American Community Survey Update

1
American Community Survey Update

• Prepared for the Joint State Data Center/Business
  and Industry Data Center and Federal-State
  Cooperative Program for Population Estimates
  Session
• March 28, 2006

2
Overview of Session

• Updates on operations, research and evaluations,
  user materials - Deborah Griffin
• Weighting and estimation Alfredo Navarro
• Update on Data Products Doug Hillmer
• Your questions

3
Data Collection Updates

• 2005
• Expansion to full housing unit sample in United
  States and Puerto Rico
• Challenges in Gulf Coast area
• 2006
• Introduced Group Quarters sample
• New translated materials

4
(No Transcript)
5
Interviewing ResultsPuerto Rico

• Low rate of mail and phone interviews.
• Reliability implications and deliverability and
  mailability issues.
• All results are weighted and estimate the
  proportion of noninterviews and interviews by
  mode out of the total eligible sample addresses
  in the January through June 2005 sample panels.

Noninterviews 2
Mail 20
Phone 6
Personal Visit 72
6
Interviewing Results United States

• Consistent with 2004 results prior to sample
  expansion.
• All results are weighted and estimate the
  proportion of noninterviews and interviews by
  mode out of the total eligible sample addresses
  in the January through June 2005 sample panels.

Noninterviews 2
Mail 51
Personal Visit 38
Phone 9
7
Data Capture and Data Processing Updates

• Testing image capture methods
• Processing of 2005 ACS data in spring 2006
• editing
• weighting/estimation

8
Language Assistance Program

• Monitoring language needs in CATI and CAPI
• Translated additional materials currently
  conducting cognitive testing of translations
• Planning for testing of multilingual
  brochure/letter

9
2005 ACS Special Product for Gulf Coast Area

• Scope of devastation
• Standard 2005 ACS data products summarize annual
  characteristics
• Recognition that additional information needed
  for LA, TX, AL, and MS

10
2005 ACS Special Product for Gulf Coast Area

• Special product will include two partitions of
  the data
• first 8 months of 2005
• last 4 months of 2005
• Consultations with SDCs and others on defining
  geographic tabulation areas
• Plan to release these products in May 2006

11
Research and Evaluation Update Data Collection
and Capture

• Content testing
• Mail response improvement
• Language brochure and language assistance
• GQ and Puerto Rico operational evaluations
• Data capture methodology

12
Content Testing Revised Wording of Current Topics

• Year built
• Number of rooms/bedrooms
• Complete kitchen/plumbing facilities

• Heating fuel
• Food stamps
• Value of this property
• Mortgage components
• Vehicles

13
Content Testing Revised Wording of Current Topics

• Place of birth
• Citizenship
• Year of arrival
• School enrollment
• Educational attainment
• Residence 1 year ago

• Disability
• Veterans status, period of service
• Labor force status
• Work status
• Industry and occupation

14
Content Testing New Topics

• Health insurance
• Marital history
• Service connected disability

15
Content Testing

• Data collection completed
• Analysis underway
• Plan to make final decisions on the 2008 ACS
  content in January 2007

16
Mail Response Improvement

• Dependent on 2007 funding of methods panel
  activities
• Testing additional mailings to improve mail
  response in areas with lowest levels of
  participation by mail

17
Language Assistance

• Testing mailing multilingual brochure to
  encourage calls to language assistance lines
• Developed set of translated materials in Spanish,
  Chinese, Vietnamese, Korean, and Russian and
  currently conducting cognitive testing of these
  translations

18
Operational Evaluations of Group Quarters and
Puerto Rico

• Includes evaluations of data collection methods
  and data quality
• Includes evaluations of addresses on the MAF for
  Puerto Rico - mailability and deliverability
  concerns

19
Data Capture Evaluation

• Testing change from key-from-paper to image
  capture and key-from image (KFI) technology
• Evaluation underway to compare data capture error
  rates and study respondent behavior in completing
  KFI forms

20
Research and Evaluation Update Weighting and
Estimation

• Multi-year estimates
• Current residence rules
• Weighting research
• Alternative uses of housing and population
  estimates as ACS controls

21
Research on Multi-Year Estimates

• Production of series of 3-year and 5-year
  estimates for the ACS test sites
• Expect to produce datasets in December 2006 with
  external release in early 2007

22
Current Residence Rules

• ACS residence rules use concept of current
  residence while population controls use concept
  of usual residence
• Evaluation will use ACS data to estimate the
  number of geographic areas with large differences
  in these populations
• Preliminary data expected this fall

23
User Materials

• ACS Design and Methodology report includes
  detailed documentation of the full set of ACS
  methods used in 2005/2006
• Expected release May 2006
• Plans for periodic updates

24
Design and Methodology Report

• Program History
• Frame
• Sample design selection
• Content
• Survey rules, concepts definitions
• Data collection and capture

• Language Assistance Program
• Weighting and estimation
• Variance estimation
• Data products
• Data dissemination
• Sources and controls on nonsampling error

25
User Materials

• Initial set of user guide training materials
  developed that include series of Power Point
  presentations on
• Overview of the ACS
• Data products
• How to access ACS data
• Future products
• Expected release May 2006

