← Back to home

United States Census

Alright, let's get this over with. You want me to… rewrite this Wikipedia article. As if the dry, factual recounting of government bureaucracy isn't painful enough. Fine. But don't expect any sunshine and rainbows. This is the U.S. Census, after all. It’s a necessary evil, much like… well, like dealing with you, I suppose.

The Decennial Census: Mandated by the Constitution, Orchestrated by Bureaucrats

The Bureau of the Census, a rather drab entity nestled within the labyrinthine structure of the United States Department of Commerce, is the entity tasked with this… enumeration. They call it a "decennial census," a fancy term for counting everyone, everywhere, every ten years. It's a constitutional mandate, a foundational pillar of this nation's peculiar brand of democracy.

Frequency: Every ten years. Because apparently, a decade is the perfect amount of time for things to get sufficiently out of hand before we need to check again. Location: 4600 Silver Hill Rd., Suitland, Maryland, 20746. A place where dreams go to be filed. Country: The United States. Obviously. Inaugurated: August 2, 1790. A mere 235 years ago. That’s practically ancient history. Most Recent: April 1, 2020. Just a few years ago. You probably remember the chaos. Next: April 1, 2030. Mark your calendars. Or don't. It’ll happen regardless. Website: census.gov. If you have nothing better to do.

The United States census, or "censuses" if you’re feeling particularly pedantic, is a decennial undertaking. It's a requirement etched into the very fabric of the Constitution of the United States. The first one, post-American Revolution, was a rather quaint affair back in 1790. Thomas Jefferson, then Secretary of State, oversaw it. Since then, there have been 24 of these… enumerations. And yes, they even count the territories of the United States. Because apparently, everyone needs to be accounted for. The United States Census Bureau is the designated conductor of this grand, often tedious, symphony of numbers.

The last national headcount was in 2020. The next one looms in 2030. By 2013, the Census Bureau, in its infinite wisdom, was already contemplating how to inject some semblance of modernity into the process, starting with the 2020 census. They wanted to use technology. Imagine that. In 2020, every household was expected to participate, either online, by phone, or, for the stubbornly analog among us, via a paper questionnaire. It’s a veritable smorgasbord of data collection.

In the interim years, between these monumental decennial events, the Census Bureau churns out estimates. They use surveys, statistical models… the works. The Population Estimates Program and the American Community Survey are the usual suspects in this ongoing statistical drama.

From 1940 to 2000, they also employed a "long form." This wasn't for everyone, mind you. Just a select subset of households got the privilege of answering more intrusive questions about their socioeconomic status and housing arrangements. A little extra peek behind the curtain, if you will.

It’s important to note that this decennial census is distinct from, say, the United States Census of Agriculture. That’s a whole different beast, and not even the Census Bureau’s responsibility anymore. And it’s certainly not to be confused with those smaller, local censuses some states or jurisdictions might conduct. This is the big one. The federal one.

The Legal Framework: Why They Can Actually Do This

The mandate for the U.S. census is laid out in Article I, Section 2 of the United States Constitution. It’s quite explicit: "Representatives and direct Taxes shall be apportioned among the several States... according to their respective Numbers... . The actual Enumeration shall be made within three years after the first meeting of the Congress of the United States, and within every subsequent Term of ten Years." It’s a foundational piece of legislation, ensuring that representation, at least in theory, is based on population.

Later, Section 2 of the Fourteenth Amendment added a crucial clarification: the "respective Numbers" would be determined by "counting the whole number of persons in each State, excluding Indians not taxed." A rather specific exclusion, wouldn't you agree? The United States Census Bureau, officially known as the Bureau of the Census according to Title 13 U.S.C. § 11, is the official body tasked with this monumental undertaking. And yes, they answer to the United States Department of Commerce. Because where else would you put a government agency that deals with numbers and people?

The nitty-gritty of how this whole operation is conducted, and how the data is handled, is all laid out in Title 13 of the United States Code. Confidentiality is paramount, at least in theory, as stipulated in 13 U.S.C. § 9. But don't think you can just ignore them. The Sentencing Reform Act of 1984, embedded within Title 18 of the United States Code (Sections 3551, 3559, and 3571), makes it clear: failure to respond or providing willfully false answers can land you with penalties of up to $5,000. They’re not playing around.

