In order to
make population projections by white and nonwhite racial groups using 1990
Census to 2000 Census trends, I had to come up with a definition of
"White" based on 2000 Census definitions which matched the definition
of "White" used for the 1990 Census MARS file. For the 1990 Census MARS file, the
distribution of "Other Race" into the four categories used at the
time, "White", "Black", "Asian", and
"Indian", resulted in most of the "Other Race" population
being added to the "White" population. A major reason for this result was that the overwhelming majority
(over 93%) of
the "Other Race" population was Hispanic and the vast majority (78.9%) of the Hispanic
population which was not "Other Race" was "White". One could easily assume that many people
were confused, thinking that "Hispanic" was a racial category, a
subcategory under "Other Race".
Since over 3/4 of the Hispanic population which listed one of the other
four categories listed themselves as "White", one could assume that
the vast majority of Hispanics who listed themselves as "Other Race"
were actually "White".
The 2000 Census data show a similar pattern. Similar
to the results in 1990, the overwhelming majority (over 95%) of the one-race "Some Other Race" population was
Hispanic and the vast majority (88.4%) of the one-race Hispanic population which was not
"Some Other Race" was "White". In considering a definition for 2000, I also
noted that calendar year 2000 births to Hispanic mothers in North Carolina were
overwhelmingly (more than 98%) "White". Over 93.5% of calendar year 2000 Hispanic deaths in North
Carolina were "White", as well.
Hence, I assumed that the fraction of one-race Hispanics who actually
were "Some Other Race" was
the same as the fraction of one-race Non-Hispanics who listed themselves as
"Some Other Race".
From the
population listing themselves as of one-race, the number included in my true "White"
population is the listed
"White" total plus the number of Hispanics listed as "Some Other
Race" minus
the number of Hispanics who actually were of "Some Other Race". This latter value is assumed to be the total number of one-race Hispanics
multiplied by the fraction of one-race Non-Hispanics who listed themselves as
"Some Other Race".
For the population listing themselves as of two-races, in most
cases, I assumed that half of the population was counted in one race group and
one half in the other race group.
Specifically, I added one half of the sum of the
"White/Black", the "White/Indian", the
"White/Asian", and the White/Pacific Islander" populations to my
true "White" population.
An exception was made for
the "White/Some Other Race" population. One half of the sum of the Non-Hispanic "White/Some
Other Race" population was added to my true "White"
population. For the Hispanic
"White/Some Other Race" population, the same assumption was made as
for the one-race Hispanic "Some Other Race" population, with
the variation that many Hispanic people who listed themselves as two-race
"White/Some Other Race" were just one-race Hispanic
"White". This assumption is
supported by the fact that the majority (62.4%) of those who listed
themselves in 2000 as "White/Some Other Race" were Hispanic. Thus, first, I added all of the listed Hispanic White/Some Other
Race" population to my true "White" population. Then, I estimated the true Hispanic
White/Some Other Race" population as the total number of two-race
Hispanics multiplied by the fraction of two-race Non-Hispanics who listed
themselves as "White/Some Other Race". Finally, I subtracted out half of this estimate from my true
"White" population.
For the population groups of
three races, if one of the races was "White", I added 1/3 of the
population of each group to my true "White" population. I did not attempt to add any portion of
population groups of four or more races to my true "White"
population.