Where to draw the line? Probably comes down to what's your tolerance for error in
informing decisions you want to make...
Earlier in my career, I worked in a State labor market research shop. In survey work for
US BLS, we would only publish survey-based stats where relative standard error (std error
as % of the central estimate) was less than 30% (for average wage) or less than 50% (for
employment levels).
The wage statistics had a more stringent standard than employment levels because of the
nature of the data (there was less variation and uncertainty around wages) and because of
perceived criticality (data user audiences expected and needed greater accuracy).
Hope that helps.
--Todd Graham
Metropolitan Council Research
From: ctpp-news-bounces(a)chrispy.net [mailto:ctpp-news-bounces@chrispy.net] On Behalf Of
Mara Kaminowitz
Sent: Wednesday, December 22, 2010 10:11 AM
To: ctpp-news(a)chrispy.net
Subject: RE: [CTPP] ACS 2005-09 data
What I would like to know is how much MOE is too much to be useful? I have been looking at
the census tract MOE's for certain variables in my region and they range from 25% to
200%. It's not pretty. And that's not even the obscure sub-categories. Where do
other people draw the line when determining data reliability?
Mara
Mara Kaminowitz
GIS Analyst
.........................................................................
Baltimore Metropolitan Council
Offices at McHenry Row
1500 Whetstone Way
Suite 300
Baltimore, MD 21230
410-732-0500 ext. 1030
mkaminowitz@baltometro.org<mailto:mkaminowitz@baltometro.org>
www.baltometro.org
<http://www.baltometro.org>