Can 19th century US Census records be reliably used to identify a family of origin?

Question

I am trying to establish the birth family of Martha (L.) Stewart, born in Ohio 1832-1840 (1860 & 1900 census records). She is identified in the 1860 census living in Pease Township, Belmont Co., Ohio, with her husband, Henry Morrison, and three young children, the oldest born in Aug-Sept 1852 in Ohio. Family oral history claims that her father was James Stewart and that he was born in Ireland in 1804. To begin looking for Martha’s family of origin, I searched the US 1850 census

Assuming that: 1) the detail about Martha’s father being from Ireland is correct 2) Martha would have still been living with her family of origin at the time of the 1850 US census (October 8th). 3) she would have been 16 – 20 years old in 1850, judging from later census records and the timing of the birth of her first child, which I believe was legitimate. 4) her maiden name Stewart is probably correct, or at least some variant of it (Stuart/Steward/Stwart), but her father’s first name could vary. 5) the age of her father could also be uncertain

Search terms used in the 1850 US census on familysearch were: bp = Ireland, b 1790-1810, with first name and state of residence changed, depending on the search.

James Stewart, res. = Ohio resulted in 9 matches, only one of which has a Martha in the family, age 16. This seemed to fit, but further research (findagrave & written local histories) suggests that this Martha married a Durand and moved to Cincinnati by 1860 with her new husband.

A search for John Stewart, res. = Ohio returns 2 matches, neither of which are age-appropriate or have a Martha in the family.

Jas Stewart gives no matches.

Expanding the search to neighboring states, there are no matches using the search criteria above for Virginia or Illinois (James or John). For Pennsylvania: 12 results for James, 11 for John; only 3 with Marthas in the family and one is a 60 year old married woman. One of the remaining Martha Stewarts (yes, I know its funny) is with her family in Philadelphia:

John Stewart, b 1810, living in Kensington, Philadelphia; ALL BORN IN IRELAND: Elizabeth, 48 Martha Jane, 17 Robert, 15 Elizabeth, 13 Margaret, 12 John, 10 BUT they where all born in Ireland. I’ve seen place-of-birth mistakes on census records before, but tending towards listing a US place of birth for an immigrant, not the other way around. I also cannot find this family in PA or OHIO, or anywhere in the US in 1860 census. It is certainly possible that the parents had died and the children all married, but this is just a preliminary search and by no means exhaustive.

The third match is for a family in North Slippery Rock, Lawrence Co., PA: John STEWART: 53, male, farmer, real estate valued at $5200, b. in Ireland ~1797 Mary, 25, farmer, b. Ireland ~1825 Hugh, 23, farmer, b. Ireland ~1827 John, 22, Saddler, b. Ireland ~1828 David, 8, at school, b. prob. Penn ~1842 Martha, 16, b. Ohio ~1834 Sarah, 12, b. Penn ~1838 I am very tempted to say that this is my Martha Stewart’s family of origin and pursue other records for them. A preliminary census and land record search does locate the elder sons in the same area through 1900, but I still have not located Martha’s marriage record, and her death record is ordered, but not arrived yet. Cannot find a death record or will for John the elder. Again, I realize that this is just the beginning of the search, but my question is, can I reliably use the 1850 Census records above to identify Martha’s family of origin?

Answer 1

Questions of this kind are a sign that the researcher needs to consider making the transition from person-based to source-based genealogy.

In a post dated Tuesday, March 22, 2011 titled The Chasm, the blogger The Ancestry Insider talked about different stages of doing genealogy research. He says:

There are three time frames or stages of ancestral research. Think of them in this fashion:

1. When we start, we fill our pedigree with people we know. It is easy.

2. We extend our pedigrees with people we know only through vital records. Vital records provide a complete picture of an ancestor. Genealogy is still mostly easy.

3. As we push further back, things take a distinct turn for the worst. Research becomes like something from the Truman Show movie. Truman doesn't have a photograph of the mysterious Sylvia, so he clandestinely tears scraps from magazine advertisements to recreate her likeness. One ad provides eyes, another her mouth. Gradually he builds a complete picture.

That's a bit like what we must do as genealogists once we push beyond vital records.

Part of the reason we get ourselves into this pickle is that we are focusing on finding people in the records, and/or finding records that match what we are looking for, instead of treating the source itself as the focus of our research. We cherry-pick records to suit ourselves during the Insider's Stage 1 and Stage 2, then we get to Stage 3, where we can't depend on a single record to hold direct answers to our question, and we're at a loss for what to do next. In many cases, we haven't been analyzing our information, or writing proof arguments all along, so now when we can't find things in the easy records, we don't know what to do.

