Scientific census-taking in the United States began with the decennial census of 1850, when the scope and methods were greatly improved by making the individual the unit of study. In 1902 the Bureau of the Census was established in the Dept. of the Interior; the following year it was transferred to the Dept. of Commerce and Labor and remained in the Dept. of Commerce when the Dept. of Labor was separated (1913). In addition to being a vital source of statistical data about the nation, information from the U.S. census is also used to allocate federal resources.
The government was criticized and also sued for undercounting the homeless and minorities in the 1990 census. In 1996 the Supreme Court ruled that the decision to adjust the count is left to the discretion of the secretary of commerce. The government proposed remedying the problem of undercounting through the use of statistical adjustments to the 2000 census, but the Supreme Court ruled (1998) against the plan, and the traditional head-count method prevailed. In 2001 the government again decided to use unadjusted census figures. About 3.3 million people, largely minorities, were estimated to have been missed by the 2000 census; a smaller number were thought to have been counted twice. Unadjusted census figures are generally believed to favor Republicans in the drawing of districts for the House of Representatives.
See W. S. Holt, The Bureau of the Census (1929, repr. 1974); F. Yates, Sampling Methods for Censuses and Surveys (4th ed. 1980); M. J. Anderson, The American Census (1990); S. Roberts, Who We Are: A Portrait of America Based on the 1990 Census (1994).
The census can be contrasted with sampling in which information is only obtained from a subset of a population. As such it is a method used for accumulating statistical data, and also plays a part in democracy (voting). Census data is also commonly used for research, business marketing, planning purposes and not at least as a base for sampling surveys.
It is widely recognized that population and housing censuses are vital for the planning of any society. Traditional censuses are, however, becoming more costly. A rule of thumb for census costs in developing countries have for a long time been $1 USD per enumerated person. More realistic figures today are around $3 USD. These approximations should be taken with great care since a variable amount of activities can be included in different countries (e.g. enumerators can either be hired or requested from civil servants). The cost in developed countries is far higher. The cost for the 2000 census in the U.S. is estimated to $4.5 billion USD, more than $15 per enumerated person. Alternative possibilities for retrieving data are being investigated. Nordic countries Denmark, Finland and Norway have for several years used administrative registers. Partial censuses ‘Micro censuses’ or ‘Sample censuses' are practiced in France and Germany.
It is typical for census data to be processed in a way so as to obscure individual information. Some censuses do this by intentionally introducing small statistical errors to prevent the identification of individuals in marginal populations; others swap variables for similar respondents. Whatever measures have been taken to reduce the privacy risk in census data, new technology in the form of better electronic analysis of data poses increasing challenges to the protection of sensitive individual information.
Census in Egypt are said to have been taken already during the early Pharaonic period. 3340 BC and in 3050 BC.
One of the earliest documented censuses was taken in 500-499 BC by the Persian Empire's military for issuing land grants, and taxation purposes.
In India, censuses were conducted in the Mauryan Empire as described in Chanakya's (c. 350-283 BC) Arthashastra, which prescribed the collection of population statistics as a measure of state policy for the purpose of taxation. It contains a detailed description of methods of conducting population, economic and agricultural censuses.
Rome conducted censuses to determine taxes (see Censor). The word 'census' origins in fact from ancient Rome, coming from the Latin word 'censere', meaning ‘estimate’. The Roman census was the most developed of any recorded in the ancient world and it played a crucial role in the administration of the Roman Empire. The Roman census was carried out every five years. It provided a register of citizens and their property from which their duties and privileges could be listed.
The world's oldest extant census data comes from China during the Han Dynasty. Taken in the fall of 2 AD, it is considered by scholars to be quite accurate. At that time there were 59.6 million living in Han China, the world's largest population. The second oldest preserved census is also from the Han, dating back to 140 AD, when only a few more than 48 million people were recorded. Mass migrations into what is today southern China are believed to be behind this massive demographic decline.
