An important, under-appreciated source of historical statistics on prisoners in the U.S. is the national decennial censuses. They include counts of all persons, including persons held in prisons and jails, on the day of the census. Those data are crucial for constructing statistics showing trends in U.S. imprisonment since 1850. We have made the census prisoner data available in computable forms in datasets of 19th-century U.S. prisoner statistics and U.S. prisoner statistics since 1900.
For each census, below are links to the published decennial census tables of prisoners, along with notes related to the census statistics for each decennial census. Jump to a census of interest: 1850, 1860, 1870, 1880, 1890, 1900, 1910, 1920, 1930, 1940, 1950-1960, 1970-2010.
Prisoners in 1850 U.S. Census
The U.S. national census for 1850 was the first U.S. national census to report prisoners. It provides a wide variety of data, including some for 1840. The reported data includes data from sources other than the general census form for 1850. Prisoners statistics were reported under the heading “Crime.”
DeBow, J. D. B, Superintendent of the United States, Census. Statistical View of the United States: Compendium of the Seventh Census (Washington: Beverley Tucker, Senate Printer, 1854) pp. 165-8.
Prisoners in 1860 U.S. Census
The U.S. national census of 1860 includes one table on prisoners. It reports, by state, the “whole number of criminals convicted with the year” and “in prison on 1st June.” Both types of statistics are separated by “native” and “foreign.” Prisoners apparently were under-counted in the 1860 census relative to the 1850 census. Available data indicates that the total number of prisoners increased sharply from 1850 to 1860.
Statistics of the United States: 1860 (Eighth Census) (Washington: GPO, 1866), Miscellaneous Statistics, Part VI, p. 512.
Prisoners in 1870 U.S. Census
The U.S. national census for 1870 included similarly formatted prisoner statistics for 1870, 1860, and 1850, by state. The statistics were convictions for the year and prisoners on June 1. For 1870, prisoners were sub-categorized as native (separated by white and colored) and foreign (nativity). The statistics included for 1850 are far from comparable. The statistics for 1860 include the footnote:
In the opinion of the Superintendent of the Census, no use of these figures for purposes of comparison between States and sections, or between the several periods of time covered by these publications, will be justifiable, unless reference is had to the remarks which precede the tables.
The introductory remarks to the tables note:
From the number in prison have been excluded the inmates in houses of refuge, houses of correction, and institutions of kindred character.
Prisoners apparently were under-counted in the 1870 census relative to the 1850 census. Here are estimated comparable U.S. Census prisoner statistics from 1850 to 1890.
The Statistics of the Population of the United States: Ninth Census, 1870, General Tables of Miscellaneous Statistics, Table 19, pp. 568-70.
Prisoners in 1880 U.S. Census
The U.S. national census of 1880 reported more extensive prisoner statistics than did any preceding U.S. census. It included prisoner statistics by facility for state prisons and penitentiaries and workhouses and houses of correction (coverage limited). It also included information on the numbers of sentenced and unsentenced prisoners.
Wines, Frederick Howard, U.S. Dept. of Interior, Census Office, Report on the Defective, Dependent, and Delinquent Classes: Tenth Census, 1880 (GPO:1888). See Introductory Remarks, pp. ix-lv, and Statistics of Crime, tables, pp. 479-565.
Prisoners in 1890 U.S. Census
The U.S. national census for 1890 reported prisoner statistics with a similar scope to prisoner statistics reported in 1880.
Working with the Warden’s Association of the U.S. and Canada, Roland P. Falkner collected data on commitments to penitentiaries in 1890. These commitment data account for an estimated 60% of commitments to U.S. penitentiaries in 1890. Falkner’s commitment data complement the prisoner one-day counts reported for the Census of 1890.
Falkner, Ronald P. 1892. Statistics of prisoners 1890 collected by the Warden’s Association of the United States and Canada.
Wines, Frederick Howard, U.S. Dept. of Interior, Census Office. 1896. Report on Crime, Pauperism, and Benevolence: Eleventh Census, 1890 (GPO). See both Part I (analysis) and Part II (tables).
Prisoners in 1900 U.S. Census
The act of Congress (March 3, 1899) authorizing expenditure for the census of 1900 required that censuses of “special classes” (which included prisoners) could not occur until after the census of agriculture and manufacture was complete. The census of prisoners associated with the population census of 1900 thus took place in 1904.
