Plans and Status for 2010 Census Coverage Measurement - Census

Document Sample
Plans and Status for 2010 Census Coverage Measurement - Census Powered By Docstoc
					Plans and Status for 2010 Census Coverage Measurement
Harland H. Shoemaker, Jr.

Decennial Statistical Methods Division

Prepared for: The 2007 Annual Meeting of the Census Information Centers October 10, 2007
1

2010 Census Coverage Measurement (CCM) Goals
• Help to improve coverage in future censuses
– coverage error components

• Understand differential coverage in the 2010 census
– net coverage error

2

CCM Definitions
• Coverage Error Components
– Omissions
• Persons or housing units not enumerated in the census, but should have been

– Erroneous enumerations
• Persons or housing units enumerated in the census who should not have been
– – – – Duplicates Non-residents Born after or died before Census Day Enumerated in wrong location (not an issue for U.S. estimate)
3

CCM Definitions (2)
• Net Coverage Error
– Net error = CCM estimate – Census count
• Used for persons or housing units • CCM estimate uses dual-system estimation methodology

– Net error  overall undercount or overcount

4

Major CCM Research Issues
• Improve measurement of Census Day residence • Improve techniques to detect and resolve duplicate census enumerations • Develop methods for measuring components of coverage error • Improve techniques to estimate net coverage error
5

CCM Program Limitations
• No estimates of coverage error for persons in group quarters or for group quarters facilities • Scope for estimating how component errors relate to census operations depends on information available from the final census files

6

2006 CCM Census Test
• Included:
– data collection and matching for persons only – automated person interview (PI, laptop)
• people living or staying at sample address on Interview Day

– paper followup
• non-matches, possible matches, cases needing more information

– computer and clerical matching
7

2006 CCM Census Test (2)
• Evaluated ability to determine correct Census Day residence
• Tried to determine if data collected met needs for estimating components • No production of coverage estimates

8

2008 CCM Dress Rehearsal
• Will include person and housing unit operations:
– Independent listing of housing units in CCM sample areas
• 2 sites: San Joaquin, CA; South-Central NC

– Housing unit matching to the census (preliminary address list)
• computer and clerical matching

– Housing unit followup (paper)

9

2008 CCM Dress Rehearsal (2)
– Automated PI (hand-held computer) – Person matching
• computer and clerical

– Person followup (paper) – Housing unit matching to the census (final address lists)
• clerical only

– Housing unit followup (final)

10

2008 CCM Dress Rehearsal (3)
• Will include Operational Assessments:
– Initial Housing Unit operations – Person operations – Final Housing Unit operations

• Will produce estimates of coverage (limited by nature of site test)
– net error – components of error
11

2010 CCM Survey
• Plan to list about 1M HUs in about 13,000 sample areas (U.S. and Puerto Rico) • Plan Person Interview in about 300,000 HUs in U.S.
– (+15,000 HUs in Puerto Rico)

• State sample allocation
12

2010 CCM Survey (2)
• Automated CCM person interview (hand-held computer)
• Nationwide computer match for persons
– to detect possible census duplicates

• Produce estimates of coverage error
– net error – components of error
13

?? Questions ??

14

Thank You
harland.h.shoemaker.jr@census.gov

donna.l.kostanich@census.gov

15