Examines the evolution of US census schedules from 1790 to the present. From 1790 to 1840, the census schedule changed very little and was difficult to use. The format was adjusted in 1850 so that more detailed information could be gathered, but the decision came only after a heated congressional debate as the new schedule became embroiled in the sectional controversy. The 19th century also saw social scientists and reformers begin petitioning the federal government to further revise the schedule so that specific types of information could be gathered. The Census Bureau created the Joint Census Advisory Committee in 1920 to handle the multitude of suggestions generated by interest groups. Also, beginning in 1930, the Census Bureau sponsored conferences so that any persons interested in revising the format of the schedules could voice their opinions. While interest groups in the 20th century have helped to determine how the census is formulated, Congress has had much influence as well. Congressional concerns, however, have usually reflected immediate political concerns.

Examines the procedures that the US Census Bureau has used to collect census data since 1850. From 1850 to 1870, federal marshals were charged with appointing assistants to collect census data. Because the primary duties of marshals concerned law enforcement, census taking became burdensome, and the assistants they chose were usually ill-trained political appointees. This changed in 1880 when the Department of Interior appointed administrators of the census who could be more closely supervised. Census administrators then selected their own enumerators. This system served as a model for later censuses, and refinements were introduced in order to improve the quality of the enumerators. Such reforms included written exams, the use of street books to insure more accurate counts, and training seminars. The 1960 and 1970 censuses ushered in the age of self-enumeration whereby census questionnaires were mailed to each household, and the data was then processed with electronic equipment. These reforms dramatically reduced census undercounts.

The purpose of this dissertation is twofold: first, to analyze who and what influenced the content of the decennial census of population schedules from 1790 to 1940; and second, to examine census-taking procedures across the same time period. The audience of this work will also likely have a serious interest in the enumeration, correction, and tabulation of the United States census, albeit for slightly different reasons; to gain a sense of the relative accuracy of census data in relation to public use samples. The Making of a Modern Census: The United States Census of Population, 1790-1940, will fill two gaps in the literature of United States census history. These gaps are relevant not only to those with a casual or incidental interest in the United States census of population, but also to those users of aggregate and public use sample census data. Users of census data, no matter what their discipline, will be interested in who and what influenced the census schedule as well as the procedures in place to collect the data. The first gap pertains to a treatment of who and what influenced population schedule content and the second to the lack of procedural histories for U.S. population censuses prior to 1950. The Making of a Modem Census is divided into two parts. Part I analyzes which individuals and groups most affected the population schedule from 1790 through 1940 and, in relation to these actors, discusses some notable changes in the scope and content of the population schedule. The ambitious goal of Part II is to provide an integrated source of census procedure from 1790-1940. In order to delineate the process of enumeration for the first sixteen decennial censuses of population, four issues of procedural comparability are considered: the selection and quality of local administrators and canvassers; training and oversight provided by the national census office; public preparedness for enumeration; and provisions for correcting final returns. The final section, 'trials and tribulations of canvassers,' provides an enumerator-level perspective of the procedures in 'action.'

There are three sources of information about undercounts in nineteenthcentury U.S. censuses: demographic analyses of net undercounts by age, sex, and race at the national level; record-linkage studies of gross undercounts for local communities; and contemporary testimony of the types and bases of underenumeration. This article reviews the strengths and limitations of each of these sources, assesses the extent of their agreement, and discusses the bases of their disagreement. Our main conclusions, in brief, are as follows. Each of the three sources offers valuable, albeit different, insights into historical undercounts. Demographic analyses are the best guide to the overall level of the undercount; community-centered linkage studies provide additional information (beyond age, sex, and race) about the persons the enumerators missed; and contemporary testimony both confirms the general conclusions from linkage studies and suggests why some social groups were disproportionatelyundercounted.

