Page 37 - IJPS-11-1

International Journal of
Population Studies Indian Census Rolls: Data for U.S. tribes

used the child-woman ratio (Baker et al., 2017, pp. 46 – 48) information. For illustration, we draw on the 1937 census
to represent fertility. The CCM also requires information roll for the Hopi tribe, located in Northeast Arizona, to
on survivorship by age and gender, so for the mortality create a historical series for the Hopi population from
component, Swanson (2022) employed “life tables” specific 1900 to 1940 by age and gender and work in the form of a
to birth cohorts by year from 1900 to 2100 (Bell & Miller, backcast that provides decennial age-gender information
2005) to survive them to extinction. No information on for the Hopi from 1940 to 1900 (Swanson & Tayman, 2024).
migration was needed because the population is essentially Eight steps were used complete these estimates:
closed to migration—the data represent everybody who (1) Assembled an age-gender distribution from the
is a member of the tribe. If the projection involved only 1937 Hopi Indian census roll along with a projection
the Hopi Tribal Reservation area, migration on and off the of this population to 1942, found in Swanson (2022);
reservation would have to be accounted for in the cohort- (2) interpolated between the 1937 and 1942 data to
component model (Greene, 2021). Using these data, the obtain an estimated age-gender distribution for Hopi
CCM was run in 5-year cycles to the year 2017, with an tribal members in 1940; (3) developed decennial rates of
80-year projection horizon. To assess their accuracy, the change by age between 1900 and 1940, assembled in an
projections for 1992, 1997, 2002, 200, 2012, and 2017 earlier study from the Hopi tribal data collected by Census
were compared to Hopi Tribal membership data (available Bureau in the 1900 census (Johansson & Preston, 1978);
annually from 1989 to 2018). These projections generally (4) generated the 1910, 1920, and 1930 tribal populations
performed very well (Swanson, 2002: 1842 – 1844). by age using the decennial linear change calculated in step

The specifics of the preceding overview are found in (3); (5) computed the ratios of females/total population by
an Excel workbook available from the authors, “HOPI age for 1900 and 1940; (6) applied the interpolated ratios
LEXUS DIAGRAM & FORECAST V5.xlsx.” It contains to estimate the female populations by age for 1910, 1920,
worksheets that include the initial 1937 Hopi population and 1930; (7) subtracted the females by age from the total
by age and gender along with the life tables used to survive population by age to estimate the male population by age
the 1937 population by age and gender to extinction as for 1910, 1920, and 1930; and (8) assembled the decennial
well as the subsequent birth cohorts (in 5-year cycles, i.e., age-gender data and present them for 1900 to 1940 by
1942, 1947, 1952,…, 2017) to extinction. The birth cohorts decade. A paper by Swanson & Tayman (2024) contains
were generated in 5-year cycles by applying, as stated details of the data and methods used in these estimates.
above, the Child-Woman Ratio, which in the case of the Table 5 shows population estimates by age and gender
Hopi consisted of those aged 0 – 4 to females aged 20 – 20. from 1900 to 1940 by decade for the Hopi tribe, which
This workbook contains all of the input data and Excel shows an increase of 1,635 between the total population
codes needed to accomplish: (1) The generation of births; estimated for 1940 (3,487) and the 1900 number (1,852).
(2) the survivorship of the initial age structure and the As evidence of the reliability of these estimates of the total
subsequent birth cohorts; and (3) the assembly of the birth Hopi tribal population, we look at the estimates provided
and survivorship results into a series of projections from by Kunitz (1974a, p. 9) that have dates matching up closely
1942 in 5-year cycles to the projection target year, 2102. In with the dates for which we have numbers. Corresponding
addition, a Lexus Diagram is included in this workbook to the 1900 number of 1,852 produced by Johansson &
that assembles the counts for each age 5-year group (0 – Preston (1978, p. 5) is a 1901 estimate of 1,841; and our
4 to 115 – 119) over the projection horizon, which is in 1930 estimate of 3,079, is an estimate of 2,752. Finally,
5-year cycles, 1937, 1942, 1947,…, 2092, 2097, 2102. corresponding to our 1940 estimate of 3,348 is an estimate
The work by Swanson (2022) suggests that his approach of 3,444.
is worth considering where long-term projections are Between 1900 and 1940, the Hopi tribal population
required for a tribal population. The CCM process applied experienced an increase of 88%, which is consistent with
to the Indian Census Rolls done between 1934 and 1940 the findings by Kunitz (1974a, p. 9). Looking at age-specific
can also be used to generate “estimates” that would provide percent changes, all but two age groups showed population
a perspective on changes in the size of the AIAN population growth between 41% and 99%. The population size of the
in general as well as in regard to specifical tribes (Liebler age group 40 – 49 had grown by 126% and that of those
et al., 2016; Passel, 1996; Thornton, 1979). aged 70+ by 442% between 1900 and 1940. These changes
likely reflect differential errors between the 1940 estimates
4.2. Filling gaps in historical tribal data
by age and the 1900 estimates by age, as well as intervening
Indian census rolls can help fill the gap in historical events that impacted the 1900 population as it aged. Some
American Indian tribal demographic and household intervening events include the presence of endemic and

Volume 11 Issue 1 (2025) 31 https://doi.org/10.36922/ijps.3906