American Community Survey Update

1
American Community Survey Update

• Presentation made to the
• Joint CIC/FSCPE/SDC Steering Committee
• February 6, 2008
• Susan Schechter and Deborah Griffin
• American Community Survey Office

2
Areas of Emphasis in 2008

• Continue careful management of resources
• Submit to Congress by March 31, 2008 the planned
  questions for 2010 Census and ACS
• Successfully implement new questionnaire in 2008

3
Areas of Emphasis in 2008

• Finalize recommendation from 2007 Content Test on
  Field of Degree question and submit revised
  questionnaire to OMB in July
• Increase focus on data users needs
• Produce the first multiyear estimates
• Continue critical research priorities

4
FY08 Budget Update

• Budget includes full funding for conducting the
  ACS in the U.S. and Puerto Rico
• Program changes
• Increased funds for annual ACS Field Interviewer
  training sessions and observation activities
• Increased funds for enhancements to the Group
  Quarters activities

5
FY08 Budget Update

• No Methods Panel funding
• Canceled plans for test of multi-lingual brochure
• Canceled plans for test of additional mailing
  during CATI phase
• Canceled all planning activities

6
Submission of 2010 and ACS Questions to Congress

• Required by law two years prior to conducting the
  decennial census
• Subjects planned for 2010 Census and ACS
  submitted last year
• www.census.gov/Press-release/www/2007/subjects_not
  ebook.pdf
• Submission to Congress includes examples of
  questions, citations, and brief narrative
  description of uses

7
Submission of 2010 and ACS Questions to Congress

• Interagency Committee for the ACS is the
  mechanism for reaching out to agencies to obtain
  needed documentation
• Will serve as a useful tool for you in explaining
  the purpose of questions to data users will show
  you which federal agencies use the data most

8
Implementation of New 2008 Questionnaire

• Based on results of 2006 ACS Content Test
• 8 housing topics were tested
• 12 detailed person topics were tested
• 3 new questions were tested
• Health Insurance Coverage
• Marital History
• Veterans Service-connected Disability

9
Interviewer Feedback on the New Questionnaire

• Purpose is to provide additional training,
  guidance, or clarifications if needed
• Close to 50 telephone and field representatives
  will participate
• Focus groups will also be conducted

10
Interviewer Feedback on the New Questionnaire

• Interviewers are asked if respondents had
  difficulty understanding or responding to
  questions
• Also ask interviewers what strategies they found
  to be successful in addressing those concerns and
  what other tools they need

11
2007 ACS Content Test

• Tested two versions of a new question on field of
  bachelors degree
• Categorical/forced-choice
• Open-ended
• Benefits
• Annual field of degree (FOD) data for small
  levels of geography
• Assist with sample frame development for NSCG

12
Categorical
13
Open-ended
14
Outreach and Planning for Multiyear Estimates

• Engaged stakeholders in several ways
• Federal Register Notice described plans for data
  products and release schedule
• Summarized public comments and distributed to
  advisory groups and stakeholders for additional
  comment
• Convened workshop of selected experts to review
  public comments

15
Outreach and Planning for Multiyear Estimates

• We are now actively collaborating with experts to
  develop education materials for a broad set of
  data users
• Materials will focus on how to and will include
  case studies
• Education campaign will include materials on the
  web that can easily be downloaded

16
Looking Ahead

• Education and information
• Education materials will be finalized.
• Extensive updating of the website content.
• Outreach campaign with Federal agencies and
  stakeholders to assist in understanding multiyear
  estimates.

17
Looking Ahead

• Data release
• Release of 2007 ACS 1-year estimates beginning
  with the August release of income and poverty.
• First time users will have two years of full ACS
  data to compare
• Release of 2005-2007 multiyear estimates for
  smaller areas of geography before the end of the
  calendar year

18
Research and Evaluation Updates

• Seasonality Research Q25
• Multiyear Estimates Study

19
Background

• The ACS questionnaire in 1999 2002 collected
  information about seasonal households
• Cognitive testing found problems with
  respondents completion of this question

20
Background

• The question was revised in 2003 and data were
  collected using the revised question in 2003
  2007
• Q25 data were never published collected for
  evaluation uses only
• Q25 was dropped from the 2008 ACS

21
Background - Q25a

• Do you or any member of this household live or
  stay at this address year round?
• Yes SKIP to the questions for Person 1
  on the next page
• No

22
Background - Q25b

• How many months of the year do members of this
  household stay at this address?
• Months

23
Background Q25c

• What is the main reason members of this household
  are staying at this address?
• This is their permanent address
• This is their seasonal or vacation address
• To be close to work
• To attend school or college
• Looking for permanent housing
• Other, specify

