6 Strategies for Interpreting the Census

By David A. Fryxell Premium

Genealogists have a love-hate relationship with the US census. Yes, we love the reams of data those every 10-years population tabulations created, but we loathe inscrutably handwritten enumeration sheets. Between census appearances, ancestors often seem to switch not just residences but birth dates, too. Names can be indecipherable or all too common (is that your James Smith or somebody else’s?). Offspring mysteriously appear and disappear. Worst, in some head counts, your ancestors simply refuse to be found at all.

Such vagaries aren’t uncommon, given immigrants’ speaking accents, census takers’ recording errors and less-than-fully informed neighbors’ reports when our ancestors weren’t home. Not to mention modern indexers’ challenging task of interpreting loopy handwriting on fuzzy microfilm. But with a few insider tips and tricks, you can not only find your ancestors in Uncle Sam’s enumerations, but also tease out hidden information about their lives. Try these six strategies for reading between the lines of the census.

1. Follow the family.

The pitfalls of census research became obvious to me recently when I tried to trace little Jimmy McKinn. At age 11, or may be 12 or 13, none other than Geronimo abducted Jimmy from his family’s New Mexico ranch. Remarkably, months later he was rescued – against his will, as he’d become “thoroughly Indianized.” Perhaps his traumatic experience some how explains why in his later life (as I discovered in the census) Jimmy kept changing the facts of his birth. None the less, I knew I had my man because the family members matched in the censuses I consulted.

Normally you search the most-recent census during your relative’s life and work back. But the only residence I knew for him my was his childhood home in Grant County, NM, so I started there and traced him forward. He was rescued in 1886; because fire destroyed most of the 1890 census, I found him in 1900: James McKinn, born July 17,1873 (making him 12 when abducted), his father born in Ireland – matching details I knew. He had a wife, Victoria, born in 1878, and two daughters, Josefa (1895) and Victoria (1899).

They were key to figuring out the James McKinn I found 20 years later in Phoenix, where family lore said he’d moved, had to be the same man. The 1920 “Jim McKinn now gave his father’s birth place view Mexico and his own age as 45.

For censuses listing only ages, not birth dates, use the official census date to calculate his birth year – it was Jan. 1 for this head count, making Jim born in 1874. But he still was married to Victoria, five years his junior, and they still had a daughter, now spelled Josepha, born in New Mexico in 1895. Little Victoria might now be married or may have died in infancy. He’d added four children: Prospero, Pete, Josephina and John.

Those kids helped me match a Jim McKinn in the 1930 census, even though this guy was age 54 (so, born in 1875 or 1876) and gave his own and his kids’ birthplace as Arizona. Yet he was still married to Victoria and three of the 1920 off spring still lived with them, all with the right ages and names. Couldn’t be a coincidence. If I’d simply searched for a James McKinn born in 1873 in New Mexico, rejecting any listing that didn’t match exactly, I would’ve skipped right over him. But linking the family from one head count to the next made me sure I had the former captive of Geronimo.

2. Notice the neighbors.

Keep in mind the folks enumerated above and below your ancestors might not be merely their neighbors – they could be relatives. If I needed further proof my 1920 and 1930 Jims were the same, had to look one entry earlier: In 1930, there was his son Prospero living next door with a family of his own.

Or the neighbors may not be related yet. Before today’s mobile society, the most available candidates for marriage were people who lived near by. For example, I found my great-great-grandfather John Stowe and his family in Lee County, Ala., using the free 1880 census search on Family Search <>. That sent me to the library for the microfilm page, where I discovered my great-grandmother Jerusha Ogles by at the time, the widow of Henry Lowe listed as “J. Lowe… Keeping house.” Her family included my great-grandfather Henry “Harry” Lowe, age 14 – who eight years later would marry Nisba Stowe, daughter of John, listed 19 lines below him.

That’s a good reminder to search for widows by their married and maiden names: I wouldn’t normally have thought to search online for J. Lowe, and my search for the unusual first name Jerusha – usually a good bet – drew a blank.

3. Tally the tick marks.

Censuses from 1790 through 1840 list only heads of household. Everybody else in the family got reduced to a little tick marking a column for age, gender and race. Note these columns change from one census to the next; you can download blank census recording forms showing the column headings at <>.

Lacking a spouse’s or children’s names, how can you tell if a family really represents your ancestors? While you can’t be absolutely certain, you can test for a match – or eliminate a red herring – by comparing the tick marks to what you know about the family. For instance, I’d found a Martha Pitts in an index to the 1810 North Carolina census. But was this my fourth-great-grandmother, widow of Noel Pitts? I compared that census entry to what I knew about Martha and her offspring, using their approximate ages in 1810. Here was my guess at matching names to the tick marks:

  • Two males under age 10 = sons Jack, 6, and Nestor, 10
  • Two males “of 10 & under 16” = sons Nicholas, 13, and Solomon, about 13
  • One female under age 10 = daughter Elizabeth, about 5
  • One female age 10 to 15 = daughter Nancy, 12
  • One female “of 16 & under 26” = daughter Frances, 17
  • One female “of 26 & under 45” = Martha herself, 32

That’s about as exact a match as you’ll find, given the vagaries of early census-taking. (Again, knowing the census day is important, because a one-year birth year difference could throw someone into a different age bracket.) Don’t forget to match the head of the household and, if it’s a man, his wife, to their own marks. And don’t be jarred by a mark or two you can’t put a name to: It could be a child you didn’t know about, a boarder or live-in relative, or the enumerator simply might’ve goofed.

