Library of Congress Prints and Photographs Division Washington, D.C.
This image from the Library of Congress shows two men seated at tables taking the Census at Standing Rock Agency in South Dakota. It was taken between 1880 and 1900.

How to Make the Census Count for Natives

Alexander Ewen & Ivana Maravic

Of the many ways Natives have been abused, cheated and neglected by the United States government, one of the more subtle, but potentially most damaging for tribes today has been the shoddy and sometimes willful miscounting of tribal members in the national census. The authors investigate the long history of this appalling miscarriage, expose the many flaws in the system and point to its many dire implications.

RELATED: No Accounting: Lies, Damn Lies and Indian Country's Census Numbers

With the 1990 count complete for all but a handful of the 50 states, census officials have found once again that the number of people identifying themselves as Indians is up sharply, exceeding numbers that can be explained by biology alone.

Philadelphia Inquirer, March 10, 1991.

The census of the United States has been contentious ever since its first implementation in 1790, when President George Washington lamented that “the indolence of the people and the negligence of many of the officers” had led to an undercount of the population. Since the numbers that the Census Bureau produces every ten years have profound political, social, and financial implications, accuracy is important. For example, the apportionment of the House of Representatives, and to some extent the balance of power in this country, is based on those figures


George Washington

To make things more difficult, by law the Census Bureau was required to count everybody. It could not assume or project numbers. As the country grew larger and the task more daunting, up to a half-million workers would be required to produce a census. Since the census not only counts people, but also profiles their social and economic backgrounds, it is far and away the largest and most comprehensive scientific study conducted in the world. The numbers are used to examine many social issues, not the least of which involve American Indians.

The Conundrum

In the year 1900, the total population of American Indians reached its nadir. According to the census of 1900, the population was just 237,196, a decline of more than 10,000 since the previous decade. A century of devastating wars, forced migrations, confinement in military-run prison camps, starvation and disease had cost untold Indian lives. In the late 19th century the stated purpose of government policy was to stamp out Indians, and according to the numbers produced by the Census Bureau, the policy was working. 

At the turn of the century it was the universal consensus of government officials, scholars, and even Indian advocates that Indians were doomed. The government seemed “incapable of presenting a plan for the civilization of the Indian that does not practically involve his extermination” wrote the New York Times. “The Indian is an expiring race,” explained Senator Henry Teller on the floor of the Senate in 1906, “Whether we have acted wisely or foolishly, the time will come when the red man will disappear.”

Over the next few decades the Indian population seesawed until it began to stabilize by mid-century and remarkably, in 1950 the census recorded 343,410 Indians. Contrary to expectations, Indians were not disappearing. To some extent this was credited to the drastic change in government policy begun in the late 1920s, known as the “Indian New Deal,” which emphasized cultural preservation instead of abolition, put a stop to the unrelenting theft of Indian lands, and ended the dictatorial powers of government agents. Conditions among Indians, who were the poorest of the poor in America, were still grim. However, in the decades of the 1950s and 1960s, advances in medicine—particularly against infectious diseases, which were by then the bane of Indian life—and reforms in Indian health care led to a marked decline in Indian mortality rates and a jump in Indian population. The census in 1960 recorded 523,591 Indians and in 1970 it recorded another sharp increase, to 792,730.

But the large jump from 1960 to 1970 raised eyebrows. In 1976, a demographer in the Population Division of the Census Bureau, Jeffrey S. Passel, wrote an article in the journal Demography entitled “Provisional Evaluation of the 1970 Census Count of American Indians,” where he examined the growth of various age groups and compared this growth with vital statistics data. Passel found that the 1970 count seemed to include 67,000 more Indians than a natural increase would entail. It appeared that “many individuals who were registered as white at birth and who were counted as white in the 1960 census shifted their racial self-identification from white to American Indians during the 1960s.”

