Creating Census Magic

1
Creating Census Magic
What We Muggles Can Do
July 2003
2
Todays Presentation

• Questionnaires and geographic concepts in 2000
• Census changes over time
• Alternative data sources
• How Documents Center can help
• Field trip back to the 19th Century

3
Census History

• Survey of the United States population every 10
  years
• Mandated by Constitution
• Purpose reapportionment of 435 seats in the
  House of Representatives

4
Census Questions Vary Over Time

• Age group, sex, free or slave in 1790
• Feeble-minded in 1840-1890
• Income first asked in 1940
• Televisions surveyed, 1950-70
• Detailed ancestry beginning 1980
• Multiple races and grandparents as caregivers in
  2000

5
2000 Questionnaires100 and Sample

• 100 Questionnaire
• Age
• Sex
• Race (Multiple)
• Hispanic origin
• Household relationship
• Occupied v. vacant housing units
• Owner v. renter occupied housing

6
Race Groups in 2000

• White
• Black or African-American
• American Indian or Alaskan Native
• Asian
• Hawaiian or Pacific Islander
• Other
• (Based on Self-Identification)

7
Individual Races

• Some files break the Asian, Pacific Islander, and
  American Indian groups into 250 specific
  categories
• Includes Chippewa Indians, Hmong, Guamanians
• Whites (Israelis, Arabs, Iranians and Afghanis)
  and Blacks (Nigerian and Haitian) are
  considered ancestries in sample data
• Racial definitions appear at
• http//www.lib.umich.edu/govdocs/census2/
• sf3td/sf3tdg7.pdf

8
New Race Category in 2000

• Respondents could choose up to SIX racial
  backgrounds
• Typical mixed race is 2-3 of population
• Race data not necessarily comparable with
  previous censuses

9
Hispanic

• Not considered a race
• Can be Hispanic and any race
• Breakdowns in some tables
• Cuban
• Puerto Rican
• Mexican
• Spanish
• Various Latin American Countries

10
Household Relationships

• Relationship to Householder
• Spouse
• Child
• Stepchild
• Grandchild
• Brother/Sister
• Parent
• Non-relative
• Unmarried partner is separate category

11
Group Quarters

• College dorms
• Prisons
• Mental hospitals
• Shelters for abused spouses
• Military barracks
• Nursing homes

12
100 Questionnaire Data Reports

• Pre-Tabulated Data
• Summary File 1 most age, race, sex data to the
  smallest geographies, block and block group
• Summary File 2 same data by 250 races to
  neighborhood (tract) level

13
2000 Sample Questionnaire(generally 1/6 of
population)

• Marital status, housing value and rent (100 in
  1990)
• Grandparents as caregivers (new)
• Ancestry
• Language
• Country of origin
• School enrollment and educational attainment (and
  dropouts)

14
2000 Sample Questionnaire

• Employment
• Industry and occupation
• Transportation to and place of work
• Disability and mental illness
• Veteran status
• Income and poverty

15
Summary File 3Ancestryhttp//factfinder.census.g
ov/

• Primarily countries with which white and some
  black races identify
• Options include first ancestry or only ancestry,
  second ancestry (in dual report), and ancestries
  totaled (all first, second and only)
• List of Ancestries
• http//www.lib.umich.edu/govdocs/census2/sf3td/sf3
  tdg1.pdf

16
Summary File 3Ancestryhttp//factfinder.census.g
ov/
Distribution of Individual Arab Ancestries by
Census Tract
17
Summary File 3Occupationhttp//factfinder.census
.gov/
Industry where you work Occupation what you do
18
Summary File 3Educational Attainment

• Population over 25 is counted
• Data is not cumulative
• Even though high school graduation rate may be
  0, people who completed bachelors or masters
  completed high school as well
• Main table is P37
• Table by race is P148
• Many tables in this data set are split so racial
  breakdowns appear in a later number

19
Summary File 3Educational Attainment
20
Summary File 3Disability

• Disabilities count overall limitations,
• not individual diseases
• Sensory
• Physical
• Mental
• Self-care
• Go outside the home
• Employment

21
Summary File 3Disability
22
Summary File 3Rent

• Contract rent amount paid to landlord
• Gross rent amount paid to landlord
• AND amount for utilities

