Weve Got Questions! Whats happening with the Census? What are these new data? How do I use the data?
Decennial Census Data Decennial Census runs every 10 years Basic demographic data to the census block level is released through Summary Files 1 and 2 Detailed socio-economic data to at least the census tract level and in many cases to the census block group level is released through Summary Files 3 and 4
The Need for More Data The socio-economic data on Summary File 3 is very useful, but … The data is needed more often than once every 10 years. Point-in-time data collection (i.e. April 1, Census Year) may not provide the most accurate picture of a community.
What is ACS? Monthly, rolling survey to collect socio-economic data. Annually reports data Data reported to the at least to the census tract level. In many cases census block group data will be available Replaces Summary Files 3 and 4! 2010 Census will be short form only
Data Collection Census 2000 Summary File 3 sent a long-form questionnaire to a 1 in 6 household sample nationally. ACS sample will be smaller than the Census 2000 sample. Decennial census surveys every 10 years ACS surveys every month.
Rolling Monthly Survey Sample addresses are selected every month from the MAF and surveys are mailed. The following month, Computer Assisted Telephone Interviewing (CATI) will contact the non-responding households. A month after CATI, Computer Assisted Personal Interviewing (CAPI) will contact the non-responding households.
Jan.Feb.MarchAprilMay Sample 1 Mail Sample 1 CATI Sample 1 CAPI Sample 2 Mail Sample 2 CATI Sample 2 CAPI Sample 3 Mail Sample 3 CATI Sample 3 CAPI Sample 4 Mail Sample 4 CATI
n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n
Sampling Sampling will cover all census geographic levels. Since the sample size is smaller, it will take more time to collect adequate data from smaller population areas. This sample time requirement means that data will be reported differently for smaller population areas.
Data Reporting ACS data will be released annually. Data for areas with a population of 65,000 or more will be reported as single-year estimates. Data for areas with a population between 20,000 and 64,999 will be reported as a 3-year average. Data for areas with a population under 20,000 will be reported as a 5-year average.
Data Release Schedule Type200520062007200820092010 Annual Est. X*XXXXX 3-Year Avg. XXX 5-Year Avg. X * - Data from test areas with a population of 250,000 or more
Data Available in 2006 2006 will be the first year of data for a fully-implemented ACS. Earlier data is available for test sites, but these tests use slightly different methodologies. Group quarters are not included in the test area data. Test areas were generally based on areas with a population of 250,000 or more.
Data Releases in 2006 Demographic data was released in mid-August. Economic data was released in late August. Housing data will be released in October. Special cross-tabulations will be released in November. Data only for areas with a population of 65,000 or more.
Data Products Profiles data, narrative, multi-year Selected Population Profiles race or ethnic group, ancestry Tables subject, detailed, custom Ranking Tables states, counties, places Maps reference, thematic
Data Profiles DemographicSex and age, race, relationship, household by type SocialEducation, marital status, veteran status, place of birth EconomicIncome, employment, occupation, industry HousingOccupancy and structure, housing value and costs, utilities
Narrative Profiles Text based profile Select information from each of the data profiles Bar charts of select data items Quick links to the Ranking Tables main page
Selected Population Profile Single profile combining demographic, social, economic and housing data items Data is shown for a specified race/ethnic group including some multiple race and tribal groups or for a specified ancestry group
Tables Detailed Multiple data items organized into a static table Subject Data items from different detailed tables are organized into static, topic-related tables and organized under general subject headings Custom User selects data items from desired detailed tables to create a unique data table
Ranking Tables 2005 ACS compares selected data item across all states. Previous ACS data ranked counties and places. Data items are organized by subject. Sorts are dynamic and may be resorted by rank or alphabetically by geography name.
Geographic Comparison Tables Shows select data as percentages or medians for all geographies within a state. GCT tables list geographies alphabetically or numerically and, unlike ranking tables, are static. There are approximately 65 GCT tables.
Geographic Comparison Table – 2005 Geographies States Congressional Districts (109 th Congress) Counties Places Public Use Microdata Areas School Districts Urban/Rural and Inside/Outside Metropolitan and Micropolitan Statistical Areas
Maps Reference maps show location of geographies with no data. Thematic maps show selected data items, similar to those shown on the ranking tables, across user-selected geographies. Thematic maps allow user to define data classes, boundaries, features, and titles.
