2012-2016 Language - no local geographies

OMG you're right! Why wasn't this noted on the modified tables page? I'm going to have to redo all of our 2012-2016 profiles.

It's not mentioned in the tech doc either: www2.census.gov/.../2016_SummaryFile_Tech_Doc.pdf

The State Data Center and Census Information Center networks received a note about on Nov. 30. I've copied it at the end of this message. This was the only source I saw; the link to the language note doesn't mention anything about geographic restrictions in the five-year data. (It's also in the table note in American Factfinder, but of course you can't get to a table that's not available for your selected geographies!)

If it helps, language data are still available for PUMAs, which in some cases can be aggregated to yield Table B16001 for counties or cities. This is still a big gap, though, since we in the Twin Cities have lost almost all local-level data on three of our six most commonly spoken language groups other than English (African languages collectively, Hmong, and Russian).

I'm really hoping the Census Bureau can figure out a less heavy-handed approach to disclosure avoidance for next year's 2013-2017 release. Does anyone know if that's a possibility?


Here's the note to SDCs/CICs:

Next week the 2016 ACS 5-year data will be released. With this release, table B16001 will no longer be available at the tract level. This is due to new Disclosure Review Board restrictions applied to B16001 only. These have been applied out of concern for respondent privacy protection. All of the other language tables will still be available at the tract level, but do not give as high a level of detail for as many specific languages.


Table note:

Additional geographical restrictions have been applied to Table B16001 - LANGUAGE SPOKEN AT HOME BY ABILITY TO SPEAK ENGLISH FOR THE POPULATION 5 YEARS AND OVER for the 5-year data estimates. These restrictions are in place to protect data privacy for the speakers of smaller languages. Geographic areas published for the 5-year B16001 table include: Nation (010), States (040), Metropolitan Statistical Area-Metropolitan Divisions (314), Combined Statistical Areas (330), Congressional Districts (500), and Public Use Microdata Sample Areas (PUMAs) (795). For more information on these geographical delineations, see www.census.gov/.../geographic-reference-files.html. County and tract-level data are no longer available for table B16001; for specific language data for these smaller geographies, please use table C16001. Additional languages are also available in the Public Use Microdata Sample (PUMS), at the State and Public Use Microdata Sample Area (PUMA) levels of geography.
Here is an example of a B16001 table at tract level (in Iselin, NJ). Notice the break-out of South Asian languages:

factfinder.census.gov/.../1400000US34023001409

Here is a link to the technical documentation PDF:

www2.census.gov/.../2016_Language_User_Note.pdf

That's a terrible way to communicate a fairly major change! Like most of you (probably), I get dozens of emails a day.

Note that the SDC email says B16001 will no longer be available at the tract level, but the table note says "County and tract-level data are no longer available for table B16001." 

Can we please have all such changes posted to the "Table and geography changes" page?

I sent the original note above to my SDC group in NY state to let them know what was about to happen. Sorry that I didn't word everything exactly correctly. It was simply posted to a listserv, and then I guess sent to other SDC's. I only found out because a colleague here in Northern NJ said he had found out about the language suppression with this release. I do outreach here in the greater NYC area, and this loss of data is not good for many who need it in this area. We have a large South Asian community in NJ, and the loss of the language variations is problematic.

Thanks for bringing this to our attention. C16001 is better than nothing, but it doesn't calibrate to the number of speakers of different languages in different geographies. We have more Hindi and Japanese speakers than Vietnamese, and now we will no longer know which areas of the City need materials translated into these languages. I understand the privacy issue, but where there isn't a privacy issue, the data should be made available. I guess it comes down to not being able to suppress data at the scale of an estimate. Why isn't that possible currently?

What would be really nice is a table that lists the top five or ten languages other than English spoken in a given tract. That table structure could be standardized, but Census tables don't usually have text for values, and of course we would want the estimates along with the names of the languages. Is a solution possible?

Extremely disappointed that B16001 won't be available for small areas. The consolidated C16001 is available, but only has 12 non-English languages over the detailed 42 non-English.

I came across this problem yesterday and was quite surprised in light of a Census Bureau press release that featured the break out of additional languages in this table. I'll be able to work around this problem once the PUMS data is available, but it is very disappointing in light of the complicated fabric of languages used in many areas.

Cliff Cook

A solution would be possible if (1) we could have locally-defined tract-aggregated geography for areas of 65K+ population, and (2) the bureau would agree to publish data for characteristics (such as language) reported by a specific defined population, such as 5%, or e.g. 200, speakers of that language. This requires programming that the bureau has not done in a long time. We did once get such data for commuting and/or migration from specific nearby cities or counties, but that was a long time ago (1980?).

Since I started this thread, since Census people read this, I'd just add that it's amazing that this much data gets released with the quality it does every year, particularly with the challenges the bureau is facing right now. We'll always want more and have complaints, but overall they've done a very good job.

Thank you for your feedback about the lack of documentation on these language table changes.  We updated the following pages to document the geographic restrictions for Table B16001:

• 2016 Language Tables Update User Note
• Tables & Geography Changes
• 2012-2016 ACS Summary File Tech Doc 5-year Appendices (update available by Monday)

Does anyone know if for 2013-2017, this data will be brought back at local geographies?

I checked with the subject matter staff, and the data will not be brought back at local geographies. They pointed me to the following table note:

Geographical restrictions have been applied to Table B16001 - LANGUAGE SPOKEN AT HOME BY ABILITY TO SPEAK ENGLISH FOR THE POPULATION 5 YEARS AND OVER for the 5-year data estimates. These restrictions are in place to protect data privacy for the speakers of smaller languages. Geographic areas published for the 5-year B16001 table include: Nation (010), States (040), Metropolitan Statistical Area-Metropolitan Divisions (314), Combined Statistical Areas (330), Congressional Districts (500), and Public Use Microdata Sample Areas (PUMAs) (795). For more information on these geographical delineations, see the Metropolitan Statistical Area Reference Files. County and tract-level data are no longer available for table B16001; for specific language data for these smaller geographies, please use table C16001. Additional languages are also available in the Public Use Microdata Sample (PUMS), at the State and Public Use Microdata Sample Area (PUMA) levels of geography.

Thanks for the quick response!