In the Middle Ages, the most famous census in Europe is the Domesday Book, undertaken in 1086 by William I of England so that he could properly tax the land he had recently conquered. In 1183, a census was taken of the crusader Kingdom of Jerusalem, to ascertain the number of men and amount of money that could possibly be raised against an invasion by Saladin, sultan of Egypt and Syria.
A very interesting way to record census information was made in the Inca Empire in the Andean region from the 15th century until the Spaniards conquered their land. The Incas did not have any written language but recorded information collected during censuses and other numeric information as well as non-numeric data on quipus, strings from llama or alpaca hair or cotton cords with numeric and other values encoded by knots in a base 10 positional system.
More about census, see: National Institute of Statistics and Census of Argentina
Since these first accounts Bulgarian authorities had organized several population censuses: 1892, 1900, 1905, 1910, 1920, 1926, 1934, 1946, 1956, 1965, 1975, 1985, 1992 and 2001.
The data provided in the Bulgarian censuses from 1888 until WW-II is regarded as highly reliable according to the standards of the time. The Bulgarian leading statisticians of the period were generally educated in Western universities and participated vividly in the international cooperation, therefore insisted and succeeded in introducing the best practices of the time. The quality of the data provided of later censuses is a matter of debate.
Censuses in Canada are conducted in five-year intervals. The last two censuses were conducted in 2001 and 2006. Censuses taken in mid-decade (1976, 1986, 1996, etc.) are referred to as quinquennial censuses. Others are referred to as decennial censuses. The first quinquennial census was conducted in 1956.
For the 2006 Census of Canada, respondents were able, for the first time, to choose to complete their census questionnaire online. Other options for answering the questionnaire include postal mail (using a pre-paid envelope) and telephone (using an 800 number).
In the Province of Alberta, Section 57 of its Municipal Government Act (MGA) enables municipalities to perform their own censuses on any given year. An official municipal census must be conducted no earlier than April 1 and no later than June 30 of the same year, according to the MGA's Determination of Population Regulation If municipalities choose to make their census count official, the new population must be submitted to the Ministry of Municipal Affairs and Housing prior to September 1 of the year the census was performed. The latest census counts for Alberta's municipalities are released in the Ministry's annual Official Population List publication.
AltaPop (Alberta Population) is a very useful website that builds upon the data provided by the Province and Statistics Canada. Visit AltaPop to compare municipal and federal census results by municipality, to analyse historic population trends by municipality, and to view detailed annual population summaries either by size of municipality or sorted alphabetically.
Following these early undertakings, the first census to attempt completely covering all citizens (including women and children who had previously been listed only as numbers) of Denmark-Norway was taken in 1769. At that point there were 797 584 citizens in the kingdom. Georg Christian Oeder took a statistical census in 1771 which covered Copenhagen, Sjælland, Møn, and Bornholm.
After that, censuses followed somewhat regularly in 1787, 1801, and 1834, and between 1840 and 1860, the censuses were taken every five years, and then every ten years until 1890. Special censuses for Copenhagen were taken in 1885 and 1895.
In the 20th century, censuses were taken every five years from 1901 to 1921, and then every ten years from 1930. The last traditional census was taken in 1970.
A limited population census based on registers was taken in 1976. From 1981 and each year onwards information that corresponds to a population and housing census is retrieved from registers. Denmark was the first country in the world to conduct these censuses from administrative registers. The most important registers are the population register (Det Centrale Personregister), a Building and Dwelling Register and an Enterprise Register. The central statistical office, Statistics Denmark is responsible for compiling these data. This information is available online in the Statbank Denmark
Most of the census in 2007 was taken in August, while the Somali Region and the Afar Region were not covered. The northern Afar region is a remote, hot and arid area. The eastern Somali region (Ogaden) hosts a large nomadic Somali population and is a conflict area where Ethiopian regular forces are fighting against Ogaden National Liberation Front (ONLF).