24
Results

• Responses to Q25 in the 2005 ACS were summarized
  for this analysis
• Data were weighted using the final ACS housing
  weights, no editing or adjustments were made for
  item nonresponse
• Separate estimates were produced for all areas
  that were published in 2005

25
2005 National Results Q25a Universe Occupied
Households
26
2005 National Results - Q25bUniverse Occupied
households responding No to Q25a
27
2005 National Results - Q25cUniverse Occupied
households responding No to Q25a
28
Using Q25 to Identify Seasonal Housing Units

• For purposes of this analysis, seasonal housing
  units are defined as all occupied housing units
  providing
• a response of No to Q25a and
• a value of 1-5 months in Q25b

29
2005 National Results - Q25

• Based on this definition, in 2005 the percent of
  occupied housing units classified as seasonal
  housing units was less than 0.5 at the national
  level
• Variation seen at the state level

30
2005 ResultsState-level Rates of Seasonal
Housing Units
31
2005 ResultsStates With Highest Rates of
Seasonal Housing Units

• Arizona 1.8
• Florida 1.5
• Maine 1.0

32
2005 ResultsCounty-level Rates of Seasonal
Housing Units
33
2005 ResultsCounties With Highest Rates of
Seasonal Housing Units

• Yuma, AZ 8.9
• Cape May, NJ 5.9
• Highlands, FL 5.5
• Indian River, FL 4.4

34
2005 ResultsCounties With Highest Rates of
Seasonal Housing Units (cont.)

• Pinal, AZ 4.4
• Collier, FL 4.3
• Lee, FL 4.2
• Manatee, FL 4.1
• Madison, KY 4.0

35
2005 ResultsOther Geographic Areas

• Congressional Districts - 4 of 436 had rates
  above 3 (highest 4.4)
• Places - 5 of 509 had rates above 3 (highest
  7.5)

36
2005 ResultsOther Geographic Areas

• Metropolitan or Micropolitan Statistical Areas -
  10 of 482 had rates above 3 (highest 8.9)
• Public Use Microdata Areas (PUMAs) - 30 of 2,071
  had rates above 3 (highest 9.4)

37
Observations

• Preliminary assessment is that seasonal housing
  units do not appear to be an issue for areas
  receiving 1-year estimates
• Seasonal housing units are expected to cluster at
  smaller geographic levels

38
Next Steps - Spring 2008

• Produce estimates based on Q25 for areas of
  20,000 and greater based on 2003 - 2006 ACS data
• Since 2003 and 2004 have reduced sample sizes,
  this approximates a 3-year estimate
• Analyze results to estimate the number of areas
  receiving 3-year estimates with high rates of
  seasonal housing units

39
Next Steps Fall 2008

• Produce estimates based on Q25 for smallest areas
  based on 2003 - 2007 ACS data
• Since 2003 and 2004 have reduced sample sizes,
  this approximates a 5-year estimate
• Analyze results to estimate the number of areas
  receiving 5-year estimates with high rates of
  seasonal housing units

40
Additional Research

• At what level is the rate of seasonal housing
  units high enough to be problematic?
• Additional analysis is needed to assess if the
  characteristics of the population living in
  seasonal housing differ markedly from the
  population in non-seasonal housing

41
Your Feedback

• Goal of todays presentation is to share
  information about our approach to studying this
  issue
• Your feedback is welcomed

42
Findings from the Multiyear Estimates Study
43
Multiyear Estimates Study

• Included research conducted by Census Bureau
  staff and external researchers
• Preliminary evaluation summaries are accessible
  from the ACS Multiyear Estimates webpage, final
  reports will be available soon

44
Multiyear Estimates Study

• Provided an opportunity to dress rehearse and
  improve methods planned for this years
  production of 3-year estimates
• Raised issues about thresholds and data release
  rules
• Gave us real examples and case studies to better
  understand issues

45
Multiyear Estimates Study

• Reliability
• Research by Starsinic and Tersine includes tables
  with median CVs by topic crossed by size of
  geographic area and by size of estimate
• G-weights found to reduce SEs for some key
  estimates, no negative effects found
• Reliability concerns raised about tract-level
  estimates (aggregating results suggested)

46
Multiyear Estimates Study

• Stability issues noted with 1-year estimates
  (improvements found with use of 3-year estimates
  instead)
• Analysis of data by race and ethnicity showed
  consistency in the differences in poverty and
  home ownership rates

47
Multiyear Estimates Study

• Additional research is underway on quality issues
  coverage, item nonresponse, and unit
  nonresponse rates for small areas
• Guidance on usability developed to help users
  choose between 1, 3, and 5-year estimates

48
For More Information
American Community Survey www.census.gov/acs/www c
mo.acs_at_census.gov
ACS Alert http//www.census.gov/acs/www/Special/Al
erts.htm