Transcript Document

But what about the ACS?
What is the American
Community Survey?
Replacement of the Long Form
Continuous Survey Methodology
Conducted Monthly
According to the Census Bureau
The ACS is…
A large, continuous demographic survey
• Produces annual and multi-year estimates of
the characteristics of the population and
housing
• Produces characteristics, not a population
count
• Produces information for small areas
including tracts, block groups and
population subgroups - updated every year
• Key component of reengineered 2010
census
ACS Facts
About 1 in 40
Households Sampled per year
250,000 Households Sampled
per month
Same Transportation questions
as 2000 Long Form
Decennial Long Form is History, Kaput, Gone!
But why change and have
an ACS?
Eliminate the Census Long Form for
2010 and Thereafter
Eliminate the peak in Congressional
funding
Eliminate the peak in hiring and training
Census takers
How does the ACS work?
Three Methods of Data Collection
1. Mail
Self-administered mail-out/mail-back
2. Telephone
Computer Assisted
Telephone Interviewing (CATI)
3. Personal Visits
Computer Assisted
Personal Interviewing (CAPI) 1:3
When will it start?
It already has
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
ACS Test Sites
4 Counties
8 Counties
9 Counties
36 Counties
36 Counties
36 Counties
Supplemental
Survey
1,203 Counties
1,203 Counties
1,203 Counties
1,203 Counties
1,203 Counties
All 3,141 counties (no group quarters)
All 3,141 counties (with group quarters)
All 3,141 counties (with/out group quarters)
All 3,141 counties (proposed group quarters)
American Community Survey
Data Release Timetable
Type of
Data
Population
Size of Area
Annual
Estimates
65,000+
3-year
averages
20,000+
5-year
averages
Down to Census
Tract and Block
Group
Data for the Previous Year Released In the Summer of:
2006
2007
2008
2009
2010
2011
2012+
In August of
2006 you got
2005 data
Aug. 2010
Aug. 2008
Example 1:
Percent Foreign Born Population
Areas with 65,000+ population
2005 2006 2007 2008 2009 2010
Single-year estimates
3-year estimates (2005 – 2007)
3-year estimates (2006 – 2008)
20.0
21.2
23.3
28.6
32.6
35.1
21.5
24.8
3-year estimates (2007 – 2009)
28.6
5-year estimates (2005 – 2009)
25.9
3-year estimates (2008 – 2010)
32.2
5-year estimates (2006 – 2010)
28.9
Aug. 2008
Example 1:
Percent Foreign Born Population
Areas with population between
20,000 and 65,000
2005 2006 2007 2008 2009 2010
3-year estimates (2005 – 2007)
3-year estimates (2006 – 2008)
21.5
24.8
3-year estimates (2007 – 2009)
28.6
5-year estimates (2005 – 2009)
25.9
3-year estimates (2008 – 2010)
32.2
5-year estimates (2006 – 2010)
28.9
Aug. 2010
Example 1:
Percent Foreign Born Population
Areas with less than 20,000
population
2005 2006 2007 2008 2009 2010
5-year estimates (2005 – 2009)
25.9
5-year estimates (2006 – 2010)
28.9
2010
April 1
August
Dec.
(PL 94-171)
Main differences between Census Long Form
and ACS
Decennial Census
Sample: 1 in 6 Hhlds
(17% Sample)
Time reference: April 1,
2000 or April 1, 2010
(Last Week)
One time data collection
every 10 years
ACS
Sample: 1 in approx 40
households each year
(2.5%/yr 12.5% @ 5 yrs)
Time reference:
Continuous including all
12 months
Requires accumulation
over multiple years for
small area reporting
Some Emerging Issues
People might be counted twice, e.g. once at
summer cabin and once at a “permanent home”.
Or once at university and once at parent’s home.
Areas with seasonal populations, e.g. due to
“snow bird” migration, and due to school
enrollments, with housing occupancy that varies
from 95% to 40% might show up as 80%.
Mail back rates from Minority
populations are significantly lower
Source: C2SS test
Tracts with 75% or more of % of maila population reporting…
backs
White
60.5%
African American
34.9%
American Indian/Alaska
Native
Asian
16.6%
Hispanic
34.2%
58.6%
Different Non Response Follow-up
rates based on mail back returns
l
l
That is, in areas with low mailback returns, a higher NRFU
rate will be implemented
Thus, standard errors will be
more consistent across
neighborhoods
Understanding Sample Error
Sample Error is larger because the number of
census forms collected each year are smaller.
Changes of plus or minus 2% may be due to Sample
Error and do not reflect measurable change.
Comparing ACS directly to 2000 Data can be iffy
Statistical Significance Tests
Standard Error = Margin of Error / 1.65
Z
AB
(SE(A))2  (SE(B))2
But what about MOEs for 2000 data?