The 1960 and 1970 censuses also coincided with a shift in Census Bureau methodology. Before 1960, enumerators had been sent to interview each head of a household. The enumerator would fill out the questionnaires and so indicate a person’s race or heritage.  In 1960, the census mailed the forms to most people, who filled them out themselves, and the enumerators simply picked them up. In 1970, the transition went further, and most people were asked to send their forms back through the mail, although enumerators were still used in rural areas.

If the 1970 census appeared to have overcounted Indians, the 1980 census sent demographers flying off their chairs. According to the Census Bureau the Indian population rocketed to 1,364,033. The explosive growth did not stop there, and in 1990 the census recorded an Indian population of 1,878,285 (not including Eskimos and Aleuts, who are not Indians). The numbers simply did not add up. Passel summed it up in the Population Research and Policy Review journal in a 1997 article entitled, “The Growing American Indian Population, 1960–1990: Beyond Demography.”

Since 1960, the Native American population has exhibited explosive growth, increasing from 552,000 to 1,959,000, or 255 percent. The average annual growth rate of 4.3 percent, extending over a 30-year period, is demographically impossible without immigration—in fact, of the 1.4 million growth only 762,000 comes from natural increase, whereas 645,000 comes from non-demographic factors.

Since presumably thousands of American Indians were not pouring into America on ocean liners and shuffling through Ellis Island, it meant that the 645,000 extra Indians counted by the census had come out of nowhere. Far from being a “vanishing race,” as the 20th century drew to a close, there were too many Indians.

The Cherokee Problem

Following the 1980 returns, researchers quickly zeroed in on the Cherokee, or at least those who claimed to be Cherokee, for being a major part of this numerical puzzle. As explosive as the overall Indian population growth had been in the decade between 1970 and 1980, for the Cherokee, the growth was even more dramatic. From 66,150 in 1970, the population of Cherokee had ballooned to 232,080 in 1980, a whopping 251 percent increase in only ten years, and, according to Joan Nagel in her book American Indian Ethnic Renewal, “a figure that surprised researchers since it outnumbered the historically dominant Navajo tribe by 32 percent.” The Cherokee had suddenly become the largest tribe in the US.

Russell Thorton, in his book The Cherokees: A Population History, analyzed the census figures and estimated that of the 232,000 Cherokees enumerated in 1980, “137,476 represented errors of closure,” in other words could not be accounted for by natural increase. That meant that more than half the Cherokee were not who they said they were, at least not according to the previous census.

Over the next decade the Cherokee population continued to explode, growing at a rate 19 percent higher than the rest of the Indian population, reaching a total of 369,035 in 1990. Nancy Shoemaker, in her book American Indian Population Recovery in the 20th Century argued that:

The spectacular growth [among Indians] since 1970 is primarily due to the popularity of Cherokee ancestry among people reclaiming an Indian identity. Historically high rates of white and Cherokee intermarriage partly explain why so many people today talk about “their Cherokee grandmothers.” But the Cherokees have a well-known history and brand-name appeal (Cherokee Jeans and Cherokee Jeeps, for instance), which probably attracts some people who know they have an Indian heritage but do not know for sure which tribe.

Why people would choose to identify themselves as Indians was not clear. According to the article “Thousands More Claim Indian Roots” written in 1991 by Andrew Cassel for the Philadelphia Inquirer, the experts believed that:

Rediscovered roots and ethnic pride are possible explanations, but so is a fashionable romanticism about Indians. Attendance at powwows and other cultural events is up, and movies and books dealing sympathetically with Indians have been big sellers… Some also see a growing desire to take advantage of programs and laws aimed at boosting Indians economically.

In some respects, the explanation made sense, largely because there was no other explanation. But it seemed odd that people would use the census to make a statement about one’s Indian heritage, since by law the individual’s answers are private and kept secret. Although never enforced, it is also against the law to provide false or misleading information on census forms. And if the theory that a romantic whim was driving the switch from non-Indian to Indian, one might assume that other tribes with “brand name” appeal, such as the Sioux or the Apache, would also show vastly disproportionate growth, but their growth was less than the overall Indian average. The tribes with the next highest rates of growth were the Choctaws and the Chickasaws, not exactly household names. (see Table 1)

Given the ironclad numbers the census was churning out, there was no alternative explanation for the explosion of Indians, except that thousands of non-Indians (with presumably only a distant connection to Indians if that) were willing to lie to the federal government to pretend to be something they were not. It was not possible that the census numbers could be wrong, and even if they were wrong, there was no way to check against them—unless someone else was counting Indians.