So when trying to decide if the clues in a particular record are worth following up, you could start by assessing your own skills.

• Can you evaluate records using the 3 x 3 evidence analysis process map, as described in Elizabeth Shown Mills' QuickLesson 17: The Evidence Analysis Process Map?
• When you analyze a record, are you aware of the kind of pitfalls outlined by Dr. Thomas W. Jones in his article Perils of Source Snobbery?
• Are you familiar with the elements of the Genealogical Proof Standard? Do you understand how they can help your research?

The next step is to learn more about the 1850 Census itself. When I am starting out on a hard problem, I build a research notebook about the record group I am about to study so I can make sure I am not missing out on clues. For the 1850 Census, my list of things to review would include:

• The Overview of the 1850 Census from the US Census Bureau's website.
• Documentation for the 1850 Census from IPUMS, including the list of questions and the instructions given to the enumerators.
• Claire Kluskens' article "Who Talked to the Census Taker? [1790-1870]," by NGS NewsMagazine, Vol. 31, No. 4 (Oct.-Dec. 2005): 32-35.
• a checklist of all of the available schedules for that census year -- so I can remember to search for the slave schedule, the agricultural schedule, the industrial schedule, and the mortality schedule -- and the blank forms for each schedule.
• If I am searching on FamilySearch.org, I review the article on Known Issues for the 1850 Census in the FamilySearch Wiki. I also look for information at the beginning of the microfilm roll and on archives.gov to see if there are any geographical areas whose schedules have not survived.
• articles about the census like the FamilySearch Wiki's United States Census, 1850 (FamilySearch Historical Records)

These resources and others like them can help me evaluate the information contained in any given household record in a census record, and can help me watch out for mistaken assumptions (such as assuming a record group holds a complete set of people for that area and time -- record groups rarely do) that I may be making about what I am seeing in a record. But like all records, a census is most valuable when used in conjunction with other record groups.

In the case of the 1850 Census, it's important to remember that the 1850 census did not ask for relationships of the household members to the head of household. It's easy to assume that these are parents and children, but we don't know that simply by looking at that one record. To identify the parents, we must use other records in conjunction with the census.

It's important to remember that the census, like most records we use for genealogy, was not created for the purpose of studying family history. We can learn a lot by reading studies from scholars in other disciplines, like archivists and those doing population studies. Reading the reports from the census bureau can give us a "big picture" view that often gets forgotten as we look for the fine details.

---

Other resources that may be useful:

• The Hidden Half of the Family: A Sourcebook for Women's Genealogy by Christina K Schaefer
• QuickLesson 11: Identity Problems & the FAN Principle (a case study which uses the 1850 Census)
• FamilySearch Wiki:United States Record Selection Table
• The Ancestor Hunt: Ohio Online Historical Newspapers Summary
• US Newspaper Directory at the Library of Congress
• Webinar: Complex Evidence - What is It? How Does it Work? And Why Does it Matter? by Warren Bittner, CG, for purchase or by subscription at Legacy Family Tree Webinars or via pay-per-view or purchase on Vimeo.
• Genealogical Proof Standard by Christine Rose (FamilySearch Lessons)
• Webinar: Inferential Genealogy by Dr. Tom Jones, CG, CGL, FASG, FUGA
• the handout for Dr. Jones' video
• FamilySearch Research Wiki: Inferential Genealogy
• 3 Ways to Unravel the Mysteries of Women in Your Family Tree by Alison J. Herzog, posted on the FamilySearch blog on March 16, 2015.

Answer 2

You may need to revisit your assumptions, particularly #2. Consider the following alternate scenarios:

(a) you don't indicate a DOD for James Stewart - if it's possible that he died before 1850, then Martha might be listed in the household of her mother (and possibly step-father), her maternal grandparents, or a guardian;

(b) Martha may have working as a domestic in another household;

(c) Martha may have been already married by Oct 1850 - although the oldest known child may have been b.1852, there may have earlier-born non-surviving children (children who died in infancy aren't always accounted for in the 1900+ census "# children/# living" entries).

Although 1850 census records may help in any of these cases, again, other record sources (church/cemetery records, probate/orphan's court records, etc.) might help to support or disprove these alternate scenarios.