Chapter 8: Accuracy of the Data
Estimate Standard Errors
Adjust for Survey Design
Factors - Table C
http://www.census.gov/prod/cen2000/doc/sf3.pdf
Things to consider when
trying to compare ACS directly
to Decennial data
Seasonality
12 months instead of "April 1") (different jobs
and workers, especially in summer
Group Quarters currently not in 2005 ACS
Non-response follow-up differences
Data collection period ACS is very long, compared to
very short in decennial
Sample Sizes and MOEs ACS is about 1 in 40 households
compared to about 1 in 6
Data Release Schedule
1-year data in 2006 (every year thereafter)
3-year data in 2008 (every year thereafter)
5-year data in 2009 (every year thereafter)
Plus the within-year release schedules
Within-Year Data Release
Schedule for 2005 Data
August 15, 2006
Demographic and Social Characteristics
August 29, 2006
Economic Characteristics; Public Use Microdata File
October 3, 2006
Housing Characteristics; Narrative Profiles; Workplace
Base Tables
November 14, 2006
Selected Population Profiles
August 15th
Demographic and Social
August 29th
Economic Characteristics
October 3rd
Housing ---- Etcetera
i.e. Workplace Tables
October 3rd
Etcetera
… For Workplace Geography
November 14
Selected Population Profiles
Where do I get ACS?
FactFinder
There is a lot here
Even built in significance tests
Note difference between DC and NJ
A place to start mining ACS
Go to Detailed Tables
Check out all the Tabs
Pick Your Geography
State . . .
Places . . .
Then Tables . . .
Note the Tabs
Keyword Search
“B” and “C” Tables
Full Table – PASSED FILTERING
Statistically
too Small
Collapsed Table
You Can Also Download the Results
Download Options . . .
My Selection
What I took home
Brand New
CTPP Data Product
ACS CTPP Data Profiles
1. Resident Person Information
2. Resident Household Characteristics
3. Worker/Workplace Data
Resident Person Information for…
States, Large Cities, MSAs
Mode to work
Travel Time to work
Departure Time for work
Florida Statewide data
Mode to work by Travel Time
Mode to work by Average Travel Time
Resident Household Characteristics for..
The Nation, States, All PUMAs,
Counties and Places over 65k
Wisconsin Statewide data
Mode to work by Vehicles Available
Vehicles Available in Households
Household Size by Vehicles Available
Vehicles Available by Workers in Household
Persons in Households by Workers in Household
Worker / Workplace Data for…
States, Counties and
Places over 65k
Total Workers -- Travel Time to work -- Mode to Work
Get ACS CTPP Data Profiles
online or from the Class
Workbook CD
On-going concerns about ACS
• Will the sample be sufficient for small
geographic area reporting? Will tract and BG
and TAZ reporting occur after 60 months of
data collection?
• Will Congress continue to fund the survey
continuously, especially if CB asks for
increasing budgets to meet the sample size
requirements? (or will CB decrease the
sample size, resulting in loss of small area
tabulation?) Group Quarters
What Were the Transportation
Concerns in 2005?
• Will there be flow data? Flows are in serious
jeopardy due to thresholds and disclosure
proofing?
• Will there be a CTPP like package? Who will
organize it, negotiate with the CB and contract
with the CB. Timing of release? Cost?
• Will there be TAZs? How might they change,
what tables will be available, for what
geographies?
Consolidated
CTPP
Purchase
On-Demand
Technical Assistance
$930,000
Training
$625,000
Research
$920,000
Data Products
$2,980,000
Oversight Activities
$465,000
TOTAL
$5,920,003
AASHTO SCOP Census Data Work Group
CTPP Technical Group
Current Expectations
• Will there be flow data? Flows are in serious
jeopardy due to thresholds and disclosure
proofing?
• Will there be a CTPP like package? Who will
organize it, negotiate with the CB and contract
with the CB. Timing of release? Cost?
• Will there be TAZs? How might they change,
what tables will be available, for what
geographies?
ACS related activities
currently underway
NCHRP 8-48 Guidebook Using the American Community
Survey Data for Transportation Planning ($350K)
Expected May 2007
NCHRP 8-36 Project Disclosure Avoidance Techniques to
Improve ACS Data ($75) Expected Next Year
Census Geography and Participant Statistical Areas
Program — Hoping to have TAZs included
ACS Use Capacity Building Peer Exchange, SCOP
Working Group/FHWA, Spring 2007
Census Data for Transportation Planning
Preparing for the Future
Irvine, California
May 11-13, 2005
http://www.trb.org/conferences/censusdata
Now let’s go do some
Statistical Significance Testing!