A Tale of Two Bureaus

Although the first census was held in 1790, the Census Bureau did not start enumerating all of the Indians in the United States until 1890, one hundred years later. The reason was simple, before 1890 most Indians were not a part of the U.S., they were citizens of their own nations. These Indian nations, such as the Cherokee, had their own governments, educational systems, judiciary, lands, and lived apart from the regular American population. Back then relatively few Indians were assimilated into the American way of life. Only starting in 1870 did the census even begin to include these assimilated Indians in their count of the general population.

With the end of the Indian wars in the 1880s and the final subjugation of the Indian nations, Congress ordered the Census Bureau to include all Indians in the census of 1890. Since the Census Bureau had no experience in enumerating Indians living on their homelands or on reservations, they turned to the agency that did, the Office of Indian Affairs (Indian Office or now the Bureau of Indian Affairs). The Indian Office was started in 1824 under the War Department (now known as the Department of Defense) and its role was to deal with those Indians defeated in war and oversee newly formed Indian reservations.

The War Department played a central role in clearing the way for the expansion of the U.S. from its original 13 states all the way to the West Coast and they had been keen to gather intelligence on Indian populations from the get-go. The department provided its first estimate of Indian population in 1789, the same year it was founded. In 1820 the War Department made the first comprehensive attempt to enumerate the Indian population in the continental U.S. (outside of California and other southwestern states that were still a part of Mexico) arriving at an estimate of 471,036.

In 1836, the Indian Office was given the responsibility of gathering Indian population data. The Indian Office developed a system of “rolls” that kept track of each tribe and their members. The rolls were used to dole out rations or other relief to the Indians, most of whom were in dire straights after their encounters with the War Department. Although the Indian Office justly acquired a reputation as one of the most corrupt, inept, inefficient, and patronage-ridden bureaucracies in the world, the iron grip it held over Indian lives allowed it to collect exceedingly accurate information when called upon to do so. In 1880, the Office counted 240,136 Indians under its jurisdiction, and when added to the census figures for the assimilated Indians, the total number was 306,543–a decline from the previous decade.

Ten years later in 1890, now given the responsibility to enumerate Indians, the Census Bureau hired agents from the Indian Office to help on Indian reservations and they cross-checked their numbers against the Indian rolls. They both arrived at almost identical numbers, 248,300. But after 1890, the Census Bureau’s focus shifted to other areas and they subsequently worked only sporadically with the Indian Office. The counting of Indians involved complex political, legal, as well as racial aspects, and the lack of dedication by the Census Bureau led some, in particular the Indian Office, to question the accuracy of the census figures.

By 1950 the Bureau of Indian Affairs (or BIA as the Indian Office was now called), was convinced that the census was badly undercounting Indians. As far as the BIA was concerned, the Census Bureau could not be troubled to count Indians properly unless it was under political pressure to do so. As the BIA reported in 1956:

In censuses for the years 1890, 1910, and 1930, a special Indian schedule was used and all enumerators were instructed to watch for Indians in the general population. In alternate censuses—1900, 1920, and 1940, for which no such special effort was made—the figures for Indians were disproportionately low.

By 1950, the census figures were 20 percent below BIA estimates. (See table 2)

But the BIA estimates only reflected those Indians that were on BIA rolls or their descendants. There were many Indians that had either never been under BIA jurisdiction or had chosen to leave it.

To make things even more complicated, the BIA recognized that they had lost their ability to track the population of the largest tribe in the country, and the numbers being produced for them were woefully incorrect.