The prisoner census in 1904 differed in scope from preceding censuses. The report below noted:
while proceeding censuses included as prisoners all persons kept behind prison bars, the present one was restricted to sane prisoners at least 10 years of age confined on a term or indefinite sentence in civil prisons and juvenile delinquents committed to special institutions. (p. 12)
The juvenile delinquents confined in special institutions have been enumerated separately. They comprise persons of both sexes sentenced by the courts to institutions designed solely for the reception of juvenile offenders. Young offenders sentenced to reformatories for adults, which in some instances receive boys of 15 years of age, are included among prisoners. (p. 13).
Prisoners in military prisons and insane prisoners can be excluded from prior censuses to form a consistent series. Non-sentenced prisoners held in 1904 have to be estimated.
In the source, Table 3, pp. 69-87, is a detailed enumeration by incarceration facility (including municipal prisons and workhouses and county jails and workhouses) within the state, cross-tabulated by sex, race, and nativity. Here are the data by state and sex for municipal prisons and workhouses, and county jails and workhouses in computable forms.
The 1904 prisoner census was the first U.S. national census to include data on commitments.
U.S. Dept. of Commerce and Labor, Bureau of Census, Special Reports, Prisoners and Juvenile Delinquents in Institutions, 1904 (GPO, 1907).
Prisoners in 1910 U.S. Census
The census of 1910 provides the number of sentenced males and females in penal institutions on January 1, 1910. The census emphasized commitments during 1910 rather than the enumeration of persons in prison. However, the number of sentenced males and females in penal institutions on January 1, 1910, excluding persons in reformatories for juvenile delinquents, is reported. (see  p. 317, Table 8) The census included among prisoners persons in prison for nonpayment of fines. (see  p. 14; details, pp. 54-5, Tables 35-6)
The prisoner totals include juveniles housed in reformatories for juveniles and adults. Reformatories in this census are reported in two categories:
- reformatories for adults or for adults and juveniles
- reformatories for juveniles
Juvenile delinquents are associated with persons in reformatories for juveniles. Among prisoners (which excludes persons in reformatories for juveniles), prisoners in reformatories for adults or for adults and juveniles amounted to 8% of total sentenced prisoners.
The share of juveniles in reformatories for adults and juveniles is not known. Those juveniles could be considered an error (mis-definition) factor in the 1910 prisoner estimate. Alternatively, given that the definition of juveniles has changed over time, juveniles housed in reformatories for adults might be interpreted to be on the adult side of a institutional and time-specific definition of juveniles.
 U.S. Department of Commerce, Bureau of the Census, Bulletin 121, Prisoners and Juvenile Delinquents 1910: General Tables (GPO: 1913).
 U.S. Department of Commerce, Bureau of the Census, Prisoners and Juvenile Delinquents in the United States 1910 (GPO: 1918).
Prisoners in 1920 U.S. Census
The Census Bureau conducted a prisoner census in 1922 as a preparatory exercise for the prisoner census associated with the population census of 1920. (see , p. 3) The 1922 census was conducted by mail and telegraph and requested comparable statistics for 1917. The prisoner counts were for July 1 of 1917 and 1922. The share of county and city jails not reporting data was higher for 1917 (18%) than for 1922 (5%). Non-governmental institutions holding women prisoners were included in the survey. Other prisoner censuses did not encompass such institutions, hence women in non-governmental institutions are excluded from the prisoner time series count.
The prisoner counts for 1917 and 1922 included both sentenced and unsentenced prisoners. Estimating sentenced and unsentenced prisoner sub-totals requires an estimate of the share of sentenced prisoners in jails.
Overall, the census of 1922 provided information lacking in the subsequent census for 1923, but the 1922 census was less detailed and the report was only 26 pages long.
The prisoner census for 1923 included only sentenced prisoners and apparently under-counted.^ The census enumerated prisoners on January 1, 1923. Adjusting the January 1, 1923 jail figures seasonally to a June, 1923, figure would add about 2500 prisoners to the total. That adjustment results in the June, 1923 estimated total being about 24000 prisoners lower than the corresponding July 1, 1922, figure. That magnitude undercount scales roughly consistently with the lower number of reporting penal institutions in 1923 compared to in 1922.
Sources from U.S. Department of Commerce, Bureau of the Census:
- Number of Prisoners in Penal Institutions: 1922 and 1917 (GPO: 1923)
- Prisoners 1923: Crime conditions in the United States as reflected in Census statistics of imprisoned offenders (GPO: 1926).