23
Summary File 3Labor Force

• Labor Force either have a job or are looking
  for a job
• Not in Labor Force dont have job or want one
  (e.g. retiree)
• Unemployed in labor force but dont have job

24
Summary File 3Labor Force
25
Summary File 3Income

• Income calculated for households (related,
  unrelated, single), families (related), and
  individuals
• Income includes salaries, interest, social
  security, retirement, public assistance
• Median means half earn below and half earn above
• Aggregate means all of the income in that
  geographic area

26
Summary File 3Income
27
Summary File 3Poverty

• Calculated variable based on income, size of
  family, and federal poverty guidelines
• Poverty guidelines for 1999 appear at
• http//www.census.gov/hhes/poverty/histpov/hstpov1
  .html
• Example of ratio of income to poverty
• Poverty for one person in 1999 was 8501
• Under .50 earned less than 4250
• Earned 1.5 of poverty12751

28
Summary File 3Poverty
29
Sample Data Products

• Pre-Tabulated Data
• Summary File 3 socio-economic and housing
  characteristics to census tract or block group
  level
• Summary File 4 same data as Summary File 3 for
  210 racial and 125 ethnic groups to tract level

30
Public Use Microdata Samples

• Samples of the samples at 1 and 5 levels
• Only 1 currently available
• Create your own tables using raw data
• For 2000 currently requires statistical package
  such as SAS, SPSS, STATA
• Larger geographies states, metro-areas, PUMAS
  and super PUMAS

31
Importance of 100 and Sample Surveys

• 100 and sample data on the same subject (e.g.
  race, age, total pop, housing) dont always match
• Ann Arbor 100 114,024 Sample 114,110
• If one variable comes from sample data, try to
  get all variables from sample data
• Only 100 data used at block level
• Sample data to only to tract or block group level
• Larger the area, the more detailed the subject
  variables

32
Census Geography

• Legal Areas
• Nation
• State
• Counties
• Cities
• Townships
• Congressional Districts
• School Districts

33
Census Geography

• Census-Designated Areas
• Metropolitan Statistical Area
• Urbanized Area and Urban Cluster
• Census Tract
• Block Group
• Block
• Zip Code Tabulation Area
• PUMAS/Super-PUMAS

34
Census Geography Map
35
Urban Areas

• Urbanized Densely settled area, 50,000
• Urban Cluster Densely settled area,
• 2500-50,000 can be outside metro area

South Central Michigan has a surprising number of
urban clusters
36
Metropolitan Statistical Area

• Central city of 50,000 or more
• Its own county, and
• Surrounding counties with
• heavy commuting patterns

37
Metropolitan Area Definitions

• Metropolitan Statistical Area - stand-alone metro
  area
• Primary Metropolitan Statistical Area - metro
  area which is component of larger metropolitan
  area
• Consolidated Metropolitan Statistical Area - two
  or more contiguous metropolitan areas

38
Detroit Consolidated Metropolitan Statistical
Area
1990 Lapeer

Livingston Detroit PMSA Macomb
Monroe Oakland
St. Clair
Wayne Ann Arbor PMSA Washtenaw
2000 Lapeer

Macomb Detroit PMSA Monroe
Oakland St. Clair
Wayne
Lenawee Ann Arbor PMSA Livingston
Washtenaw Flint PMSA
Genesee

39
Detroit Metropolitan Area2000-2003





40
MSA Definitions

• This is very complicated
• Just consult the definitions when you need them
• http//www.census.gov/population/www/estimates/met
  rodef.html

41
Census Tracts

• Areas of about 4000 people
• Approximate neighborhoods

42
Ann Arbor Tract
43
Block Group
Two eight block groups per tract All 2000s
(2001, 2002, 2003) are BG 2 Smallest area for
sample data
44
Blocks
All blocks in 2000 have 4-digit numbers Some 100
data but no sample data
45
PUMAS and SUPER-PUMAS
For Public Use Microdata Samples http//ftp2.cen
sus.gov/geo/maps/puma/
46
PUMAS and SUPER-PUMAS
American Factfinder Reference Maps http//factfin
der.census.gov/
47
Converting Geographic Codes