The Cherokee Problem Revisited

In 1890, when census workers traveled to Indian Territory (now the state of Oklahoma) to count the “Five Civilized Tribes” as the Cherokee, Creek, Choctaw, Chickasaw, and Seminole were then called, they found themselves in the midst of a political tinderbox. The Census Bureau reported that:

The citizens of the Five Civilized Tribes watch with a jealous eye each movement of the Unites States and its agents, as questions of vast moment are pending. This made them chary of answering questions proposed by enumerators or special agents.

In Washington, political pressure was building for the breakup of Indian Territory and the dissolution of the Five Civilized Tribes. More than 100,000 whites were squatting on Indian lands and more were pouring in at a rate of 2,000 per month. Railroad companies had lobbied Congress for concessions to Indian lands and minerals, and Wall Street had sold millions worth of railroad bonds in anticipation of the Indians getting the boot. The Census Bureau noted that, “The unsettled condition of the Indian Territory and the constant clashing between the whites, called intruders, and the Indians or their authorities produced a prejudice against the census that was hard to overcome.”

To further complicate matters, the Five Civilized Tribes were in every respect nations with their own laws regarding citizenship.

A serious difficulty was met with the answer to “Are you an Indian?” Under the laws of the Five Tribes or nations of the Indian Territory a person, white in color and features, is frequently an Indian… Negroes are frequently met who speak nothing but Indian languages, and are Indian by tribal law and custom…

The Census Bureau quickly realized that they were hopelessly incapable of sorting through all of these issues. They chose to count only those Indians who were Indian by blood, that is, those that looked like they might be Indians. The decision was fateful, for more than 18,000 people who were citizens of the five tribes were left out of the census count, a total greater than the Navajo population at the time. Although they may not have been Indian by blood, they had been Indian by law. Most of them spoke the native languages, participated in the tribal customs, and, except for their color, were indistinguishable from their Indian neighbors.

The Cherokee population enumerated by the census in 1890, including the Eastern Cherokee and others scattered around the US, was 22,724. Despite the census having excluded 5,394 citizens of the Cherokee Nation who were not deemed to be Indian, the Cherokee were still by far and away the largest tribe in the country. The Navajo, estimated at 17,201, were a distant second.

But that was the last enumeration of the Cherokee that the Census Bureau would undertake with the Indian Office. The 1900 census did not use a special Indian schedule, and by 1910, the Cherokee Nation was gone.

The Lost Tribes

The end came to the Five Civilized Tribes in 1906, when their governments were dissolved and their vast lands broken up. The “legalized robbery” unleashed an “orgy of graft and exploitation,” as Yankton Sioux author Gertrude Bonnin wrote in 1924. More than 1.5 million whites, according to census figures, poured into Indian Territory between 1890 and 1907, leaving a trail of theft and murder that has no parallel in American history.

In the mayhem that preceded the dissolution, thousands of whites petitioned to be registered as Cherokee to share in the spoils, while thousands of traditional, full-blood Cherokee refused to participate in the travesty and many were left off of the last Cherokee roll. Despite the confusion, in 1907 the Indian Office gave their last reasonably reliable figures for the Cherokee using the final Cherokee roll. It reported 36,390 Cherokee by blood, 4,925 black freedmen, and 286 white citizens of the now defunct Cherokee Nation, plus an additional 1,550 Eastern Cherokee.

With the five tribes now dissolved, after this the Indian Office could only estimate the Cherokee population or rely on census figures. But the census figures were coming in low. Although the Indian Office had counted 37,940 Cherokee by blood in 1907, three years later, in 1910, the census only enumerated 31,489. Even so, the Cherokee were still by far the largest tribe, with the Navajo population far behind at 22,455 according to the census.

The Census Bureau was roundly criticized for its lackluster effort in the 1920 census, which provided little usable information about Indians. It tried to make up for it by going all out in 1930 and for the last time worked closely with the Indian Office. As in 1890, both departments’ figures were very similar, but the Indian Office was forced to rely on the census’s count of the Five Civilized Tribes. The census enumerated 45,238 Cherokee, making it still the largest tribe. The Navajo were closing the gap, with 39,064 members.