Test area data reports an estimate, upper, and lower bounds.
2005 data reports an estimate and a margin of error.
ACS Data All ACS data are estimates 90% confidence that the actual data point is within the margin of error While this is a new way of presenting the data, there has always been a confidence interval with Summary File 3 sample data.
ACS and Decennial Data Since ACS and Summary File 3 are sample data products and ACS replaces the long form (Summary File 3), isnt it OK to compare ACS data with decennial data? NO!
ACS and Census – Comparison Sample sizes are different Residency rules are different Legislative purposes are different Group Quarters data may be missing Geographies may be different Users should not compare ACS and SF3 data
ACS Table Names Base table data is the most detailed and is the base for all other ACS tables and profiles Census contains Population and Housing tables (i.e. P0010001, PCT159H001, H001001) ACS contains Base and Collapsed tables (i.e. B03002, C03002, B05003H)
ACS Table Naming Structure Initial letter identifies the type of table – base and collapsed Following 2 digits identify subject area 01 – age and sex 02 – race There are 26 subject areas Next 3 digits are table sequence numbers An alphabetic suffix indicates race/ethnic classifications and are not the same as the Census classifications
Collapsed Table – Quick Guide For some areas, the detailed Base Tables may have estimates that are unreliable. Reliability, in the form of median coefficient of variation, is checked for all Base Tables for each area. Coefficient of Variation = standard error/estimate If the median CV for a Base Table is less than 61%, the table passes and is published. If the median CV for a Base Table is greater than 61%, the table fails.
Collapsed Table – Quick Guide If a Base Table fails the check, some detailed estimates are collapsed into a single estimate in a Collapsed Table. The median CV reliability check is now run on the Collapsed Table. If the Collapsed Table passes the median CV check, it is published. If the Collapsed Table fails the median CV check, it is dropped and not published.
Collapsed Tables – Last Word The median Coefficient of Variation check is only done on Base Tables for single year data (i.e. for areas with a population of 65,000+) and for 3-year average data (i.e. for areas with a population from 20,000 to 64,499). For 5-year average data, all Base Tables are published. There are no dropped or collapsed tables. Standard disclosure rules apply to all ACS releases.
ACS and Dynamic Geographies Populations change Boundaries change ACS reports data as single year, 3- year average, and 5-year average depending on the population of the area. ACS accounts for these changes through sampling strategy and the Boundary and Annexation Survey.
Boundaries Governmental unit boundaries are based on annual Boundary and Annexation Survey (BAS) benchmark file Boundaries are as of January 1 of the reported year (i.e. 2005 ACS boundaries are as of January 1, 2005) Non-governmental unit census geographies will be drawn once a decade prior to the decennial census
Sampling Monthly samples include locations across census geographies. Samples are collected each month for areas that are not in incorporated places. When areas are annexed, their sampled data is added to the total for the annexing place.
n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n n
Sampling Controls Households missed in the ACS survey can skew the data. The US Census Bureau produces annual Intercensal Population Estimates for counties. These estimates for total, age, race, sex, and Hispanic Origin populations and housing unit estimates are used as controls for ACS at the county level.
Sampling Controls The Intercensal Population Estimates are NOT the same as the State Demographers estimates. County level population, age, race, sex, Hispanic Origin, and total housing unit data on ACS will equal the Intercensal Population Estimates data (without Group Quarters data for 2005). Underlying ACS survey data is not controlled and sums of data from Detailed Tables may not equal the controlled totals.
Points to Remember ACS and the Decennial Census are separate products. Group quarters data are not included in the 2005 ACS ACS data are estimates with margins of error Single year data will be more volatile than 3- or 5- year averages ACS is new and adjustments will happen
Accessing ACS American FactFinder http://factfinder.census.gov click Data Sets http://factfinder.census.gov North Carolina State Data Center http://census.state.nc.us under Survey Data from the US Census Bureau http://census.state.nc.us
Questions Bob Coats Bob.Coats@ncmail.net North Carolina State Data Center North Carolina Office of State Budget & Management (919) 807-4781