[CTPP] ACS 2005-2007 Population Estimates

[Patty Becker]

The first thing to understand is that while the ACS county estimates 
are controlled to the census population estimates, the sub-county 
estimates are not.  The numbers that appear for cities are just what 
the weighted data show. For Detroit, they are always too low.  There 
is always a lag in terms of getting new housing units into the 
sampling frame (the Master Address File, or MAF), while there is loss 
when a sample address turns out to be demolished or vacant.

I don't know if that the 2010 ACS data are going to be weighted to 
the census results. I don't think they've thought about that yet.  If 
they are, there won't be any problem updating from April to July. We 
do not usually see redone intercensal estimates after the census, and 
I'm sure that there won't be any adjustment of ACS results.

Bottom line: be very careful in using whole numbers from the ACS, and 
when you do, always round them to '00s (hundreds) so that people will 
understand that they are estimates and not counts. ACS is really best 
for the data that are expressed in percentage terms.

Patty Becker




At 05:28 PM 6/25/2009, you wrote:
>For ACS 2005-2007 population and housing occupancy/vacancy 
>estimates, are people finding discrepancies between ACS and other 
>data sources, particularly in sub-county geographies?  ACS estimates 
>are controlled at the county level and, like the ACS 1-year data, 
>ACS 2005-2007 population estimates are showing disagreement with 
>other data sources.  For example, the Oakland 3-year ACS estimate 
>shows the population at 372,000, when CA Dept. of Finance estimates 
>are over 400,000.  This runs counter to the on-the-ground anecdotal 
>experience - ACS shows a loss of 30,000 people during a period that 
>showed an increase of 10,000 housing units.
>
>Will the decennial census correct this?   My understanding is that 
>Census 2010 numbers will be used to control 2010 ACS characteristic 
>data.  There will need to be some adjustment, however, given that 
>the decennial census benchmarks population at April 1, and ACS uses 
>a July 1 population number.   It's also my understanding that 
>population estimates for 2001-2009 will be updated, though 
>retroactive adjustments for ACS characteristics will not be 
>done.  Does anyone else know something different about this?
>
>Other thoughts about this?
>
>Thanks,
>
>Shimon
>
>---------------------------------------------------------------
>Shimon Israel
>Associate Transportation Planner/Analyst
>Metropolitan Transportation Commission
>101 Eighth Street
>Oakland, CA 94607-4700
>(510) 817-5839 (office)
>(510) 817-5848 (fax)
>---------------------------------------------------------------
>
>
>_______________________________________________
>ctpp-news mailing list
>ctpp-news at chrispy.net
>http://www.chrispy.net/mailman/listinfo/ctpp-news

-------------------------------------------------------------------------------------------------
Patricia C. (Patty) Becker         248/354-6520
APB Associates/SEMCC       FAX 248/354-6645
28300 Franklin Road                 Home 248/355-2428
Southfield, MI  48034                     pbecker at umich.edu 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.chrispy.net/pipermail/ctpp-news/attachments/20090626/682fdaf4/attachment.html