• Use the Mable Geocorr 2K Search Engine to relate
  five digit zip codes to counties or census tracts
• http//mcdc2.missouri.edu/websas/geocorr2k.html

48
Converting Geographic Codes
49
Converting Geographic Codes
50
Converting Geographic Codes
51
Converting Geographic Codes
52
Locating Census Maps

• American Factfinder
• http//factfinder.census.gov/
• Census Bureau Web Site (PDF)
• http//ftp2.census.gov/plmap/
• http//ftp2.census.go/geo/maps/puma/
• Documents Center Paper Copies
• Tract maps for Michigan
• Block maps for Wayne and Washtenaw
• Geolytics CDs in Documents and Maps
• Arcview GIS Program in Map Library

53
Technical Documentation

• Geographic and subject definitions
• Lists of tables and individual components in
  tables
• Code lists (ancestry, race, occupation, group
  quarters, Hispanic, industry, language, country)
• Original questionnaire

54
Technical Documentation

• Links in American Factfinder when choose file
• http//www.census.gov/prod/cen2000/doc/
• Disaggregated versions
• http//www.lib.umich.edu/govdocs/census2/sf1td.ht
  ml
• http//www.lib.umich.edu/govdocs/census2/sf3td.ht
  ml

55
Census Data Formats2000

• FTP
• Transfer all data for all geographies in a county
  and manipulate with SAS or SPSS
• http//ftp2.census.gov/plmap/
• http//ftp2.census.gov/census_2000/
• http//www.icpsr.umich.edu/access/subject.html
  I
• American Factfinder
• http//factfinder.census.gov/
• Prepared profiles
• Data extraction to a spreadsheet,
• Reference and thematic maps

56
Census Data Formats2000

• Census Bureau CDs/DVDs
• Quicker than American Factfinder for extracting
  very large data sets
• Geolytics CD-ROMS
• Thematic mapping for all variables and created
  variables for 1970, 1980, 1990, and 2000
• Maps can be imported into more sophisticated
  programs

57
FTP Advantages

• http//ftp2.census.gov/census_2000/
• http//www.icpsr.umich.edu/ subject.htmlI
• Only option for PUMS data
• Data often available first via ftp
• Can manipulate very large data sets for all
  geographies in a state

58
FTP Disadvantages

• http//ftp2.census.gov/census_2000/
• http//www.icpsr.umich.edu/ subject.htmlI
• Must know how to use a statistical package
• Can get too much data

59
American Factfinder Advantages

• http//factfinder.census.gov
• SF1-4 data sets
• Custom tables option to choose parts of one table
  or mix tables
• Address search for all geographic codes of one
  street address
• Reference and thematic maps provide street
  boundaries and multiple geographic layers

60
American Factfinder Disadvantages

• http//factfinder.census.gov
• No PUMS files
• Limited types of geographies and number of data
  items retrievable
• Cannot choose multiple races/ancestries in SF2
  and 4
• Thematic maps can not be imported into GIS
  programs

61
Census Bureau CD Advantages

• Documents CD LAN and in Documents Center
• Numerous geographic retrieval options (e.g. all
  blocks in county, all tracts in state)
• Larger number of data cell retrievals
• Can save strategy to rerun program once geography
  or data elements have been chosen

62
Census Bureau CD Disadvantages

• Documents CD LAN and in Documents Center
• Only SF1 (nation) and SF3 (Michigan) available
  but SF2 and 3 (national) coming soon
• Not a substitute for ftp all data for all
  geographies in a state
• No thematic mapping program

63
Geolytics Advantages

• Documents CD LAN and in Documents Center
• 1990 and 2000 Sample Data also on CITRIX
• Thematic maps can be imported into ARCVIEW
• Available for 1970-2000
• ONLY non-ftp source for 1970 and 1980 digital
  census data
• Can create own variables using calculator

64
Geolytics Disadvantages

• Documents CD LAN and in Documents Center
• 1990 and 2000 Sample Data also on CITRIX
• No 100 data for 1990
• Limited number of geographic choices (e.g. no
  tracts in a city, just county)
• Maps do not show street boundaries
• Retrieval options vary slightly between censuses

65
Questions to Ask Before Starting a Time Series

• Did the Census Bureau ask the question?
• How was the question asked?
• Is the geography the same?
• What formats are available?

