This article published today on language access issues with the US Census is pretty damning:
http://www.globalwatchtower.com/...
The US Census is completely ignoring swaths of the population because of its very limited language access, rendering some statistics likely invalid and losing a chance to gather information from many minority areas. And even when it tries to get it right, it's missing the target and wasting money.
For example, the government is spending $133 million on PR campaigns to get people to fill it out, but only $51k (yikes!) on language services to make it accessible to non-English speakers. That's leading to major mistakes - some of them rendering their efforts to reach out utterly pointless.
Examples below the fold...
Even though there are 168,438 U.S. residents who speak Navajo at home, the government does not offer a form in this language. Navajo speakers could try to read an online guide in their language — but thanks to a botched localization job, all they are likely to see is a message that tells them they might need to download special fonts. Oh, and the message that tells them this? It’s only in English, not Navajo.
News flash to the US Census Bureau: there are quite a few people who live here and who can't speak, read or write in English! Oh wait, perhaps you knew that already?
Of the 44.9 million people in the United States who do not speak English at home, says the Census Bureau, 19.5 million speak English less than "very well."
Quite right, couldn't agree with you more! Pity you didn't figure out how to get more of those people to participate in the Census, eh?
Where the government did spend money making the Census available to those who can't read and write in English, they didn't exactly spend their money wisely:
The 2010 Census for the United States is available in five languages: Chinese (simplified), Korean, Russian, Spanish, and Vietnamese. Why these five languages? Apparently, they were determined to be the languages of highest demand, even though the data from the 2005 American Community Survey indicates that more people speak Tagalog and French.
And it's not like things are just taking a long time to get better. It's actually getting worse:
In the past, the U.S. Census asked individuals if they spoke a language other than English at home. No longer. The 2010 Census uses a short form, in which there is no question about language use. India takes the opposite approach - all languages and mother tongues with populations of 10,000 and above are recorded.
Perhaps we need to outsource our US Census Bureau operations to India? Seems like they know how to poll an entire multilingual population correctly!
Of course the Census debates have been pretty heated this year with every kind of nutty conspiracy theory abounding, regardless of how silly and counter-productive those ideas may be. The fact remains, however, that like any other statistical analysis the census becomes becomes more valuable and accurate the more people participate. But because of a few basic steps which weren't taken, the US Census is going to be far from it.