Re: [CTPP] how to know if Census ests and ACS have been under-estimating the population?

After sending an earlier CTPP post this morning, I was asked “*how do you know Census Bureau “badly underestimated” the top-line population totals in your metro?*” Great question… The latest news is: Census Bureau has coughed up the answer for us. This morning CB has published side-by-side comparison of Census enumeration counts vs. the Bureau’s annual estimates (from a model). You can find it at http://www.census.gov/popest/states/tables/table-2.xls It looks like there are 26 states where the Bureau’s model has been under-estimating population by > 0.5%. And 10 states where the model has been over-estimating population by > 0.5%. * Of course, my tally here assumes that the Census counts are “perfect” and free of under-count. And maybe it’s too early to make such assumption. (Biggest discrepancy: Arizona… Hmm.) Anyway, check it out at http://www.census.gov/popest/states/tables/table-2.xls Enjoy. ________________________ Todd Graham Principal Forecaster Metropolitan Council 390 Robert Street North Saint Paul, MN 55101 ph: 651/602-1322 email: todd.graham(a)metc.state.mn.us in: www.linkedin.com/in/toddgraham web: http://www.metrocouncil.org/data/ ________________________ -----Original Message----- From: ctpp-news-bounces(a)chrispy.net [mailto:ctpp-news-bounces@chrispy.net] On Behalf Of Graham, Todd Sent: Wednesday, February 02, 2011 8:47 AM To: ctpp-news(a)chrispy.net Subject: RE: [CTPP] 2000 and 2006-08 work trip comparison question Hi Frank-- Hope the new year is treating you well. You asked what you're missing: How is the 3-year or 5-year window *not* like a moving average? Some of the moving parts changing from year to year are the control totals and, as result, the weights (expansion factor) that would be attached to any given household found in the 5 annual sub-samples that comprise the period sample. The population control totals come an exogenous, completely-outside-the-survey-itself, annual estimation model. And in our region, Mpls-St Paul, CB badly underestimated the top-line population totals during early and mid-decade. I think they realized it (oops), and they've been trying to make up ground ever since. Also pertaining to control totals, there may be a time-series disruption for small areas with the vintage 2009 data. CB transitioned from having county-level control totals to having sub-county control totals. And that's likely to cause weird time-series breaks for some cities where population was previously under- (or over-)estimated. ... Expect more fun this Fall, when vintage 2010 ACS are re-benchmarked to 2010 Census Enumeration ________________________ Todd Graham Principal Forecaster Metropolitan Council 390 Robert Street North Saint Paul, MN 55101 ph: 651/602-1322 email: todd.graham(a)metc.state.mn.us in: www.linkedin.com/in/toddgraham web: www.metrocouncil.org ________________________ ________________________________________ From: ctpp-news-bounces(a)chrispy.net [ctpp-news-bounces(a)chrispy.net] On Behalf Of Frank Lenk [FLENK(a)MARC.ORG] Sent: Tuesday, February 01, 2011 10:34 PM To: ctpp-news(a)chrispy.net Subject: RE: [CTPP] 2000 and 2006-08 work trip comparison question. Call me dense, but I don’t see how this problem is any different than working with moving averages. I mean, I do understand that this data is not an average but a period estimate. Still, the issues created by dropping the first year of the period and adding the last year as the data series moves forward in time seems the same to me. What am I not understanding properly? Frank Frank Lenk Director of Research Services Mid-America Regional Council 600 Broadway, Suite 200 Kansas City, MO 64105 www.marc.org<http://www.marc.org> 816.474.4240 flenk@marc.org<mailto:flenk@marc.org> 816.701.8237 _______________________________________________ ctpp-news mailing list ctpp-news(a)chrispy.net http://www.chrispy.net/mailman/listinfo/ctpp-news