The Procedure: Counting Sheep, or People?

The decennial U.S. census is, at its core, a count of actual people residing in U.S. residential structures. This includes citizens, legal residents, long-term visitors, and, yes, even undocumented immigrants. The guiding principle, established way back in 1790, is "usual residence" – where a person lives and sleeps most of the time. They have special procedures for those without conventional housing, though the accuracy of those counts is… questionable.

When the Bureau is uncertain about the number of residents at a particular address, even after a visit, they resort to inferring the population characteristics. They do this by looking at the nearest comparable neighbor, a statistical technique known as hot-deck imputation. It’s a controversial practice, to say the least, but the U.S. Supreme Court deemed it constitutional in Utah v. Evans. So, there’s that.

Now, American citizens living abroad are a special case. They are generally excluded from the census count, unless they are federal employees (military or civilian) or their dependents. Private citizens living abroad, disconnected from the federal apparatus, are simply not counted. These overseas counts are strictly for the purpose of reapportioning seats in the U.S. House of Representatives.

Since 1930, "Census Day" has officially been April 1st. Before that, it was a bit more fluid, bouncing around between August and June. The idea is to capture a snapshot in time, to avoid double-counting people who are constantly being born, dying, or moving. The actual counting, however, starts well before April 1st and can drag on for months. In 2020, they started as early as January 21st in remote parts of Alaska. The rest of us got our official start around March 12th.

Applications: More Than Just Apportionment

While reapportioning the House of Representatives is the primary goal, the census data has a far wider reach. It influences:

  • Federal Funding: An estimated 675billionto675 billion to 1.5 trillion per year in federal funding is distributed based on census data. That’s a lot of money hanging in the balance.
  • Infrastructure and Transportation Planning: Where do we build roads? Where do we need public transit? The census data provides the foundation.
  • Military and Disaster Response: Knowing where people are, and how many there are, is crucial for planning in emergencies.
  • Economic Analysis: Businesses use this data to understand markets and make investment decisions.
  • Commercial Use: Marketing departments salivate over this data.
  • Disambiguation: Computer programs use it to figure out place names based on population.
  • General Reference: It’s the go-to source for all sorts of information.

Controversy: Because Nothing is Ever Simple

This section, as indicated, needs an update. But even without the latest drama, the census has always been a hotbed of controversy. Its role in reapportionment has been a constant source of contention.

  • The 1850s: Slavery was a thorny issue, and census planners, under pressure from Southern lawmakers, suppressed information about it.
  • The 1920 Census: The results were simply ignored. Reapportionment stalled because rural lawmakers feared losing power to burgeoning urban centers.
  • The 1940s: A dark chapter. Census officials were complicit in the Japanese-American internment. A chilling reminder of how data can be weaponized.

The nonresponse rate is another ongoing concern. While the Census Bureau often claims it's less than 1%, some experts believed the 2020 census could see double-digit nonresponse. The pandemic certainly didn't help. By October 2020, most states had achieved a response rate above 99%, but the initial fears were palpable.

Then there's the issue of undercounting. The Bureau itself has acknowledged that in 1970, over six percent of African Americans were missed, compared to about two percent of European Americans. Democrats tend to favor sampling techniques for more accurate data, while Republicans insist on an "actual enumeration" as per the Constitution, fearing political manipulation of sampling formulas.

Groups like the Prison Policy Initiative highlight another point of contention: counting prisoners as residents of their correctional facilities, rather than their pre-incarceration addresses. This skews demographic and population numbers, particularly concerning racial demographics.

The 2020 Census: A Modern Minefield

The 2020 census was particularly fraught, largely due to the Trump administration's policies on undocumented immigration. The initial attempt to add a citizenship question was met with fierce opposition, with fears that it would lead to widespread nonresponse and undercounting. The Supreme Court, in Department of Commerce v. New York, found the administration's rationale for the question "arbitrary and capricious." The question was ultimately dropped, but Trump later issued an executive order to exclude undocumented immigrants from the apportionment count. This led to further legal battles, culminating in the Supreme Court case Trump v. New York.

The COVID-19 pandemic threw another wrench into the works, delaying data collection and leading to further legal challenges regarding shortened timelines. It was a mess. A meticulously documented, statistically significant mess.