The first large-scale census in the German Empire took place in 1895. Attempts at introducing a census in West Germany sparked strong popular resentment in the 1980s since many quite personal questions were asked. Some campaigned for a boycott. In the end the Constitutional Court stopped the census in 1980 and 1983. The last census was in 1987. Germany has since used population samples in combination with statistical methods, in place of a full census.
Relaciones Geográficas of Mexico and Guatemala, 1577-1585.
On May 25, 1577, King Philip II of Spain ordered by royal cédula the preparation of a general description of Spain's holdings in the Indies. Instructions and a questionnaire, issued in 1577 by the Office of the Cronista Mayor-Cosmógrafo, were distributed to local officials in the Viceroyalties of New Spain and Peru to direct the gathering of information. The questionnaire, composed of fifty items, was designed to elicit basic information about the nature of the land and the life of its peoples. The replies, known as "relaciones geográficas," were written between 1579 and 1585 and were returned to the Cronista Mayor-Cosmógrafo in Spain by the Council of the Indies.
The setting up, in 1952, of the National Registry (Þjóðskrá) eliminated the need for censuses. All those born in Iceland, and all new residents, are automatically registered. Individuals are identified in the registry by means of a national identification number (the so-called kennitala), a number composed of the date of birth in the format ddmmyy and four additional digits, the third of which is a control digit, and the last of which indicates the century in which the person was born (9 for the 1900s and 0 for the 2000s).
The National Registry doubles as an electoral register. Likewise, all bank accounts are linked to the national identification of the owner (companies and institutions all have their own identification numbers).
The first census in India in modern times is dated 1872. It started as far back as in 1860 and was finished in 1871. Starting from there, a population census has been carried out every 10 years, latest being the fourteenth in February-March 2001.
Census is carried out by the office of the Registrar General and Census Commissioner, India, Delhi an office under the Ministry of Home Affairs, Government of India. under the Census of India Act, 1948. The act gives Central Government many powers like to notify a date for Census, power to ask for the services of any citizen for census work. The law makes it compulsory for every citizen to answer the census questions truthfully. The Act provides penalties for giving false answers or not giving answers at all to the census questionnaire. One of the most important provisions of law is the guarantee for the maintenance of secrecy of the information collected at the census of each individual. The census records are not open to inspection and also not admissible in evidence.
Census happens in two phases, first House Listing and House Numbering Operations and second actual population enumeration phase. Census is carried out by the canvassing method. In this method, each and every household is visited and the information is collected by a specially trained enumerator.
The Statistical Center of Iran carries out nationwide population and housing censuses every 10 years, the last of which occurred in 2006 (1385 AP). In the Islamic Republic of Iran, based on Article 4 of the Act of the Statistical Center of Iran (SCI), the census shall be implemented once every 10 years according to the Presidential decree. So far there have been 6 incidences of population census in Iran in the years 1956, 1966, 1976, 1986, 1996, and 2006; all taken in accordance with scientific methods.
The most recent census took place on 23 April 2006. According to the 2006 form, "any person who fails or refuses to provide information or who knowingly provides false information may be subject to a fine of up to €25,000," under the Central Statistics Act 1993. On the CSO website, instructions for non-English speaking residents of the Republic of Ireland were available. They were mock copies of the census forms, with all headings/questions etc. being translated into a particular language. These were not to be filled out, but were only a guide on how to fill out the English or Irish form. This census also asked two new questions relating to ownership of PCs and Internet connection.
Data from the 1911 Census for the island of Ireland was made publicly available in 1961, and are being published online. Subsequent census records will be made publicly available 100 years after collection.
Questions relating to the ability to speak the Irish Language are included in the census. The figured obtained have been criticised as inflated by cognitive biases, such as response bias or wishful thinking. The 2006 census included an additional question on frequency of speaking Irish.
Kosovo claimed it's independence in early 2008. USA was among the first to admit new republic. It happened during Slovenian first EU presidency.
