Enslaved Population Data Description
The data include all records transcribed from the census slave schedules in the respective years. The 1850 file includes 3,203,109 enslaved persons in 358,095 holdings; 1860 includes 3,936,602 enslaved persons in 400,898 holdings.
Each record in the data represents a single enslaved person. Person records are organized into holdings identified by the variable HOLDNUM. The combination of HOLDNUM and SLAVENUM constitutes a unique ID for each record in a census year.
The censuses collected only limited demographic information on enslaved persons. Names of enslaved persons were not recorded. The data are described in detail in this article, where their limitations and potential applications are discussed. See the variables page for a list of the available data fields.
Whenever possible, IPUMS links the records of enslaved persons to the holder in the full count censuses of the free population available through the IPUMS data extract system. In 1850, HISTID provides the linking key. In 1860, there can be as many as three holders associated with an enslaved record. The linking keys are HISTID, HISTID2, and HISTID3. The proportion of the enslaved population linked to a holder is 75.9% in 1850 and 78.9% in 1860.
Version 2: April 2026
The current version of the data was released in April 2026. This release affected the 1860 census file: adding data on disabilities and improving the variables on fugitive and manumitted status of enslaved persons.
Adding disability data was relatively straightforward, but entailed visual inspection of all census forms. Sometimes it was difficult to discern between stray and intentional marks on the census forms.
The fugitive and manumission fields required more interpretation, necessitating a comprehensive comparison of the data provided by Ancestry to the census form images. The census forms sometimes contained extraneous information in these fields, where office workers appeared to use the columns for other purposes, such as summing up characteristics from other items on the forms. Additionally, the census instructions stated that these fields should report the sum of fugitive or manumitted persons from the holding, but some enumerators appeared to record these data at the level of individual enslaved persons. In the earlier release of the full count data, we provided only holding-level counts, in accordance with the census instructions. Individual counts were available only for the cases included in the smaller Menard subsample of 2004. For the 2026 release, we tried to replicate as best as possible the approach taken in the Menard sample for the full population. The new release provides holding-level counts as well as individual indicators of fugitive or manumitted status, when so indicated on the census form. There are relatively few such individual-level cases, but we deemed them a potentially useful population for research. The new holding-level fugitive and manumission variables are all-inclusive; they sum up any individual counts for the holding.
The new 1860 data on disability, fugitive, and manumission status cover roughly 95 percent of the enslaved population. Disability data are unavailable for Oklahoma and most of West Virginia and Kentucky. Fugitive and manumission data are unavailable for Oklahoma and most of West Virginia.
Supported By
