The Boston Globe:

The proposed boycott – organized this spring by the National Coalition of Latino Clergy and Christian Leaders, a group based in Washington that represents 20,000 churches nationwide, including 300 in Massachusetts – is stirring deep divisions among immigrant communities. It faces stiff opposition from a string of advocacy groups, including the Massachusetts Immigrant and Refugee Advocacy Coalition, the Service Employees International Union, and the Brazilian Immigrant Center. To this point, the boycott effort has revolved mainly around word-of-mouth, talk radio, and blog entries by some members of participating churches.

And Wisconsin Public Radio has the same story.

It’s been the subject of vigorous debate in the Spanish-language press since at least February. The outreach people I talked to most recently were bemoaning the fact that some politicians are telling people not to participate. I would have assumed that they meant the right-wing ones, but the outreach area they work in is heavily Democratic, so who knows.

That’s not so much a ridiculous margin as it it a ridiculous interpretation of a distribution that we know won’t be quite normal because of the hard minimum. Sure, it may be more precise to state the upper and lower bounds of the error separately rather than as an error range, but I’d be willing to bet that someone has to stuff the data into a table that couldn’t accommodate the additional field.

and the left? …really? where?

I’m probably a bit touchy on this because I keep talking to enumerators who are running into anti-Census bias. It’s so easy for stuff like this to segue into “Ha HA, isn’t New Jersey corrupt, and by the way you can’t trust Obama’s census either,” and then we see something like what happened with ACORN last year, where voter registration applications that they flagged as potentially invalid somehow get defined in the US media as “ACORN trying to commit voter fraud.”

Yes, I know all this. My first thought was that the housing unit estimation method would have led to the error, but then I saw the bit at the top of the entry that the Census had just missed a bunch of people in the enumeration to begin with.

As of the census of 2000, there was one married couple living in the village. Both were white and above 65, and their median age was 67 years. Their household income was $11,250, or $6,167 per capita. The population density was 9.4 people per square mile (3.7/km²). There were 4 housing units at an average density of 18.8/sq mi (7.4/km²).

It’s very common in the ACS to see ridiculous margins of error in subpopulations—e.g., foreign born in County X are 42, with a margin of error of 106. What, there are negative 64 people living there? Really?

I know this is a humorless response, but I’ve talked a number of reporters out of relying on such baloney numbers, so I worry when I see things like this repeated uncritically. Especially in an environment in which there is boiling rhetoric from both the right and the left against the upcoming 2010 Census.