The census that was planned for 1981 was postponed and later cancelled. A call for privacy was responsible for the cancellation of any further census since 1991. Censuses are being conducted by the Centraal Bureau voor de Statistiek (Netherlands) since 1899. The censuses today are mostly (population, fiscal) register based, combined with surveying.
Censuses during the independence were taken 1963, 1973, 1991 and 2006. The results from 1973 were highly disputed. The preliminary results for 2006 indicates a population of 140,000,000. 700,000 enumerators were engaged in this operation.
Data being collected include population data (citizenship, place of residence, place of birth, position in household, number of children, religion, languages, education, profession, place of work, etc.), household data (number of individuals living in the household, etc.), accommodation data (surface area, amount of rent paid, etc.) and building data (geocoordinates, time of construction, number of floors, etc.). Participation is compulsory and reached 99.87% of the population in 2000.
Starting in 2010, the census will cease to be conducted through written questionnaires distributed nationwide. Instead, data in existing population registers will be used. That data will be supplemented with a biannual questionnaire sample of 200,000 people as well as regular microcensuses.
A census was taken in the Ottoman Empire 1831-38 by Sultan Mahmud II (1808-1839) as a part of the reform movement Tanzimat. Christian and Jewish men were counted but the female population was excluded.
Following the influence of Malthus and concerns stemming from his An Essay On The Principle Of Population the UK census as we know it today started in 1801. This was championed by John Rickman who managed the first four up to 1831, partly to ascertain the number of men able to fight in the Napoleonic wars. Rickman's 12 reasons - set out in 1798 and repeated in Parliamentary debates - for conducting a UK census included the following justifications:
The census has been conducted every ten years since 1801 and most recently in 2001. The first four censuses (1801-1831) were mainly headcounts and contained little personal information.
The 1841 Census, conducted by the General Register Office, was the first to record the names of everyone in a household or institution. However, their relationship to the head of the household was not noted, although sometimes this can be inferred from the occupation shown (eg servant). Those under the age of 15 had their proper ages listed, but for those who were older the ages were supposed to be rounded down to the nearest five years, although this rule was not strictly adhered to. Precise birthplaces were not given - at best the birthplace can be narrowed down to the county in which the person was living.
From 1851 onwards the census shows the exact age and relationship to the head of household for each individual; the place of birth was also listed, but with varying degrees of precision. Sometimes those who were born abroad have the annotation B.S. or British Subject.
The censuses are reasonably accurate. However, ages in particular are frequently shown incorrectly, though often the difference is only one year; in general the younger the individual the more accurate the age shown. Birthplaces often vary from one census to the next: a common error is to show the place where the census was taken as the birthplace, but most of the variations in birthplace can be accounted for by changes in geographical scale (for example, the nearest town being shown instead of the precise village, or a city being shown instead of the relevant suburb).
The censuses are also remarkably complete - though inevitably a small percentage of the population wasn’t recorded for one reason or another, and in some cases the records are missing or damaged (notably in 1861). Furthermore, all censuses of Ireland before 1901 have been lost or destroyed.
Because of World War II, there was no census in 1941. However, following the passage into law (on 5 September 1939) of the National Registration Act a population count was carried out on 29 September 1939, which was, in effect, a census.
The 2001 census was the first year in which the government asked about religion. Perhaps encouraged by a hoax chain letter that started in New Zealand, 390,000 people entered their religion as Jedi Knight (more than Sikhs, Buddhists or Jews), with some areas registering up to 2.6% of people as "Jedi". It was wrongly implied in emails that stating "Jedi" on the form would cause it to become an "official religion". No such thing exists in the United Kingdom. However, the director of reporting and analysis at the ONS stated that it may have helped with the collection process as it encouraged young people, who are often missed, to complete forms. (See Jedi census phenomenon.)
