Fellow CTPP afficionados:
The MOE reported by Census is based on a 90% confidence interval, which is a common standard for scientific research -- it means that the confidence interval has a 90% likelihood of including the real value (aka 9 to 1 odds). The larger the MOE in relation to the estimated value, the less reliable the estimate. Relatively large MOEs do not necessarily mean the estimates are "worthless," as Ed notes. It all depends on your needs and purposes.
It is good to strive for scientific certainty. But in the world of public policy-making, transportation planning, etc., we normally are obliged to make decisions with whatever is the best available data. We can't usually throw up our hands in despair if scientific certainty is not available. I'm retired now, but what I used to do when confronted with this quandary was to calculate just exactly how unreliable an estimate was before deciding what to do about it.
A quick and dirty alternate way to look at it is to divide the MOE by 1.645 to get the Standard Error of the Estimate (SEE). The SEE represents the standard deviation around the estimate, which has a 68% likelihood of including the real value, or about 2 to 1 odds. If you look at the estimate and its SEE rather than its MOE, are you willing to take a deep breath and make needed decisions based on 2 to 1 odds, rather than wait for 9 to 1 odds?
For another
example, I might have a number from ACS or PUMS that wasn't
significant at the 90% confidence level (e.g., has commuting between county A
and county B actually gone up between year x and year y, or is it more
accurate to say it was stable?). Using a standard built-in
formula in Excel (NORMDIST) I could determine
that the increase was significant at the 80% level. From my
point of view, 80% confidence (or maybe even 70%) might be fine, depending
on the nature of the question. Your mileage may vary.
--
Pete Swensson
retired from Thurston Regional Planning Council
Olympia, WA
Quoting Ed Christopher
Do we have a metric for useless? I have seen published data where the MOE is more than 85% of the estimate. Is this useless? What if the MOE is only 10% of the estimate. Anyone using any standards for what is considered useless. Also, at what point should this determination be put in the hands of the user/analyst?
Ed CBrian,
Tell me something. Do you want to use these data for individual block groups or do you want to aggregate BGs to larger areas, such as TAZs? I hope you know that the sampling errors for single BG data are so high as to render the data absolutely useless. I think that's why you can't find the numbers.
Patty BeckerOn Fri, Feb 15, 2013 at 3:35 PM, Brian Chenault <bchenault@hrtpo.org> wrote:Hello,
Brian Chenault from the Hampton Roads Transportation Planning Organization here. We are trying to collect several pieces of data at the block group level using 2007-2011 ACS 5-year estimates. Using the Census’ Summary File Excel Retrieval Tool, we are able to retrieve basic population data at the block group level. However, data at the block group level is not provided for poverty, language spoken at home, etc. This appeared to be the case even when examining 2006-2010/2005-2009 5-year estimates. Has this information not been “filtered down” from the census tract level yet? How is one supposed to go about gathering this type of updated information at the block group level?
Much appreciated,
Brian
Brian Chenault
Community Outreach Planner
Hampton Roads Transportation Planning Organization
The Regional Building | 723 Woodlake Drive | Chesapeake, Virginia | 23320
bchenault@hrpdcva.gov | http://www.hrtpo.org | Phone: 757.420.8300 | Fax: 757.523.4881
_______________________________________________
ctpp-news mailing list
ctpp-news@ryoko.chrispy.net
http://ryoko.chrispy.net/mailman/listinfo/ctpp-news
--Patricia C. (Patty) Becker
APB Associates/Southeast Michigan Census Council (SEMCC)
28300 Franklin Rd, Southfield, MI 48034
office: 248-354-6520
home:248-355-2428
pbecker@umich.edu_______________________________________________
ctpp-news mailing list
ctpp-news@ryoko.chrispy.net
http://ryoko.chrispy.net/mailman/listinfo/ctpp-news
_______________________________________________
ctpp-news mailing list
ctpp-news@ryoko.chrispy.net
http://ryoko.chrispy.net/mailman/listinfo/ctpp-news