1 / 54

February 16, 2011

American Community Survey Steering Committee Updates State Data Centers Census Information Centers Federal-State Cooperative for Population Estimates. February 16, 2011. Plan for today. Introductions and data release updates Content and Internet Test updates

eloise
Download Presentation

February 16, 2011

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. American Community SurveySteering Committee UpdatesState Data CentersCensus Information CentersFederal-State Cooperative for Population Estimates February 16, 2011

  2. Plan for today • Introductions and data release updates • Content and Internet Test updates • Margin of error issues: aggregating estimates, calculation of margins of error • Group quarters small area estimation • Population estimates and controls

  3. 2009 ACS data releases • PUMS files were released 1-2 months after estimates

  4. 2010 ACS data releases(tentative) • PUMS files will be released 1 month after estimates

  5. Plan for today • Introductions and data release updates • Content and Internet Test updates • Margin of error issues: aggregating estimates, calculation of margins of error • Group quarters small area estimation • Population estimates and controls

  6. Update on ACS Content and Internet Tests Anthony Tersine Decennial Statistical Studies Division SDC-CIC-FSCPE Steering Committees

  7. Content Test – Background • Two new topics: • Computer ownership/Internet access • Parental place of birth • Four revisions to current topics • Food stamps • Veteran’s identification and period of service • Public assistance income • Wages/salary and interest/dividends

  8. Content Test – Where are we? • Data collection completed • Finishing up coding of write-ins • Analysis will begin shortly • Results out Fall 2011 • New content in January 2013 or later

  9. Internet Test – Background • Testing various ways of offering the Internet response option in the mailings • Metrics – response rates, form completeness, internet break off rates, etc. • Collecting paradata to analyze instrument design • Test delayed from March 2011 to April 2011

  10. Paradata – Example Research Topics • Problematic screens/questions • Usefulness of help • Problems with authentication procedures • Behavior related to exit/reentry

  11. Internet Instrument Features • Available in Spanish • Pre-fill names in questions • FAQs/instructions accessible from all screens • Question-specific help • Soft edits to combat item nonresponse • Review/edit screen

  12. Login

  13. Help and Progress Indicator

  14. Grey out Features

  15. Review and Edit

  16. Plan for today • Introductions and data release updates • Content and Internet Test updates • Margin of error issues: aggregating estimates, calculation of margins of error • Group quarters small area estimation • Population estimates and controls

  17. Aggregating Estimates and Calculating Margins of Errors Presented at the Joint FSCPE/CIC/SDC Meeting February 16, 2011 Karen King American Community Survey Variance Estimation and Statistical Support Branch

  18. Overview of this Presentation How to aggregate estimates and approximate the associated margin of error (MOE). Two examples will be presented Discuss issues associated with derived MOE

  19. Aggregating Estimates Published estimates may be aggregated to form additional estimates. With 5-year data, Tract/Block Group estimates can be aggregated to form user defined areas. Calculating the MOE for these defined areas is also important.

  20. MOEs of Aggregated Estimates To aggregate two published estimates, simply add the estimates together, but cannot simply sum the MOEs together. The actual formula is Covariance is unique and not published. The approximation is used instead:

  21. Example 1:Total number of people with income below the poverty level Total = 42,945 + 61,956 = 104,901 So the total is 104,901 with an approximate MOE of 7,376.

  22. Example 1 Total number of people with income below the poverty level Covariance = 30,679,038

  23. Example 2: Total number of males with income below the poverty level So the total is 23,001 with an approximate MOE of 3,046.

  24. The Adjusted Covariance Matrix. The values on the diagonal in blue are the squared MOEs. The off diagonals are the covariance. We can see that the covariance range in size and are non-trivial.

  25. What can be done? We have found that the approximation formula seriously breaks down when aggregating more than four estimates. So we suggest you aggregate the fewest number of estimates as possible. Other Options: Calculate the estimates using the Public Use Microdata Sample (PUMS) Request a special tabulation (fee based and certain criteria apply)

  26. Accuracy Documents The formulas used to approximate the MOEs here are available in the ACS Accuracy document. It also contains more examples. The URL for the Accuracy documents is: http://www.census.gov/acs/www/data_documentation/documentation_main/

  27. Plan for today • Introductions and data release updates • Content and Internet Test updates • Margin of error issues: aggregating estimates, calculation of margins of error • Group quarters small area estimation • Population estimates and controls

  28. Proposed Procedures for ACS Group Quarters Estimation for 2010 Prepared for the FSCPE / SDC / CIC Steering Committees February 16, 2011 Mark E Asiala U.S. CENSUS BUREAU U.S. DEPARTMENT OF COMMERCE Washington, DC 20233