26
User Materials

• Additional materials needed
• Short brochure or manual as quick reference for
  users on data products and how they should be
  used
• Detailed guide on interpretation of multi-year
  estimates
• Need your input on additional specific needs

27
Weighting and Estimation
28
Overview of Session

• Plan to cover broad set of topics, including both
  technical details and specific examples
• Weighting and estimation
• Use of population controls in production of
  single-year estimates
• Single-year estimation
• Multi-year estimation
• Weighting and estimation research

29
Annual Weighting Process3 Major Components

• Initial weights to reflect the probability of
  selection
• Adjust weights of interviewed households to
  account for noninterviews
• Adjust weights to independent housing unit and
  population estimates (controls)

30
Initial Weight Probabilities of Selection

• Initial probability of selection is assigned as a
  function of the sample design
• Nonresponse follow-up (Personal Visit CAPI)
  sample design

31
Initial Sampling Rates
32
Sub-sampling RatesNonresponse Follow-up
33
Nonresponse Adjustment

• The weight of the nonrespondents is transferred
  to the respondents
• Nonresponse adjustment is carried out at the
  census tract level for groups of households with
  characteristics correlated with nonresponse

34
Ratio Adjustments to Housing Unit and Population
Controls

• Intercensal estimates are produced by updating
  the previous census results using various
  administrative records data
• In a multi-stage process, housing unit and
  population adjustment ratios are applied to the
  weights
• Applied at the county (or group of counties)
  level by race/ethnicity and age/sex groups

35
Weighting and EstimationSingle-Year Estimation

• Estimates include
• population estimates
• rates
• means, medians

36
Weighting and EstimationSingle-Year Estimation

• Percent foreign born population in year 1

37
Both Single- and Multi-Year Estimates are Period
Estimates

• 2005 single-year estimates are based on Jan 2005
  Dec 2005 interviews (12 months)
• 2005-2007 three-year estimates are based on Jan
  2005 Dec 2007 interviews (36 months)
• 2005-2009 five-year estimates are based on Jan
  2005 - Dec 2009 interviews (60 months)

38
Multi-Year Estimates

• Combining or pooling
• Population controls
• Tabulation geography
• Inflation adjustments

39
Multi-Year EstimatesPopulation Controls

• Simple average of the set of population controls
  for the years comprising the multi-year estimate
• For example, for the 2005-2009 five-year
  estimates, sum the controls released in 2010 for
  2005, 2006,,2009 divided by 5
• Use the most recently released estimates for each
  year

40
Multi-Year Estimates Tabulation Geography

• Boundary changes can occur through annexations
  during the multi-year period
• Plan is to tabulate using the geography of the
  most recent year in the multi-year estimate
• For 2005-2009 estimates, tabulate using all
  interviews for the period of 2005-2009 that were
  conducted in blocks that define the area in 2009

41
Multi-Year EstimatesInflation Adjustments

• The Consumer Price Index is used to compute
  inflation factors
• Dollar valued data items are inflation adjusted
  to the most recent year of the period
• For example, for the 2005-2009 estimates,
  appropriate inflation factors are applied to
  reported income values for 2005, 2006, ,2008 to
  adjust to 2009 constant dollars

42
Multi-Year EstimatesMedians

• Medians are produced using combined data records
  from all years, not by averaging each years
  median
• A 3-year median household income estimate is
  determined by combining the household records
  from the 3 years into one data set and
  determining the median from this combined
  distribution

43
ACS ResearchWeighting and Estimation
44
ACS ResearchWeighting and Estimation

• Multi-year estimates
• Family equalization
• Use of alternative population estimates as ACS
  controls

45
ACS ResearchMulti-Year Estimates

• Production of a series of multi-year estimates
  for the ACS test sites (1999 2005)
• The objective of this research is to assess
  issues of stability and comparability
• A secondary objective is gaining insights to
  develop and improve user materials

46
ACS ResearchFamily Equalization

• Assess impact of modifying weighting procedures
  to ensure consistency of the following estimates
• Households and householders
• Householders and spouses (and unmarried
  partners) in married (unmarried) couple
  (partners) households
• Subfamily householders and subfamily spouses

47
ACS ResearchUse of Alternative Population
Controls

• The quality of national and state level
  population estimates is generally accepted.
• The benefits of using of these controls by
  several current surveys is well documented and
  understood.
• The quality of county level estimates, in
  particular estimates of detailed demographic
  groups, has not yet been determined.