The 1930 census would be the last enumeration of the Cherokee for the next forty years. When in 1970, the census again released tribal population figures, the Navajo were number one, with a population of 96,743. The tables had turned and the Cherokee were now a very distant second, at 66,150. For some reason the Cherokee population growth had slowed to a crawl.

Indeed for the two decades after 1930 the entire Indian population stopped growing. Whereas in the decade between 1920 and 1930 the census had recorded a gain of almost 88,000 Indians, in the next decade, from 1930 to 1940, the growth was less than 2,000. Yet, as opposed to the 1920s, when starvation among Indians was rampant, the period after 1930 was relatively prosperous. The Indian New Deal gave thousands employment opportunities, as did the mobilization for World War II. The land grabs were largely ended, corruption in the Indian Office was reduced, the brutal boarding schools phased out, and the dictatorship of government agents done away with. Indians in the 1930s and 40s were once again allowed to form their own governments, practice their religions, and live more freely than they had since the days before their conquest. One might have expected a baby boom.

Something was very wrong with the census numbers.

The Undercount

It was never easy to enumerate Indians. In the first place, the definition of who is an Indian varied among different government departments and changed over time. Starting in 1960, the census added the indigenous people of the new state of Alaska to the total count, skewing the numbers and making comparisons difficult between the 1950 and 1960 census. The census also grouped Eskimos and Aleuts, who are considered different peoples, with the Indians, fostering much confusion.

For most of the century the majority of Indians lived on isolated lands featuring few paved roads that were difficult to get to, even after the widespread introduction of the automobile. Many did not have addresses to which forms could be mailed to or to help enumerators locate them. Being an Indian was often a matter of culture as well as race. For much of the last century, the view of many Americans (and thus many census enumerators) was that real Indians only lived in teepees and wore their traditional dress.

In addition, Indians are not one people, but are made up of hundreds of different nations, speaking different languages, with a variety of cultures, physical features, and degrees of assimilation to American society. Even the catchword “American Indian” is controversial, but no other substitute has been found. The newer term “Native American” is also applied to any American citizen born in the U.S., and so is not suitable when discussing demographics.

The census was hamstrung by its requirement to count every person. For most of the early part of the century, it knew full well it had failed to enumerate thousands of Indians, because it could compare its numbers directly with the Indian Office and the tribal rolls. But it was required to publish the numbers its census enumerators had come up with, whether they had done a good job or not, and was not allowed to provide estimates.

Therefore the undercounting started from the very beginning, with the census of 1890. A comparison with the previous census and with Indian Office tribal rolls shows that the 1890 census missed more than 56,000 Indians, or about 23 percent of the total. Things only went downhill from there.

Another Look at the Numbers

By 1910, the Indian Office estimate for Indians under its supervision was almost 40,000 above the census total for that year. Yet the Indian Office numbers did not include those Indians not under its jurisdiction; meaning in 1910, the census had failed to count more than 100,000 Indians, or almost 40 percent of the total. And this figure did not include the growing disparity in the count of The Five Civilized Tribes.

Before 1906, The Five Civilized Tribes represented, even without their non-Indian citizens, a full 20 percent of the total Indian population. The Cherokee were the largest tribe in the country and the Choctaw (4th largest), Creek (7th largest), and Chickasaw (11th largest) were not far behind. But their dissolution, and the subsequent fragmentation of their populations, made it difficult to track their numbers.

The Citizen Potawatomi, based in Indian Territory, also suffered the same fate as The Five Civilized Tribes, but because their tribal roll system was left intact, the Indian Office managed to continue to track their numbers after 1906. In the years when both the Indian Office and the Census Bureau enumerated the Citizen Potawatomi and the numbers can be compared, the results are enlightening. In 1910, the census only counted 866 Citizen Potawatomi, whereas the Indian Office had 1,768 on their rolls. In 1930, the census only counted 636, whereas the Indian Office counted 2,572.

