Projections of the Voting-Age Population for States: November 1998 - Methodology
November 3, 1998 voting-age population projections by state are provided
of those aged 18 and over, 18 to 24, 25 to 44, 45 to 64 and 65 and over by
gender for the three race groups, White, Black, Other, and the two origin
groups, Hispanic and Non-Hispanic. The total of the three race groups or
the total of the two origin groups add up to the total voting-age
population. The voting-age population is projected to November 3, 1998,
the actual date of the election, rather than November 1, 1998 to
distinguish this product from many other products that utilize a date of
November 1, 1998.
A brief description of the methodology used to construct these projections
follows. The exact manner of construction was dependent upon the relevant
data available at the time of creation (January 1998) as well as expert
- The "vintage 96" 1995 and 1996 detailed (single year of
age-gender-race-origin) population estimates as of July 1 for each
year are aggregated into the required age/gender/race/origin
cells. This is done for each state and the District of Columbia.
The required aggregations are the five age groups, 0 to 17, 18 to
24, 25 to 44, 45 to 64 and 65 and over by gender, race and
origin. Estimates for July 1 are used.
- The "vintage 96" aggregated estimates are then proportionally
adjusted to the "vintage 97" state totals for 1995 and 1996.
- Each adjusted July 1996 cell value, sometimes referred to as
796R, is carried forward to July 1, 1997 by applying a ratio (1996
adjusted cell value/1995 adjusted cell value) to the adjusted July
1996 cell. In this way, the population trends exhibited between
1995 and 1996 are carried forward to 1997.
- These "initial" July 1997 state cell values are then
proportionally adjusted by the ratio of the "vintage 97" state
total population on July 1, 1997 to the sum of the "initial" July
1, 1997 cell values for that particular state. The resultant
numbers are sometimes identified as 797R.
This state-level data has not yet been adjusted to be consistent
with the national-level data. This adjustment is accomplished
through the following two steps.
- The adjusted July 1996 (796R) and July 1997 (797R) state
values for each cell are then used to create estimates for
November 1, 1996 and November 1, 1997. The formulas are shown
November 1, 1996 = ((123/365) * (797R - 796R)) + 796R
November 1, 1997 = ((123/365) * (797R - 796R)) + 797R
- The sum of the 51 values (50 states plus the District of
Columbia) of each cell on November 1, 1996 and November 1, 1997
are then adjusted to be consistent with the national cell values
for that date. These values are referred to as 1196R and 1197R.
The 18 and over population of each state in 1196R is the November
1, 1996 voting-age population.
- To obtain the voting-age population for November 3, 1998, the
following formula is applied:
November 3, 1998= ((367/365) * (1197R -1196R)) + 1197R)
- As a final adjustment, each state age-gender-origin cell value
is adjusted to be consistent with the age-gender-race values.
- These projections are based upon the results of the 1990
Census as enumerated. They include all residents of the United
States, whatever their voting status.
- This method ignores cohort trends and implicitly assumes that
population cells will display short-run stability in their rates
- It assumes the November 3, 1998 population can be approximated
by the July 1996 to July 1997 state population trends.
For more information, contact the Population Projections Branch, 301-457-2428.