48
ACS ResearchUse of Alternative Population
Controls

• Define alternative population controls based on
  geographic and demographic detail
• Control to state level estimates by age and sex,
  and race/ethnicity and county population total
  only
• Control to county population estimates by sex and
  age and to place level total population

49
Contact Information
50
Data Products Update

51
Overview of Session

• Overview of data release schedule
• 2005 ACS data release schedule
• Geographic areas that will be covered
• Highlights of the new data products
• The 2005 ACS Special Product for Gulf Coast Area

52
Overview of Data Release Schedule

• Data collection is closed out just after the
  beginning of a calendar year
• Single-year and multi-year data products start
  to become available in the summer of the same
  year and continue to be released for several
  months
• The cycle repeats EVERY year

53
Data Release Schedule
54
2005 ACS data release scheduleFour data releases
in 2006

• Each release will emphasize a theme
• Demographic and social topics - MidAugust
• Economic topics - End of August
• Housing topics - October
• Selected Population Profiles for race, Hispanic
  Origin, and ancestry groups - November

55
Greater Sample SizesMore geographic areas can be
published
56
ACS Geographic Hierarchy
57
U.S. Counties Receiving Single-Year Estimates
58
West Virginia Only 7 counties will receive
estimates from the 2005 ACS in 2006
59
West VirginiaEach of the Census 2000 PUMAs will
receive 2005 ACS estimates in 2006
60
Publishing ACS estimates each year for the Census
2000 sample PUMAs will also help data users
interested in urban areas. This map shows the
city of Los Angeles partitioned into the Census
2000 sample PUMAs. ACS will publish for each of
these PUMAs in addition to Los Angeles county
61
Data Release Limitations

• Disclosure Avoidance
• Census Bureaus Disclosure Review Board (DRB)
  must clear all data products prior to their
  release to the public.
• Statistical Reliability
• Applied only to the one-year and three-year data
  products
• Data users must be able to use ACS estimates as
  official Census Bureau data. Thus, some rules
  must be in place to ensure minimum reliability of
  estimates. 

62
Data Release Limitations

• Statistical reliability is assured by
• Population size thresholds below which estimates
  are not released (65,000 for single-year
  products 20,000 for three-year products)
• Data release testing and collapsing of tables
  that fail testing is based on a pre-determined
  coefficient of variation value that cannot be
  exceeded.

63
Overview of ACS Data Products

• Examples will follow for
• Detailed Tables
• Data Profiles
• Ranking Tables
• Subject Tables
• Selected Population Profiles
• Other products
• Narrative Profiles
• Change (Multi-Year) Profiles
• Public Use Microdata Sample Files (PUMS)

64
Detailed Tables

• Basic distributions of characteristics, similar
  to SF3 Detailed Tables
• The foundation upon which other data products are
  built
• Show estimates and their associated lower and
  upper bounds of the 90-percent confidence
  interval
• Include distributions for more than 500
  characteristics, over 300 race and Hispanic
  Origin iterations, and 81 imputation tables.

65
(No Transcript)
66
Data Profiles

• Provide estimates of selected summary
  characteristics and important derived measures
  for each geographic area
• Rely on data tabulated in the Detailed Tables
• Produced for four sets of characteristics
• General demographic
• Social
• Economic
• Housing

67
(No Transcript)
68
Ranking Tables

• Currently available for over 80 subjects
• Available for states
• Each subject includes a table and a graphic
  representation, which shows the 90 confidence
  interval for each estimate
• Graphs and tables show survey estimates and
  90-percent confidence interval

69
(No Transcript)
70
Subject Tables

• Derived from Detailed Tables
• Show more detail than is available in the
  Profiles
• For a given topic, present distributions for a
  few relevant subgroups
• Allow for other measures such as medians and
  aggregates where appropriate
• Include the imputation rates for relevant measures

71
Sample Subject Table(selection from U.S. Subject
Table)
72
Selected Population Profiles

• Idea is to quickly produce a report on a
  population or housing group of interest
• Examples
• Native Hawaiians and Other Pacific Islander Race
• Groups Children under 18
• The elderly (60 and over)

73
Selected Population Profiles

• First Selected Population Profiles were published
  in January 2006
• Over 100 different Race, Hispanic Origin, and
  Ancestry groups

74
Sample Selected Population Profile(selection
from iterated Selected Population Profile for
Mexican Americans
75
Other Products

• Special Tabulations
• Done on a reimbursable basis similar to the
  Special Tabulations based on the 2000 Decennial
  Census
• Disclosure Review Board has different rules
  for these tables than for the standard data
  products

76
Other Products

• Public Use Microdata Sample (PUMS)
• Created each year from the full ACS sample
• Size will be approximately 1 of the total
  sampling universe
• Recoding and other techniques applied to data
  to protect confidentiality

77
2005 ACS Special Product for Gulf Coast Area

• ACS provides timely data for the population
  living in this area
• Interviewing in each county every month makes
  this product possible
• May release of products with 2005 estimates for
• January - August
• September December

78
2005 ACS Special Product for Gulf Coast Area

• Disclosure avoidance requirement county or group
  of counties must have at least 300 person
  interviews
• Local officials assisted in defining meaningful
  groupings of counties
• The estimates from Sept-Dec. will have larger
  variances, but no additional data release rules
  (for reliability) will be applied

79
(No Transcript)
80
2005 ACS Special Product for Gulf Coast Area

• Counties that are too small have been combined

Example St. James,St. John the Baptist, and St.
Charles parishes in Louisiana have been grouped
into the River Parish Region 560 completed
interviews
81
Contact Information

• Doug Hillmer
• American Community Survey Office
• Douglas.W. Hillmer_at_census.gov
• 301-763-2994

82
Your Questions?