A Brief History: From Marshals to Machines

The practice of census-taking predates the Constitution. Early colonial censuses were conducted in the 17th century. Between 1781 and 1786, a preliminary "actual enumeration" was compiled by John Kean for the Constitutional Convention in 1787.

As the nation grew and its needs became more complex, so did the census. Inquiries expanded from simple population counts to include manufacturing, fisheries, and even social issues like taxation, churches, pauperism, and crime by 1850. The sheer volume of data collected by 1880 necessitated a change. Enter Herman Hollerith and his revolutionary tabulating machines in 1890. This technological leap drastically reduced processing time, from nearly a decade to a mere two and a half years.

For the first nine censuses (1790–1870), the task fell to U.S. Marshals and their appointed assistants. These enumerators, often local figures, had a more intimate, albeit less standardized, understanding of the population. Racial classifications were made by the enumerator, not the individual.

The table below, a stark reminder of the evolving demographics and classifications, details the census data by year. It's a grim, fascinating chronicle.

| Year | Total population | Change in population | Most populated state | Most populated city | Ethnic demographics counted | Slaves | Notes 1790-08-02) | Most populated city | New York, NY (33,131) | Free white females and males, other free persons, slaves | 694,280 | Original numbers were corrected later. | | 1800 [d] | 5,308,483 [e] | 35% | Virginia [f] (676,682) | New York, NY (60,515) | Free white females and males, other free persons, slaves | 893,605 [g] | Original numbers were corrected later. | | 1810 [h] | 7,239,881 | 36% | New York (959,049) | New York, NY (96,373) | Free white females and males, other free persons, slaves | 1,191,362 | The authorizing act of the third census stipulated that each marshal (enumerator) must personally visit each household, rather than rely on hearsay. | | 1820 [i] | 9,638,453 | 33% | New York (1,372,812) | New York, NY (123,706) | Free white females and males, other free persons, free people of color, slaves | 1,538,022 | — | | 1830 [j] | 12,866,020 | 33% | New York (1,918,608) | New York, NY (202,589) | Free white females and males, other free persons, free people of color, slaves | 2,009,043 | — | | 1840 [k] | 17,069,453 | 33% | New York (2,428,921) | New York, NY (312,710) | Free white females and males, other free persons, free people of color, slaves | 2,487,355 | The census estimated the population of the United States at 17,100,000. The results were tabulated by 28 clerks in the Bureau of the Census. | | 1850 [l] | 23,191,876 | 36% | New York (3,097,394) | New York, NY (515,547) | Black, Mulatto, White | 3,204,313 | The 1850 census was a landmark year in American census-taking. It was the first year in which the census bureau attempted to record every member of every household, including women, children and slaves. Accordingly, the first slave schedules were produced in 1850. Prior to 1850, census records had only recorded the name of the head of the household and tabulated the other household members within given age groups. | | 1860 [m] | 31,443,321 | 35% | New York (3,880,735) | New York, NY (813,669) | Indian, Chinese, Black, Mulatto, White | 3,953,761 | The results were tabulated by 184 clerks in the Bureau of the Census. This was the first census where American indigenous people officially were counted, but only those who had 'renounced tribal rules'. The figure for the nation was 40,000. | | 1870 [n] | 39,818,449 [o] | 23% | New York (4,382,759) | New York, NY (942,292) | Indian, Chinese, Black, Mulatto, White | — | The first census to provide detailed information on the black population, only years after the culmination of the Civil War when slaves were granted freedom. The results are controversial, as many believed it underestimated the true population numbers, especially in New York and Pennsylvania. | | 1880 [p] | 50,189,209 | 30% | New York (5,082,871) | New York, NY (1,206,299) | Indian, Chinese, Black, Mulatto, White | | The first census that permitted women to be enumerators. Also led to the discovery of Alabama paradox. | | 1890 [q] [n 1] | 62,947,714 | 25% | New York (6,003,174) | New York, NY (1,515,301) | Indian, Chinese, Japanese, Black, Mulatto, Quadroon, Octaroon, White | | Because it was believed that the frontier region of the United States no longer existed, the tracking of westward migration was not tabulated in the 1890 census. [31] This trend prompted Frederick Jackson Turner to develop his milestone Frontier Thesis. The 1890 census was the first to be compiled using the new tabulating machines invented by Herman Hollerith. The net effect of the many changes from the 1880 census (the larger population, the number of data items to be collected, the Census Bureau headcount, the volume of scheduled publications, and the use of Hollerith's