph 24]If the American Community Survey were made voluntary, experts say, the census would have to spend significantly more money on follow-up phone calls and in-person visits to get enough households to answer.

[Paragraph 25]But Congress also plans to cut the census budget, making such follow-ups prohibitively expensive.

[Paragraph 26]“If it’s voluntary, then we’ll just get bad data,” said Kenneth Prewitt, a former director of the census who is now at Columbia University’s School of International and Public Affairs. “That means businesses will make bad decisions, and government will make bad decisions, which means we won’t even know where we actually are wasting our tax dollars.”

Census surveys aren’t down for the count
Steven Dinan | Washington Times
May 20, 2012

In a previous post, the Washington Times was a bit more positive in its coverage of Webster, et.al. This article is more muted but covers most of the Republican talking points: the questions are not constitutional; the census shouldn’t enumerate non-citizens; maybe businesses could pay for the data [no discussion of other interested users like the public].

Oddly enough, the article sounds like it would be in favor of the use of statistical methods for correcting for the problems of a voluntary ACS:

Some lawmakers in the House have even floated the idea of making businesses pay for the data as a way of recouping costs, while others said the Census Bureau — which has occasionally advocated using statistical sampling for the decennial census — could make use of that technique to correct for any errors in a voluntary census.

Without survey, guess who suffers?
Mary Sanchez | Kansas City Star
May 20, 2012

A regional newspaper provides an excellent summary of the issue.

[Selected Excerpts]

[Paragraph 1]Here’s a fiscal mess in the making that has slipped by most Americans.

[Paragraph 2]This one is pushed by the folks who claim that government is too big and intrusive.

[Paragraph 3]The U.S. House of Representatives is attempting to eliminate funding for an annual socioeconomic survey of the U.S. population. Sounds like a little harmless nipping at the overblown federal budget, right?

[Paragraph 9]Clearly, this is an issue where current political mantras overtook rational thinking. The attitude driving this ill-considered move is that the government has no business asking questions about the ages of your household members, income, marital status, military service, mortgage, rent payments or other personal information.

[Paragraph 10] But without the data, government funding would be based on guesswork. That is hardly a fiscally responsible approach.

0 Responses to “The ACS survey is “Random” rather than “Scientific””


Comments are currently closed.