Discusses the development of computer software for the Public Use Microdata Samples (PUMS), which are random samples taken from the US censuses from 1850 to 1920. Because each census was done differently, the Integrated Public Use Microdata Series (IPUMS) was developed so that various PUMS could more easily be compared. Once the IPUMS was finished, the University of Minnesota's Social History Research Laboratory developed the computer software needed to utilize the PUMS data sets. This was a two-step process. First, data-entry software packages were created to upload the information into microcomputers. Once this was done, a larger UNIX workstation was used to do the post-entry processing so that the data could be checked, verified, and recoded into numeric form.

Describes the process involved in creating the occupational codes for the 1850, 1880, and 1920 Public Use Microdata Samples (PUMS). PUMS are random samples of households taken from the US decennial censuses and coded into computer format for use by historians and social scientists. Much of the work for the 1850, 1880, and 1920 PUMS has been done by the Social History Research Laboratory at the University of Minnesota. The process for encoding occupational data was especially difficult because each census recorded such information differently. The laboratory solved this problem by using a common classification scheme and creating a data dictionary to translate the manuscript census information into codes. Additionally, they used occupational groupings of the 1950 census to create a supplementary coding scheme that provided more detailed information. In many instances the PUMS clearly demonstrate that the Census Office's shoddy procedures often led to dramatic undercounts of many occupational categories, particularly female domestic servants.

A quiet demographic revolution has occurred during the twentieth century in the United States: the decline in intergenerational household sharing. Why were these living arrangements so common for older women early in the century? We examine the characteristics of adult kin who shared intergenerational households in 1910. Two nationally representative samples, of elderly mothers and their co-resident biological adult children were taken from the 1910 Census P.U.S. and linked to test general hypotheses relating to the determination of living arrangements. We find that kin availability influenced co-residence in two ways: by increasing the pool of children available and by facilitating strategic processes of kin selection based on quality of children. As kin availability increased, mothers chose security (especially the retention of headship) and a child's lack of competing obligations.

Public Use Microdata Samples (PUMS) are random samples taken from the decennial censuses since 1850. Because PUMS use the coding systems from the censuses from which they are taken, comparison between PUMS is often impossible. For this reason the Social History Research Laboratory of the University of Minnesota developed the Integrated Public Use Microdata Series (IPUMS), which uses compatibility codes that allow such comparisons to be made. To create IPUMS codes for occupation and income, the 1950 census and its occupational grouping scheme was adopted. PUMS taken from all other censuses were recoded according to this scheme, which now allows for comparison over time. A similar IPUMS coding system was done so that median income could be compared between various PUMS.

This essay reexamines the revisionist argument about the history of the family in light of new evidence about long-run changes in American family structure. In particular, I use the new Integrated Public Use Microdata Series, a national database incorporating consistent individual-level data from the U.S. Census over the period 1850 to 1990. I also report findings from the only eighteenth-century American census of sufficient size and quality to permit a consistent analysis of family composition, the 1776 census of Maryland. The evidence suggests that the revisionist interpretation needs revising. In fact, a form of extended family structure was dominant in nineteenth-century America and quite probably in the eighteenth century as well. The American preference for extended family structure disappeared in the twentieth century, and I...offer a brief analysis of some explanations for this change.

I use a new data source, the Integrated Public Use Microdata Series, to trace race differences in family structure between 1880 and 1980. Analysis confirms recent findings that the high incidence among African-Americans of single parenthood and children residing without their parents is not a recent phenomenon. From 1880 through 1960, black children were two to three times more likely to reside without one or both parents than were white children. In recent years, however, the race differential in parental absence has grown. Also, blacks have had a consistently higher percentage of extended households than have whites, but until 1940 this was the result of single parenthood and parentlessness among children: Extended households were more common among whites once the effects of absent parents were controlled

Explores wealth accumulation among European immigrants who arrived in the United States between 1840 and 1850. Uses a new sample of immigrants linked from passenger-ship records to the 1850 and 1860 federal census manuscripts. These immigrants rapidly accumulated real and personal wealth. Their real wealth grew 10% with each year's residence in the United States. This was not because immigrants arriving in the early 1840's were wealthier at arrival than later arrivals, nor was the rapid accumulation of wealth confined to one nationality or occupation. Rather, it reflects these immigrants' abilities to adapt to new circumstances after their arrival.