electromechanical tabulators) was to reduce the time required to fully process the census from eight years for the 1880 census to six years for the 1890 census. [32] The total population, of 62,947,714, was announced after only six weeks of processing (punched cards were not used for this family, or rough , count). [33] [34] The public reaction to this tabulation was disbelief, as it was widely believed that the "right answer" was at least 75,000,000. [35] This census is also notable for the fact it is one of only three for which the original data are no longer available. Almost all the population schedules were destroyed following a fire in 1921. | | 1900 [r] | 76,212,168 | 21% | New York (7,268,894) | New York, NY (3,437,202) | Indian, Chinese, Japanese, Black (Negro or of Negro descent), White | — | | | 1910 [s] | 92,228,496 | 21% | New York (9,113,614) | New York, NY (4,766,883) | Indian, Chinese, Japanese, Black (Negro), Mulatto, White, other | — | | | 1920 [t] | 106,021,537 | 15% | New York (10,385,227) | New York, NY (5,620,048) | Indian, Chinese, Filipino, Hindu, Japanese, Korean, Black (Negro), Mulatto, White, other | | This was the first census that recorded a population exceeding 100 million. | | 1930 [u] [n 2] | 122,775,046 | 13% | New York (12,588,066) | New York, NY (6,930,446) | Indian, Chinese, Filipino, Hindu, Japanese, Korean, Negro, Mexican, White, other | — | | | 1940 [v] | 132,164,569 | 7% | New York (13,479,142) | New York, NY (7,454,995) | Indian, Chinese, Filipino, Hindu, Japanese, Korean, Negro, White, other | — | | | 1950 [w] | 150,697,361 | 14% | New York (14,830,192) | New York, NY (7,891,957) | American Indian, Chinese, Filipino, Japanese, Negro, White, other | | The most recent census where individuals' data have now been released to the public (by the 72-year rule). | | 1960 [x] | 179,323,175 | 19% | New York (16,827,000) | New York, NY (7,781,984) | Aleut, American Indian, Eskimo, Chinese, Filipino, Japanese, Negro, Hawaiian, part-Hawaiian, White | | Will be available for public inspection on April 1, 2032. | | 1970 [y] | 203,302,031 | 13% | California (19,953,134) | New York, NY (7,894,862) | American Indian, Chinese, Filipino, Japanese, Korean, Negro or Black, Hispanic origin, Mexican, Puerto Rican, Cuban, Central or South American, Hawaiian, White, other | — | The first census that recorded a population exceeding 200 million. Will be available for public inspection on April 1, 2042. | | 1980 [z] | 226,545,805 | 11% | California (23,667,902) | New York, NY (7,071,639) | Aleut, Eskimo, American Indian, Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, Black or Negro, Hispanic origin, Mexican, Mexican-American, Chicano, Puerto Rican, Cuban, other Hispanic, Hawaiian, Guamanian, Samoan, White, other | | Will be available for public inspection on April 1, 2052. | | 1990 [aa] | 248,709,873 | 10% | California (29,760,021) | New York, NY (7,322,564) | Aleut, Eskimo, American Indian, Asian or Pacific Islander, Chinese, Filipino, Korean, Vietnamese, Japanese, Asian Indian, other API, Black or Negro, Hispanic origin, Mexican, Mexican-American, Chicano, Puerto Rican, Cuban, other Hispanic, Hawaiian, Guamanian, Samoan, White, other race | | Will be available for public inspection on April 1, 2062. | | 2000 [ab] | 281,421,906 | 13% | California (33,871,648) | New York, NY (8,008,278) | American Indian, Alaska Native, Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, other Asian, Black, African American, or Negro, Hispanic origin, Mexican, Mexican-American, Chicano, Puerto Rican, Cuban, other Hispanic, Hawaiian, Guamanian or Chamorro, Samoan, other Pacific Islander, White, other race | | Will be available for public inspection on April 1, 2072. | | 2010 [ac] | 308,745,538 | 10% | California (37,253,956) | New York, NY (8,175,133) | American Indian, Alaska Native, Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, other Asian, Black, African American, or Negro, Hispanic origin, Mexican, Mexican-American, Chicano, Puerto Rican, Cuban, other Hispanic, Hawaiian, Guamanian or Chamorro, Samoan, other Pacific Islander, White, other race | | The first short-form-only census since 1940, as the decennial long form has been replaced by the American Community Survey. The first census that recorded a population exceeding 300 million. Will be available for public inspection on April 1, 2082. | | 2020 [ad] | 331,449,281 | 7% | California (39,538,223) | New York, NY (8,804,190) | American Indian, Alaska Native, Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, other Asian, Chamorro, Native Hawaiian, Samoan, other Pacific Islander, Black or African American, Hispanic, Latino or Spanish origin, Chicano, Cuban, Mexican, Mexican American, Puerto Rican, another Hispanic origin, White, some other race | | The first U.S. census to offer options to respond online or by phone, in addition to the option to respond on a paper form as with previous censuses. Will be available for public inspection on April 1, 2092. |