All of the British censuses from 1841-1901 have been transcribed and indexed and are available online; there is a joint project between the National Archives of Ireland and Library and Archives Canada to digitize the 1901 and 1911 censuses for the whole of Ireland, and it is possible this will be completed by the end of 2007. The next UK census is planned for 2011.
The first U.S. Census was conducted in 1790 by Federal marshals. Census takers went door to door and recorded the number of people in each household, along with the name of the head of the household. Slaves were enumerated, but for apportionment purposes each counted as only three-fifths of a citizen. American Indians, being neither taxed nor considered during apportionment, were not counted in the census. The first census counted 3.9 million people, less than half the population of New York City in 2000; the 2000 census counted over 281 million people. In 1902, Congress established the Census Bureau as a permanent Federal agency.
In recent times, there have been two forms of questionnaire – long and short. The long form and its additional questions about matters such as daily commute times, housing unit factors, etc. has been replaced by the American Community Survey (ACS). Computer algorithms (based on complex sampling rules) determined which form was mailed to a given household (in practice, of those households whose locations are on the Census Master Address List), with one in six receiving the long form. This was supplemented by census workers going door to door to talk to people who failed to return the forms. In addition to a simple count of residents, the Census Bureau collects a variety of statistics, on topics ranging from ethnicity to the presence of indoor plumbing. While some critics claim that census questions are an invasion of privacy, the data collected by every question is either required to enforce some federal law (such as the Voting Rights Act) or is required to administer some federal program. The United States Congress gives approval to every question asked on the Census.
Despite a massive effort, the Census Bureau has never been able to count every individual, leading to controversy about whether to use statistical methods to supplement the numbers for some purposes, as well as arguments over how to improve the actual head count. The Supreme Court has ruled that only an actual head count can be used to apportion Congressional seats; however, cities and minority representatives have complained that urban residents and minorities are undercounted. In several cases, the Census Bureau has recounted an area with disputed figures, provided the local government paid for the time and effort. The State of Utah protested the figures of the 2000 decennial census because it stood to gain a seat in the House of Representatives, but North Carolina gained it instead. Had the Census Bureau been mandated to count the numbers of Utahns living overseas, including many Mormon missionaries, Utah might have gained the seat.
To minimize the burden on individuals and to provide improved data, the Bureau has prepared several alternate methods for gathering economic, demographic, and social information, including the American Community Survey and record linking of depersonalized administrative records with other administrative records and Census Bureau surveys.
By law (92 Stat. 915, Public Law 95-416, enacted on October 5 1978), census records are sealed for 72 years. This figure has remained unchanged since before the 1978 law, reflecting an era when life expectancy was under 60 years, and thus attempts to protect individual's privacy by prohibiting the release of personal information during individuals' lifetimes. Thus, the most recent Census released to the public was the 1930 Census, released in 2002.
Indexes to some of the U.S. Censuses have been produced over the years, making the process of searching old census records much easier. Some indexes of census records have been produced by amateur volunteer genealogists. Due to the sheer volume of information, and the manual methodologies involved, the indexing is sometimes limited to the head of household. Some indexes have been published in bound volumes, which are often available in regional libraries, along with microfilm rolls that can be researched.
While valuable, indexes produced from these censuses can be problematic to use. The original census records from this era were completed by hand by census enumerators; this leads to problems in handwriting recognition and variations in spelling of surnames within the original documents.
The 1880 to 1920 censuses have indexes of last names, produced using the Soundex system; the indexing project was performed by the Works Progress Administration. The Soundex system is tolerant of variations in spelling; names with similar sounds but different spellings have the same encoding. The chief motivation in producing the Soundex name indexes was to assist citizens in finding census records to provide evidence of age, especially for those born before the advent of governmentally-approved birth certificates. (Verification of age was needed to establish eligibility for Social Security). Partial Soundex indexes of the 1930 census are available; resources from the Works Progress Administration were diverted towards support of World War II efforts before the project was completed.
With the advent of the Internet, census indexes are available on many genealogy websites. Several commercial services also provide access to images of the original census pages.