Correlations

Each person in the 1850 census was entered as a separate record in a PAF database (Personal Ancestral File version 2.31, a genealogy program).

Members of a household were linked into a family wherever it seemed "plausible." NOTE: THESE LINKS CONSTITUTE A STARTING POINT ONLY! The PAF database will be constantly modified in Stage II as new information comes to light.

Second marriages are especially likely to be problematic, as illustrated by the following family:

    Jacob Herndon   60
    Charlott [sic]  22
    Nancy           19
    Ruth            15
    Robert          13
    Joseph          11
    Susan           78

We interpreted this as Jacob, a widower with 5 children living at home plus his mother. I received a Herndon database, and it turns out that Charlotte is Jacob's second wife. Jacob's first wife died in 1840, after bearing him 12 children. Jacob and Charlotte married in 1850 and had another 14 children! Nonetheless, we can expect a good percentage of family links will later prove to be correct.

Records were entered in the same order as the census listings. No spellings were changed, even when it seemed likely that the census taker was mistaken. Certain conventions were followed (the "field" and "level" terminology below is from PAF, but most genealogy programs will have equivalents):

• Wives were entered with their husband's surname and the title field was flagged with "Mrs." to indicate that the birth surname was unknown.

• Ages were subtracted from 1850 and entered as "Abt" (about or circa) dates in the birth date field. For infants where the age was given in months, the birth dates were entered as "Abt 1850", since the census enumeration took place over a period of several months and more precision might be misleading.

• The state of birth (for example "Tenn." or "N.C.") was abbreviated with the 2-letter postal code and entered into Level 3. There were seven (!) residents who were born abroad; the country was entered into Level 4. Levels 1 and 2 of the birthplace field (for city and county) are empty; they may be filled in during Stage II.

• The ID field contains a reference to the 1850 census in the form HHHH-NN, where HHHH is the household number and NN is the person number listed within the household: for example, 0479-03 is the third person in household #479. Numbers are padded with zeroes so the computer can easily sort the records in census order. The ID field also provides a unique identification number which will not change when imported into other databases, in contrast to the RIN (Record Identification Number).

• Notes tagged with "OCCU:" for occupation and "LAND:" for land value complete the individual records. No source documentation is given in the notes for Stage I; the implicit source for all records is the census transcription cited above.

Certain fields of the PAF database were then extracted (using a custom program I wrote) into a comma-delimited record format for importing into a general purpose database program. No family links are present in this file; it is essentially a census database. The first line contains field names, in more or less the same order as the Womack transcription:

• RIN (Record Identification Number in the PAF database; assigned consecutively in census order)

• ID (HHHH-NN format)

• SEX (not printed in Womack's book; checked against census microfilm when given name was ambiguous)

• FIRST

• MIDDLE (only one middle initial is included for brevity)

• LAST

• AGE (converted back from birth date in the PAF database)

• OCCUPATION

• LAND (value)

• BIRTHPLACE

The first field, RIN, is provided as a convenience in cross-referencing the general purpose database with the genealogy database. It may be stripped without any adverse effects, since the records will sort the same way by using the ID field as a key.

Some advantages of a database program are the ability to search and sort on various fields, such as

• given name (patterns of unusual names occurring in several different families might provide clues)

• date of birth (candidates for parents might be born in a certain time frame)

• state of birth (migration patterns might become more obvious)

Various statistical summaries can also be performed with a database program, such as the number of children in school at various ages or the ratio of lawyers to farmers. If any readers would like to prepare such reports for the web site, I would be pleased to add them here or link to your home page.

Click here to download the database WCTN1850.ZIP (104K).

The author gratefully acknowledges the assistance of Kensey Alsman, Jim Davis, Doris Good, James Matthews, Sam Powell, Howard Powers, and Cynthia Whitlock in preparing Stage I. Together they entered and proofed the 8468 names in the 1850 census.

Home