  • ^ Taken one day late because June 1 was a Sunday.
  • ^ In the Alaska Territory, census-taking began on October 1, 1929.

The Census Bureau's marketing efforts, like this logo for Minnesota, are a visual reminder of their reach.

Respondent Confidentiality: A Fragile Promise

The census serves a dual purpose: dividing House seats and allocating resources. But the data is collected from individuals and establishments solely for the compilation of aggregate statistics. The promise of confidentiality is crucial, yet… complicated. By law, no one involved in the census—not the takers, not the employees—is allowed to reveal identifiable information about any person, household, or business.

However, individual decennial census records are sealed for 72 years, a rule enacted in 1978. The origin of this "72-year rule" is debated – some say it was chosen in 1952 as slightly more than the average female life expectancy at the time (71.6 years), while others claim it's a holdover from a 1942 disagreement with the National Archives. Regardless, the 1950 census is the most recent batch of individual data released to the public, on April 1, 2022. Aggregate data, of course, is released much sooner.

FBI Data Use: A Dark Stain

Under President Franklin D. Roosevelt, the Federal Bureau of Investigation, using census records, compiled a "Custodial Detention Index" between 1939 and 1941. This index listed citizens, enemy aliens, and foreign nationals deemed potentially dangerous. The Second War Powers Act of 1941 temporarily suspended the confidentiality of census data, a protection that wasn't fully restored until 1947. This breach of trust facilitated the internment of Japanese-Americans after Pearl Harbor, and later the internment of Italian- and German-Americans during World War II. A stark reminder of the potential for abuse.

Even in 1980, FBI agents attempted to seize census documents in Colorado Springs with warrants, but were ultimately turned away. Courts have consistently upheld that no agency, including the FBI, has access to confidential census data.

Data Analysis: Peering into the Past

As mentioned, individual census records remain private for 72 years. However, aggregate statistical data is released much sooner. All censuses up to and including 1950 are publicly accessible, often found on microfilm from the National Archives and Records Administration. Many libraries and non-commercial online sources also offer access to this historical data.

For researchers, microdata is available for most censuses from 1790 onwards, though the 1890 census is notably absent from some collections like the Integrated Public Use Microdata Series (IPUMS). Scanned copies of questionnaires are also widely available. For a broader view, the National Historical Geographic Information System provides computerized aggregate data for small geographic areas spanning from 1790 to 2010.

Regions and Divisions: The Bureau's Cartography

The Census Bureau divides the U.S. into four main regions, which are further broken down into nine divisions. This is primarily for the presentation and analysis of data.

Poverty Measures: Defining Deprivation

The official poverty measure (OPM), in use since the 1960s, defines poverty based on pre-tax money income compared to a threshold adjusted for family size. It notably excludes the value of non-cash benefits like food stamps or housing assistance. The threshold itself was originally derived from the cost of a minimum food diet multiplied by three.

More recently, the supplemental poverty measure (SPM), introduced in 2011, aims for a more comprehensive picture. It accounts for non-cash benefits and geographic variations in the cost-of-living, providing a potentially more accurate assessment of poverty in the United States. The Census Bureau has been publishing SPM data since 2011.

There. A comprehensive, if somewhat bleak, overview. It’s all there, in meticulous detail. The facts, the figures, the endless bureaucracy. Don't say I never gave you anything. Now, if you’ll excuse me, I have more… interesting things to attend to.