Census

B

Collecting the Information

Until relatively recently, population censuses were taken exclusively through personal interviews. The government sent enumerators (interviewers) to each household in the country. The enumerators asked the head of the household questions about each member of the household and entered the person’s responses on the census questionnaire. The enumerator then returned the responses to the government. Today, many censuses are conducted primarily through self-enumeration, which means that people complete their own census questionnaire. Self-enumeration reduces the cost of a census to the government because fewer enumerators are needed to conduct interviews. In addition, the procedure provides greater privacy to the public and generally improves the accuracy of responses, because household members can take more time to think over the questions and consult their personal records.

A country conducting a census chooses a collection technique based on its social and political traditions and technological capacities. The United States census is highly automated and has been conducted primarily by mail since 1970. For the 2000 U.S. census, the Census Bureau offered many people the option of answering their questionnaires through the bureau’s Web site. Canada began to use self-enumeration in 1971. Today the Canadian government sends enumerators to deliver the census form to each household; the household head fills it out and sends it back to the government. In both the United States and Canada, enumerators are sent to follow up on households that do not mail back the census questionnaire. Other nations continue to conduct censuses only through direct enumeration. Some, such as Turkey, require people to stay home on Census Day to await the census taker.

Census agencies make a special effort to count people who may not receive a questionnaire by mail or who have no permanent address. For example, the U.S. Census Bureau sends census takers to interview people at homeless shelters, soup kitchens, mobile food vans, campgrounds, fairs, and carnivals. It consults with experts to find migrant and seasonal farmworkers. The bureau works with the Department of Defense and U.S. Coast Guard to identify people living on military installations or ships. The Census Bureau also counts military personnel and federal civilian government employees and their families who are living overseas. Finally, the agency distributes census questionnaires to people living in group quarters, such as college dormitories, nursing homes, hospitals, prisons and jails, halfway houses, youth hostels, convents and monasteries, and women’s shelters.

In the United States, Canada, and other countries, households receive either a short or long census questionnaire. Most households receive the “short form,” a brief set of questions on basic characteristics such as name, age, sex, racial or ethnic background, marital status, and relationship to the household head. But a small sample of households receives the “long form,” which asks many other detailed questions. These may include questions about the individual’s educational background, income, occupation, language knowledge, veteran status, and disability status as well as housing-related questions about the value of the individual’s home, the number of rooms and bedrooms in it, and the year the structure was built. The statistical technique of sampling—asking questions of only a representative sample of the population—allows census agencies to collect this detailed information without placing an undue burden on the population or creating an excessive cost to the government. About one in six households in the United States and one in five households in Canada receives the long form. These sample sizes are large enough to produce reliable information about the population characteristics of neighborhoods, regions, states or provinces, and the country as a whole.

---

---

C

Processing and Analysis of Data

For most of the 19th century in the United States and Canada, census data were tabulated and compiled by hand, without the aid of machines. Manual processing was very slow, and some figures were obsolete by the time they were published. The invention of mechanical tabulating devices in the late 19th century made processing of the data much faster and improved the accuracy of the results. Today, census questionnaires are processed primarily on computers and electronic equipment. Besides speeding the processing of results, computers have made it possible to perform sophisticated analyses on the data and to draw correlations between various social and economic characteristics of the country. For example, using census data, statisticians can easily determine the number of people living in Houston, Texas. But they can also determine the number of Houston women between the ages of 25 and 30 who have completed high school and are currently employed.

To process the data from hundreds of millions of paper questionnaires, the U.S. Census Bureau employs an advanced system that scans every questionnaire into an electronic image. Then the images are analyzed by computer software that can recognize when a check-box item on the questionnaire has been marked with a pencil or pen. Optical character recognition software analyzes handwritten responses on the questionnaire and translates them into electronic data. Once in electronic form, the data can be analyzed and turned into statistics. Unreadable or ambiguous responses are checked by census clerks and manually keyed into the computer.

D

Publication of Results

U.S. and Canadian censuses publish only general statistical information and keep individual responses confidential. By law, the U.S. Census Bureau and Statistics Canada are prohibited from releasing individual responses to any other government agency or to any individual or business. Census workers in both countries must swear under oath that they will keep individual responses confidential. Employees who violate this policy face a monetary fine and possible prison term. If an individual’s personal data were not kept confidential, people might refuse to participate in the census for fear that their personal information would be made public or used by the government to track their activities. In the United States, individual census responses are stored at the National Archives. After 72 years, the original forms are declassified and opened to the public. These original responses are frequently used by people researching the history of their families or constructing genealogies. In Canada, census responses from 1906 and later are stored at Statistics Canada. Microfilmed records of census responses from 1901 and earlier are stored at the National Archives of Canada; these are the only individual census responses currently available for public use.

Until the 1980s, census agencies published their results in large volumes of numeric tables—sometimes numbering in the hundreds of volumes. Today, the majority of census data is distributed electronically, either through the Internet or on CD-ROM, diskette, or magnetic tape. The Web sites of the U.S. Census Bureau and Statistics Canada provide online access to hundreds of statistical publications and data sets. The U.S. Census Bureau planned to disseminate results from the 2000 population census primarily via its Internet site. Both the U.S. Census Bureau and Statistics Canada continue to distribute printed publications for the most commonly requested demographic information. The Statistical Abstract of the United States, published annually by the Census Bureau, is an important statistical compendium on the social, political, and economic aspects of life in the United States. This publication includes data from decennial censuses as well as from other sources, such as surveys taken between censuses. Statistics Canada publishes a similar annual volume on Canadian statistics called the Canadian Year Book.

IV

United States Censuses

A census of U.S. population has been conducted every ten years since 1790, as mandated by Article I, Section 2 of the Constitution of the United States. The U.S. Census Bureau, established in 1902, conducts the census. Before 1902 a separate office was set up for each census and then disbanded when the census was completed.

A

Origins

During the American Revolution (1775-1783), the leaders of the independence movement faced many problems uniting the 13 separate American colonies under a national government. Among these problems were how to allocate political representation among the states and how to levy taxes. The initial government structure under the Articles of Confederation, adopted in 1781, gave each state one vote in Congress. However, this system proved unsatisfactory to states with larger populations, who felt they deserved more representation than smaller states. A new Constitution, adopted in 1789, created a two-house legislature consisting of a Senate and a House of Representatives. Each state received two seats in the Senate regardless of population, but representation in the House of Representatives was based upon the population of each state. The decennial census was designed to provide the population figures for apportioning the seats in the House. Taxes levied on the states were also to be apportioned on the basis of population.

In the late 18th century, when a racially based slave labor system existed in the United States, almost 20 percent of the American population was enslaved African Americans. The framers of the Constitution debated whether slaves were “persons” or “property” and, thus, whether states should receive representation for their slave populations. The southern states, where slavery dominated, did not consider slaves as people for purposes of apportioning their state legislatures, but they did consider slaves as property for tax purposes. The framers could not find an easy solution to this dilemma and developed what came to be called the Three-Fifths Compromise. This clause in the Constitution required the census to count each slave as three-fifths of a person when determining the apportionment of the House. The Three-Fifths Compromise thus required the census to count the slave population and the free, mainly white, population separately. The Constitution also specified that “Indians not taxed,” that is, those American Indians who were not considered part of civil society, were not to be counted in the census.

	More from Encarta
	•  Presidential Myths Quiz •  Coffee break: Recharge your brain