Re. the Census: Will Republicans Suddenly Discover Statistical Sampling?

Republicans are in a bit of a bind regarding the U.S. Census. In the past, they have opposed "statistical sampling", which would readjust the Census results to account for undersampled groups, such as the poor and minorities, which typically vote Democratic. In fact, congressional Republicans made Obama's Census director pinky-swear not to do this, and he did:

Robert Groves, director of the University of Michigan's Survey Research Center and a former Census Bureau official, is an expert on statistical sampling, the practice of extrapolating a larger population from a smaller slice of it. Proponents of sampling say it helps produce a more accurate tally of the population, especially when it comes to traditionally undercounted groups, such as minorities living in urban areas.

But many Republican lawmakers insist that sampling violates the Constitution, which calls for an "actual Enumeration" of the population every 10 years. Critics also say the use of sampling would politicize the traditionally nonpolitical Census Bureau.

Dr. Groves, during his confirmation hearing Friday, told members of the Senate Homeland Security and Governmental Affairs Committee that he wouldn't use sampling to adjust the 2010 count. Asked whether he would consider using it in a future census, he said: "There are no plans to do that for 2020."

Of course, Republicans being Republicans, this wasn't good enough. Many Republicans, and their conservative media Wurlitzer (or is it the other way around?), denigrated the entire concept of the Census:

Prominent right-wing voices decided last year that the U.S. census was not to be trusted. Rep. Michele Bachmann (R-Minn.) said the process could lead to "internment camps." Rep. Ron Paul (R-Texas) called the census "invasive." Fox News' Glenn Beck's suggested Americans may not be comfortable with "ACORN members" collecting information. Radio host Neal Boortz said some census information is "designed to help the government steal from you in order to pass off your property to the moochers."

Unfortunately for Republicans, not only did their political televangelists use this ridiculousness to move product, but the rubes actually took it seriously! Uh-oh:

Polling by the Pew Research Center finds Democrats are more likely than other Americans to view the census as "very important" to the country. Seventy-six percent of Democrats call this year's count very important, compared with 61 percent of Republicans and independents.

In Texas, some of the counties with the lowest census return rates are among the state's most Republican, including Briscoe County in the Panhandle, 8 percent; King County, near Lubbock, 5 percent; Culberson County, near El Paso, 11 percent; and Newton County, in deep East Texas, 18 percent. Most other counties near the bottom of the list are heavily Hispanic counties along the Texas-Mexico border.

There is a reason for the enthusiasm gap on the census: A number of prominent conservative and libertarian Republicans have been blasting the census for months.

It's estimated that each undercounted person will cost Texas $12,000. Part of me wants to argue for fairplay--the Census is very important and should be completed (and if you haven't done so yet, do it. Seriously, stop reading this, and do it right now). On the other hand, it is refreshing to see the blowback from all of the rightwing lunacy: they're intentionally screwing themselves over, no ACORN needed. Not only does the Census help determine funding formulas, but it's also used to apportion House representatives.

If this pattern holds, I wonder if Republicans will suddenly see the virtues in statistical sampling....

More like this

But many Republican lawmakers insist that sampling violates the Constitution, which calls for an "actual Enumeration" of the population every 10 years. Critics also say the use of sampling would politicize the traditionally nonpolitical Census Bureau.

IIRC this point was litigated last time around, and the Supreme Court upheld the Republican position that "actual Enumeration" meant "no statistical sampling". Would it be wrong of me to enjoy some schadenfreude if the Republicans are hoist on their own petard?

Oddly enough, I have heard that among the Republican-held House districts with the highest response rate is the Minnesota district represented by Michelle Bachman. Maybe her constituents have been paying attention to state officials' warnings that Minnesota was at risk of losing a seat, probably Bachman's, if their census count came in low.

By Eric Lund (not verified) on 05 Apr 2010 #permalink

Schadenfreude seems definitely called for. The sad thing? This doesn't even rise to "hoisted by their own petard" territory - this is just plain, old "didn't think matters through".

And not thinking matters through has become the hallmark of the new GOP rank-and-file. After all, they've been brought up on the simplicity of "conservative principles" - why even look at a proposed government program in detail, when it's a law of nature that everything connected to government is inevitably detrimental?

Of course, by now GOP leadership begins to realize that that's one genie they can't get back into the bottle - they no longer control the message, they have no way of stopping the mutated spawn of their indoctrination from carrying the message to ever further extremes. At the moment, they're hastily preparing a PSA featuring Karl Rove in support of the census, while at the same time Erik Erikson talks of chasing census takers away with his shotgun.

By Phillip IV (not verified) on 05 Apr 2010 #permalink

Whatever you want the Republicans to do or support, just make its opposite look like a Democratic idea, and they will fall over themselves to support it.

So if you want them not to turn in their census so they lose representation in Congress, just tell them Democrats support the census, and Acorn is doing all the counting. I wonder how many house seats Texas would lose?

Of course we don't really want that, but it sure would be fun.

@ BaldApe , seriously I'm really not at all sure that we don't really want this. At the very least *I* will not be bothered in the slightest by Texas losing house seats and money because of something like this.

The only real downside is that I think census numbers are valuable to economists and other researchers. But of course they know how to correct for bad data (which of course doesn't mean good data isn't better).

By Timothy Underwood (not verified) on 05 Apr 2010 #permalink

I have no problem with the Rethuglican party screwing themselves over on the census. I just wish it would have an effect in this Fall's election.

By DGKnipfer (not verified) on 06 Apr 2010 #permalink

Living in Utah, shortly before the 2000 census, I encountered a great deal of anti-census rhetoric. Most of the conservative Utahns I knew were spouting it. The few non-conservatives I knew were trying to convince conservative Utahns that participating in the census was a good thing. Utah ended up being 857 (later shortened to 80) people short of gaining a 4th Representative. Utah, of course, claimed some Mormon missionaries were not being correctly assigned to their proper state of residence. Probably true, but it is also likely that anti-census rhetoric contributed to the count being just short of what was needed.

The problem is not cooperation. There are thousands and thousands of reports of people not receiving the census form.
I have talked to hundreds of people who have not been surveyed or counted in 40 years.
If the population is sampled then the population will be undersampled and then undersampled again.
This is what has happened since 1970:
Simplified: There were 9 samples in a county in 1970 and all were rural white ranch families living miles apart. The "readjusters" would say..OK, there must be at least one impoverished family, one black family working on the railroad and one hispanic ranch worker family (undercount) there are 9 white families, and 3 minority and or impoverished families. Population 12 families x 4 guessed family size= 48 people adjusted for undercount and population density.
Now..there is a town in the middle of all this that was a store and gas station, missed in the sample..but before the next is developed into a modern city and has a population of thousands..but the census did not and is not sampling that areas. The census is still working off the original..24 people..and the increase ..birth, death, immigration. And the count is predicated on 1/3 poor/minority which was a statistical fiction in the first place..
Now..reality..6 of the white families have not been counted in 40 years because they have post office boxes and don't get census what is the population of the county?
There are probably 400 million people in this country and that is a fact..