The census was heavily undercounting The Five Civilized Tribes, but there was no way to know by how much. After 1930, the increasingly desultory efforts on the part of the Census Bureau made it impossible to compare census numbers for each tribe (since the census no longer reported them) with their tribal rolls. Because of the combined size of the Five Tribes, the inability to properly enumerate them skewed the entire Indian population count. In the meantime, improving economic conditions and health care was leading to a drop in Indian mortality rates, which should have been reflected in the census, but was not.

It was difficult enough to enumerate Indians when the Census Bureau made it a priority, but the lack of effort after 1930 meant that by the 1950 census more than a third of all the Indians, over 250,000, were left out.

The Need for an Accurate Census

With the conquest of infectious diseases and improved health care after 1950, Indian populations did begin to grow rapidly, but there was never an unnatural “population explosion” of Indians between 1960 and 1990. For most of the century the census simply was unable to provide accurate population statistics for Indians. It was only in 1980, when Indians were allowed to count themselves—as opposed to being counted—and following a massive campaign by the Census Bureau to get the census forms into Indian hands, did the true numbers begin to materialize.

That is why those tribes that were the most undercounted, such as the Cherokee, Choctaw, Chickasaw, Seminole, and Potawatomi, saw the highest rates of “growth” following 1960. Although there were undoubtedly some non-Indians switching their identity to become Indians, the vast majority of the growth was due to misidentified Indians or those who had never been enumerated before finally setting the record straight. They were joined by tens of thousands of the descendants of the legal citizens of the Five Civilized Tribes who were never counted by the census, but who never lost their belief in their Indianness, despite what others had decreed.

Unfortunately the 1990 census, by far the most accurate and comprehensive record of Indian population since 1930, became mired in controversy because of the perceived overcount. For 2000, the Census Bureau changed their methodology, allowing a mixed-race category that only made the enumeration process more complex and less accurate. In 2010 the Bureau reduced their systems for tracking Indian populations and their social conditions.

Sadly, the census numbers are once again virtually useless for understanding Indian populations, forcing researchers to look for other alternatives. However since the 1950s, the BIA’s ability to track Indian populations has diminished to the point that they long ago stopped releasing their own estimates.

The need for accurate figures is important to understand the many social changes that are underway among Indian people, and it is hoped that in the future the Census Bureau may once again put the time and effort to secure accurate and usable numbers. This new effort will require a better understanding of Indian people, their history, and their cultures, for the truth is that enumerating Indians is much, much more than simply a question of numbers.

Alexander Ewen, Purepecha, is the director of the Solidarity Foundation and the author (with Jeff Wollock) of the forthcoming Encyclopedia of American Indians in the 20th Century (Facts on File: 2013). Ivana Maravic is a research associate and mathematician and statistician for the Solidarity Foundation. Ewen and Maravic are working on the forthcoming Demographic History of American Indian Languages.



chahta ohoyo's picture
chahta ohoyo
Submitted by chahta ohoyo on

im going to lay this at the doorways where it belongs...the individual tribes themselves, and the BIA
because the fact is that in spite of the 'popularity' of being 'cherokee', there are probably hundreds of thousands of native peoples living in and amongst white society who, because of refusal of their tribes and the united states govt policies of not enumerating those of us who are not DIRECTLY descended from our own damn family members who ARE enumerated, are standing out here with our hat in our hands with NO possibility of ever ever ever becoming 'OFFICIAL' indians
this is sickening and disgusting...i would think we would be welcomed and wanted if for no other reason than the power of numbers as a political entity...i truly believe that our existence is the bane of OUR tribes who see their federal govt dollars diminishing exponentially by the addition of all we outsiders...i am accepted by/acceptable to the choctaw nation because i have an absolutely undeniable blood connection back to one of the founding families of that nation, but because my ggg aunt was 'officially' enumerated just because she happened to be at the right time and place to be enumerated, and, because her own blood sister was not, i never, at least at this point in time, have the faintest chance of becoming a 'tribally enumerated' choctaw...but, i have come to accept my status because, becoming 'official' carries with it a whole new set of rules and regulations that i, as a non-reservation, living freely in white society NDN dont ever have to cope with...i just want to be recognized as equal to/as good as the rest of my people...