66
Census Questions Vary Over Time

• Reflects changes in society
• In 1990 and 2000 Congress wanted to reduce
  paperwork
• Grid of questions, 1790-2000
• http//www.lib.umich.edu/govdocs/census2/censubj.p
  df

67
Census Questions Vary Over Time

• Ancestry first asked in 1980
• Poverty was first determined in 1970
• Marital status and house value switched from 100
  to sample questionnaire in 2000
• Dilapidated housing last measured in 1960

68
Census Racial Definitions Vary Over Time

• Before 1980, Hispanic was called Spanish-origin
• Races in 1960 were white, Negro and other
• Multiple races new in 2000
• Asian and Hawaiian in 2000 were same category in
  1990

69
Technical Documentation

• American Factfinder for 1990
• Earlier years search MIRLYN
• census and technical w documentation
• Printed Census volumes

70
How Was Question Asked?

• Did 25 Ann Arborites and 122 Hawaiians take the
  subway to work in 2000?
• How did the person get to work LAST WEEK?

71
Questionnaires

• Technical Documentation in 1990.
• Printed Census Volumes through 1990.
• Measuring America
• http//www.census.gov/dmd/www/pdf/d02p.pdf

72
Is the Geography the Same?

• Metropolitan Areas Change
• Detroit 1970 - 3 counties
• 1980 - 6 counties
• 1990 - 7 counties
• 2000 - 6 counties
• http//www.census.gov/population/www/estimates/met
  rodef.html

73
Is the Geography the Same?

• Census Tracts, Block Groups, and Block Numbers
  Change
• Census Tract equivalencies in paper tract reports
  through 1990
• Relationship files, 1990-2000
• http//www.census.gov/geo/www/relate/rel_blk.html
• http//www.census.gov/geo/www/relate/rel_tract.htm
  l

74
Is the Geography the Same?

• Different definitions or urban and urbanized area
• Block group data on tape but not generally
  available until 1990
• Census tracts limited to metro areas until 1990

75
Census Formats

• Primary vehicle through 1980 was PAPER
• (Grad Stacks HA 201 .year with backup on
  microfiche/film in Documents)
• 1990 a combination of paper and CD not all files
  on American Factfinder
• No paper in 2000
• 1970-80 Geolytics are only good CDS for period

76
Data Tapes

• 1970 - 2000 at ICPSR
• http//www.icpsr.umich.edu/access/subject.htmlI
• 1850-1960
• ICPSR has some data http//www.icpsr.umich.edu/acc
  ess/subject.htmlI
• IPUMS has PUMS
• http//www.ipums.umn.edu/usa/
• Historical Data Browser (state and county)
• http//fisher.lib.virginia.edu/census/

77
Citing Census Data

• Census Bureau
• http//www.census.gov/main/www/citation.html
• APA Style - Purdue
• http//owl.english.purdue.edu/handouts/research/r_
  apa.html
• Government Documents - Univ. of Memphis
• http//exlibris.memphis.edu/govpubs/citeweb.htm

78
Documents Center

• Campus representative to the Census Bureaus
  State Data Center program.
• (e.g. we often know people who can help or when
  something is being released)
• Assistance with American Factfinder (data sets
  and Factfinder mapping)
• Assistance using Census Bureau and Geolytics
  CD-ROMS
• Physical copies of CD-ROMS on the LANs
• Paper tract maps for Michigan, 1990-2000.
• Census CD-ROMS.Paper tract maps of Michigan.
• Referrals to other sources.
• Census 2000 web page
• http//www.lib.umich.edu/govdocs/cen2000.html

79
Documents Center

• All Census CD-ROMS for 1990 and alternative data
  access through the EXTRACT software.
• Assistance with historical printed reports
• SOME data interpretation.
• GUESS on the need for PUMS although we cant run
  it.
• Alternative data sources

80
Documents Center

• 203 Hatcher Library North, [phone],
  govdocs_at_umich.edu
• http//www.lib.umich.edu/govdocs/
• Summer Hours
• M-F, 1-430 p.m.
• Special arrangements possible