Ongoing Issues with the Census

1
Ongoing Issues with the Census

• Despite the very high quality of the data and the
  Census Bureaus decades of experience in
  completing this exercise, there are still major
  issues surrounding the Census.
• These include
• 1) Census Accuracy
• 2) Capturing Attributes Correctly
• 3) Political context
• 4) Future of the Long Form

2
Ongoing Issues with the Census
3
Census Accuracy

• Despite spending over 5 billion (yes Billion) on
  the 2000 Census, the Census Bureau readily admits
  that it did not complete a full enumeration of
  the nations population.
• For example, in 1990 it is estimated that the
  Census missed over 4 million people. Why?
• Difficult to count populations
• --Homeless --Immigrants --Central city
  populations
• Low response rate to census survey (67 in 00,
  65 in 90, 75 in 80)
• Errors in Census data inputs (boundaries,
  geocoding, coverage errors) For example, for
  the 2000 census, the CB mistakenly located a
  prison with 1,228 inmates that is in Gulf County
  in Franklin County.
• Consequently, the CB issued a CQR (Count
  Question Resolution) that counted the 1,228
  inmates in the correct county. However, while the
  totals for each county are now correct, the
  underlying SF1 and SF3 data have not been updated.

4
Capturing Attributes Correctly

• For each Census the CB comes under fire from
  different groups for how they capture certain
  population attributes, especially
• --Race --Ethnicity --Occupation
• Race For the first time in 2000, the census
  allowed individuals to pick several races to
  reflect their racial make-up. This has generally
  been considered to be a major step in the right
  direction in capturing race accurately.
• Ethnicity The Bureau routinely comes under for
  for only capturing Hispanic under the category
  of ethnicity. Many other groups argue that they
  are an ethnic minority (Italians, Slovakians,
  etc.). The only place this data are captured are
  in a long form question on Ancestry.
• Occupation The 2000 census used a new coding
  system (NAICS) that was implemented in 1997. It
  is generally held that this coding system more
  accurately captures the occupations of workers in
  the United States.

5
Political Context

• To Sample or Not to Sample Throughout the 1990s,
  Congress argued about whether to allow the Census
  Bureau to use sampling procedures to determine
  hard-to-count inner city populations.
• Sampling A statistical procedure that centers
  upon sampling the population and then inferring
  from this sample the total population in the
  country (and the characteristics of this
  population).
• The debate was largely a political
  one Republicans ? Against Democrats ? For
• The Argument Against The Constitution calls for
  an enumeration (count) of the population every
  ten years.
• The Argument For We know that when we enumerate
  we miss lots of people, especially people of
  color and people of lower socioeconomic status.
• The Verdict Sampling was not used for the 2000
  Census
• The Big Point The Census is more than simply
  gathering data it is intertwined with political
  power and resource allocation.

6
Future of the Long Form

• Due to the skyrocketing costs of the Census,
  Congress has looked to changes that would save
  money. One major new initiative
  Discontinuation of the Long Form and replacing
  it with the American Community Survey (ACS)
• The ACS will produce updated yearly estimates for
  all communities, eliminating the need for a
  traditional long form in the 2010 census and
  beyond.
• Starting in year one, the ACS will provide
  estimates of demographic, housing, social, and
  economic characteristics every year for all
  states, as well as for all cities, counties,
  metropolitan areas, and population groups of
  65,000 people or more.
• For smaller areas it will take three to five
  years to accumulate a sufficient sample to
  produce data for areas as small as census tracts.
• For more ACS info see http//www.census.gov/acs/w
  ww/