  29. Overview • Background of the Estimation Issues • Proposed Solution • Closing Remarks

  30. Background • GQ sample is designed to produce state-level characteristics of the GQ population • GQ population also contributes to • Characteristics of total resident population • Estimated total population for noncontrolled areas • Consequence of sampling • Lack of representation of GQ sample for some areas • Over-representation in other areas

  31. ACS GQ Sample in Tracts by Major GQ Type Group (2006‑2009 ACS Sample)

  32. Proposed Solution • Long term solutions • Alternative sample designs that are better optimized for small areas • Cost and time implications • Topic of panel by the National Academy of Sciences • Near term solutions • Solution that uses existing data

  33. Solution: Identify Problem Areas • Goal is to achieve representation in estimates of GQ population wherever we have it in our sampling frame • Specifically, we want representation for the following combinations • County by major GQ type • Tract by major GQ type

  34. Major GQ Types • 27 specific types of GQs in ACS • Grouped into seven major GQ types • Adult correctional institutions • Juvenile facilities • Nursing homes • Other long-term care facilities • College dorms • Military facilities • Other non-institutional GQs

  35. Solution: Impute GQ Persons • Select a subset of GQs from our frame that were not in sample • All GQs whose is expected size is 16 or greater • A subset of the smaller GQs to satisfy our goal of representation • Impute person records into those GQs • Donors are the interviewed persons

  36. Details on the Imputation Process • Our search for donor records accounts for • Geographic proximity • Specific and major GQ type • All male / female GQs for a subset of the GQ facilities

  37. Summary of our Research Work • Tested our methods using • Census 2000 data • Simulated 2006-2010 ACS file • 2000 data allowed detailed analysis using just the 100% data items • ACS data allowed a comparison of estimates from the 2009 methodology to the new methods

  38. Noted Improvements • Improved representation of GQ population estimates across all areas • Reduction or elimination of extreme estimates of total GQ population for substate areas • Impact of the GQ portion on characteristics of total resident population should be in line with expectations

  39. Concluding Remarks • This method is proposed to be implemented with 2010 data year. • There are opportunities to refine the method in future years. • A strong source of auxiliary data in the future will be 2010 Census.

  40. Thank you. • Mark E. Asiala • mark.e.asiala@census.gov

  41. Plan for today • Introductions and data release updates • Content and Internet Test updates • Margin of error issues: aggregating estimates, calculation of margins of error • Group quarters small area estimation • Population estimates and controls

  42. Population Estimates and Controls for the American Community SurveyVictoria VelkoffPopulation DivisionU.S. Census BureauFSCPE-CIC-SDC Steering Committee MeetingFebruary 2011

  43. Overview • Population estimates - what we produce and how • Postcensal versus intercensal population estimates • Population estimates as controls for the ACS • Types of population controls for the ACS by year of release

  44. Estimates Produced Annually • Population • Nation by age, sex, race, and Hispanic origin • States by age, sex, race, and Hispanic origin • Counties by age, sex, race, and Hispanic origin • Incorporated places and minor civil divisions (total population only) • Puerto Rico Commonwealth and municipios by age and sex • Housing units • States • Counties

  45. Producing Population Estimates • Estimates base is most recent Census with some modification (e.g., Some Other Race is recoded). • From the last Census forward, population is estimated using a cohort-component method at the national, state, and county levels. • Nation: Population2 = Population1 + Births - Deaths + NIM NIM = Net international migration • States and counties: Population2 = Population1 + Births - Deaths + NM NM = Net domestic and international migration • Subcounty estimates produced using a distributive housing unit method.

  46. Postcensal Versus Intercensal Estimates • Postcensal population estimates • Built off of the last census • “Vintage” identified by terminal year in the series • July 1 estimates, full series from last Census date forward (for Vintage 2009, series is April 1, 2000-July 1, 2009) • Intercensal population estimates • Based on two consecutive censuses • Will be produced by age, sex, race, and Hispanic origin at the county level • Will be used to control 2010 ACS products

  47. Population Estimates as Survey Controls for ACS • Population estimates are the official estimates for the nation, states, counties, cities, and towns. • Population estimates are used as survey controls for the ACS to reduce variance and coverage bias.

  48. Population Controls Provided to ACS • Population estimates provided as controls • County by age (single years 0-84,85+), sex (male, female), race (31 races), and Hispanic origin (Hispanic, non-Hispanic) • Puerto Rico municipios by age (single years 0-84, 85+) and sex • For ACS 2009 and beyond – subcounty total population estimates • Group quarters population by the 7 major types at the state level and for Puerto Rico • Housing units at the county level and subcounty level

  49. ACS Controls • ACS creates their set of controls from the population estimates for weighting areas which are counties or groups of counties • 13 age groups • 5 race alone categories (non-Hispanic) • Hispanic • Group quarters controlled at the state level by type (7 major types) • ACS uses the housing unit estimates to control the number of housing units in a weighting area

  50. Population Controls by ACS Release Year: 2010

More Related