Census records are the gateway to genealogical research for most people, and I am no exception. I can still remember hunkering over a microfilm reader in a dark corner of Davis Library in Chapel Hill, gaping at my great-great-great-grandparents’ names revealed in crabbed script in the 1910 federal population schedule. Like so many others, I squirmed impatiently for the release of the 1930 and 1940 censuses, anxious to determine what whos and wheres could be answered by the fresh infusion of data. As much as I have relied upon census data, however, I am acutely conscious of its limits. The census schedules are imperfect documents that qualify only barely as a primary resource. This is not to discount their usefulness for genealogical purposes. I’m just saying that — based as they are on a mishmash of personal knowledge, second-hand information, hearsay and rank speculation — they don’t prove much of anything about a person’s name, age, ethnicity, relationships, or occupation.

Here’s an example, courtesy of the 1920 census of Wilson, Wilson County, North Carolina:

A husband and wife with two daughters, no? If you didn’t have reason to know better, you might accept this at face value. “Hattie May” happens to be my paternal grandmother, however, so I do know better. (She was Hattie Mae, by the way.) Let’s take each person one-by-one:

“Jessie Jacobs” was Jesse A. Jacobs Jr. He was actually born in 1856, so was 63 or 64 years old, not 60, when the census taker stopped by. He is described as “B,” which is a designation he never would have provided. I am fairly certain that his wife gave information for the household, and I am equally certain that she described everyone in it as “colored.” Jesse himself might have offered “Croatan,” as the multi-racial, ethnically Native American members of the Coharie tribe were then called.

“Sara Jacobs” was Sarah Henderson Jacobs. She was, indeed, Jesse’s wife. She was born in 1872, so her age is a little off, too. She was 47 or 48, not 42.

“Mamie Jacobs” was born in 1907, so her age is basically correct. She was not, however, the daughter of either Jesse or Sarah Jacobs. Nor was she a Jacobs. She was the daughter of Bessie Henderson, who was the niece of Sarah H. Jacobs. In other words, Sarah was her great-aunt. Her mother died when she was three, and she was reared for her first eight years by her great-grandparents, Lewis and Margaret Balkcum Henderson. See:

Here, in the 1910 census of Brogden township, Wayne County, North Carolina, is a correctly described family unit. (This, by the way, is the census entry that dropped my jaw so many years ago and got me hooked.) My great-great-great-grandfather Lewis Henderson, great-great-great-grandmother Margaret, great-grandmother Bessie, and great-aunt Mamie. (Bessie was more than seven months pregnant with my grandmother when the census taker showed up on April 18. And look at how many children Margaret had lost. Only three of nine surviving. It breaks your heart.)

Back to 1920: “Hattie May Jacobs” was born in 1910, so her age is basically correct, too. She spent her first eight months or so in her great-grandparents house, but when Bessie died in the late winter of 1911, Sarah and Jesse Jacobs took her to Wilson to live with them. Mamie remained in Wayne County until her great-grandparents died, then she, too, went to Wilson. She and Hattie were known as Jacobses as a result, and for years my grandmother believed she had been formally adopted. Well into adulthood, when she learned that she had not, she reverted to her birth mother’s surname, Henderson.

Fast forward twenty years to the 1930 census of Wilson, Wilson County. Have things gotten better?

No. Sarah Jacobs was 58 years old, not 49. That was likely deliberate deception. Hattie, of course, was her great-niece, but their relationship was essentially mother-daughter and undoubtedly so reported to the census taker. Their occupations are not shown here, but Sarah was described as a laundress “at home” and my grandmother as a servant for a private family. The former accords with what I was told about Sarah’s work, but I have never heard that my grandmother worked as a maid. Most curious, however, is not what’s in this entry, but what is not. Namely, my two uncles. They were three and one in 1930, and I’ve found them listed nowhere else in the census either. A deliberate omission? A mistranscription? I don’t know, but it’s another stark example of the unreliability of census records.

So, three consecutive census schedules for one family and only the first reasonably accurate. As I’ll demonstrate in coming weeks, this was not the exception. Caveat emptor.