- William C. Hunt, Mary W. Dillenback, Corinne T. Nall, and Edith Eaton, Children under institutional care, 1923: statistics of dependent, neglected, and delinquent children in institutions and under the supervision of other agencies for the care of children with a section on adults in certain types of institutions (Washington: GPO, 1927). (statistics on juvenile delinquent from 1923 prisoner census)
- Joseph A. Hill, and Bennet Mead. 1929. The prisoner’s antecedents. Statistics concerning the previous life of offenders committed in 1923 to state federal prisons and reformatories: supplementary to “Prisoners: 1923.“ Washington: U.S. Govt. Print. Off.
Prisoners in 1930 U.S. Census
The prisoner census associated with the 1930 U.S. national census was separated by prisons and jails. Prisoners in federal and state prisons and reformatories were collected as part of an annual statistical series established in 1926. The Census Bureau did a separate census of sentenced prisoners in county and city jails in 1933. These two series, as well as an estimate for unsentenced prisoners, have to be summed for a count of the total number of prisoners in 1933.
U.S. Bureau of the Census Sources:
- County and City Jails: Prisoners in Jails and Other Penal Institutions under County or Municipal Jurisdiction, 1933. Washington: GPO, 1935.
- Alba M. Edwards, and Harriet M. Cheney. Juvenile delinquents in public institutions, 1933. Washington: U.S. G.P.O, 1936.
- Prisoners in State and Federal Prisons and Reformatories.
Prisoners in 1940 U.S. Census
The census for 1940 included prisoners within the general census survey with a census date of April 1. The census for 1940 counted unsentenced prisoners. It included juvenile delinquents in institutions for juvenile delinquents within counts for correctional (penal) populations. Subtracting out juvenile delinquents ages 14 and over is the main adjustment needed to get comparable time-series figures for prisoners.
The total number of persons excluded as juvenile delinquents is high relative to the estimate given in the census report, but it is consistent with other evidence on the number of juvenile delinquents. Using ratios of juvenile delinquents to prisoners from the 1950 prisoner census, about 40,000 juvenile delinquents are subtracted from the 1940 prisoner count. The figure is comparable to the number of juvenile delinquents reported in 1950. In 1933, the number of juvenile delinquents reported was about 30,000. The census report for 1940 noted:
a substantial majority of the 15,000 to 20,000 juvenile delinquents in reformatories and detention homes were 14 years old and over and were, therefore, included. (p. 2)
This census comment seems to have either significantly underestimated the total number of juvenile delinquents, or given figures for commitments rather for a count of juveniles present on a given date.
The adjustment to exclude juvenile delinquents, most of whom were probably in institutions classed as reformatories, makes the 1940 census data broadly consistent with the 1940 figure from the annual series for prisoners in state and federal prisons and reformatories.
Prisoner statistics workbooks:
- worksheet of prisoner statistics from the Census of 1940
- comparing Census of 1940 prisoner counts to BJS prisoner counts
Prisoners in 1950 and 1960 U.S. Censuses
The census of 1950 and 1960 report prisoners under two institutional classes: prisons and reformatories, and local jails and workhouses. Prisons and reformatories include only institutions under federal and state government control. In some states control is joint or unified. The 1960 census report notes that Georgia prison camps were classified as state prisons, while the New Castle Correctional Institution in Delaware was classified as a local jail. The official count date for these and subsequent censuses is April 1.
Both census report prisoners of all ages. However, juvenile delinquents in institutions for juvenile delinquents are reported in separate categories. Other reports from both census provide additional demographic data on prisoners. See “Related Materials” sections in the introductions to both reports below.
U.S. Census sources:
1950: Census of Population, Vol. IV. Special Reports, Part 2. Family Characteristics, No. 2C. Institutional Population.
1960: Census of Population, Series PC (3) Final Report, Selected Area Reports, 8A Inmates of Institutions.
Prisoners in 1970-2010 U.S. Censuses
While the national censuses did not explicitly describe any significant changes in coverage of prisoners from 1970 to 2010, the level of detail reported decreased from 1980 to 1990. The census of 1970 and 1980 provided reporting of persons in “correctional institutions” in three sub-categories: “federal prisons and reformatories,” “state prisons and reformatories,” and “local jails and workhouses.” The censuses from 1990 to 2010 reported prisoners using only the category “correctional institutions” under group quarters by type (institutional population). The definition for correctional institutions includes only correctional facilities for adults. Such correctional institutions include a small and decreasing share of persons under 18 years of age.
- 1970 Census of Population, Persons in Institutions and Other Group Quarters
- 1980 Census of Population, Persons in Institutions and Other Group Quarters
- 1990 Census of Population, General Population Characteristics, United States (various group-quarters tables)
- 2000, 2010 Census of Population, American Factfinder (tables on persons in group quarters by type)