4. Intuit other information.

Finding Martha Pitts as the head of her house hold in 1810 was a clue her husband, Noel Pitts, had either died or fled his family since his last census appearance in 1800. I could’ve penciled in “died 1800-1810” for Noel and started narrowing that 10-year range – but merely failing to find some one in a census doesn’t mean he died. At times, whole households were skipped or listed as erroneously you may not find them. Try to locate the rest of the family recorded without your target person. And check mortality schedules, which, from 1850 through 1880, listed people who died in the previous year.

Reading between the lines of the census can help you reconstruct whole households and their migrations. In pre-1850 head counts, this requires educated guesses about the identities behind the tick marks. For example, I’d found an 1825 North Carolina marriage record for an Abraham Stow Jr., showing bondsman Joel Stow – quite possibly my ancestor Joel Stow, born in North Carolina in 1787. Although the “junior” designation didn’t always in play family relation ship (some times it distinguished two same-named men in a town), I hypothesized that if Abraham and Joel were brothers, their father could’ve been Abraham Sr. A census hunt for Abraham stow tests this theory: I had to recreate the Stows’ house holds, decade by decade. In 1840, I found an Abraham Stow, age 80 to 90, plus a John Stow, age 50 to 60, in Surry County, NC. By that year, according to the census, my ancestor Joel had moved to Tallapoosa County, Ala. So far, so good – the elderly Abraham could be his father, and John, a brother.

Next, in 1830, I found Joel Stow in Surry County (the house hold’s tick marks neatly matched what I knew about his children) on the same page with an Abraham Stow, age 20 to 30 – the recently married brother. John showed up 10 pages later. But where was Abraham Sr.? I looked again at the junior Abraham’s listing: Besides a young wife, marks indicated a male, 70 to 80, and a female, 60 to 70 – likely Abraham’s parents (of course, it’s possible they represent his wife’s parents). In 1820, I found separate entries for the elder “Abram,” Joel and John under the spelling Stoe; Junior would be among the four young males in Abraham’s household. I’ve only been able to find John in the 1810 census, but the elder Abraham shows up again in 1800.

There, with a little more educated guesswork, the pieces fit: Joel, age 13, is the young male between 10 and 15; the possible older brother John could be the male aged 16 to 25; younger brother Abraham Jr. isn’t yet born; and Abraham Sr. is the 26-to-45-year-old male, presumably on the older end of the scale. In his early 40s, he’d be a perfect match for the 80-to-90-year-old Abraham of 1840. (There’s an Abraham in the 1790 census, too, but its age categories are so broad as to be unhelpful.)

Not all the Stow family questions are so neatly answered, of course. Where’s Abraham Jr. in 1840? Not in Alabama with brother Joel. Where’s Abraham Sr’s wife in 1800, when the only older female listed is too old at 45-plus? Is the younger wife of later years a second marriage? Where’s everybody else in 1810? New questions, new answers to look for.

5. Go for the geography.

Lacking a birth certificate, how did I know Joel Stow was born in North Carolina? Fortunately, beginning with the 1850 census, enumerators asked each person’s birthplace; the 1880 census was the fist to add birthplace’s of everyone’s parents. So even if Joel hadn’t lived long enough to have his birth place recorded with his own entry, I could’ve done some “time traveling” to figure out where he was born. Indeed, I confirmed Joel’s North Carolina origins in the 1880 and 1900 census listings of his on John. So don’t despair if you’ve been unable to find an ancestor’s native state.

Jimmy McKinn’s loose grasp of geography notwithstanding, you often can use the birthplace columns to confirm (or eliminate) ancestral “suspects.” Already knowing Joel Stow was born in North Carolina, I could be pretty sure the 1860 Joel Stowe living in Tallapoosa, Ala., with his daughter Rachel was the same man – the age and the birthplace matched. And, thinking in reverse, I knew my 1900 John Stowe was Joel’s son because the census listed his father as born in North Carolina. The addition of an to the name (family legend says John’s wife added it to make the name look “more finished”) didn’t throw me off.

6. Combine with other sources.

Adding up information from censuses and other sources can yield results. If Noel Pitts was dead before the 1810 census, when did he die? I found the answer – Nov. 7, 1807 – in the supporting material some one submitted with a Daughters of the American Revolution<> application.

You also can use census finds to leapfrog to fresh facts. Once my Jimmy McKinn chain led me to his son Prospero, I was able to find a 1985 death record for Prospero McKinn – which listed his mother’s maiden name, Villanueva.

Sometimes you’ll make break throughs by bouncing between the census and other resources, particularly land and tax records. I finally solved the quandary of my Ekstrom family, for instance, by combining census are cords and city directories. After arriving in America from Sweden, the widowed Anna Maja Pehrsdotter Ekstrom showed up in the 1880 census as – believe it or not – Mary Van Kirkhoon, married to a Belgian gentleman. Unlikely as this seemed, that Mary was the only person in my census search who fit: Born in Sweden, about 43 years old, living in Moline, Ill., first name Anna or Mary (Maja in Swedish) – and her children had the surname Ekstrom. Could I prove Mary was Anna Maja?

I turned to Moline city directories for the period and soon found her widowed again and back to her fist husband’s Ekstrom surname. I compared her children in the 1880 census with other city directory listings and with her offspring back in Sweden. After finding her in the 1900 census as “Anna M. Ekstrom,” I matched the Ekstrom children from that census to those in the 1880 Van Kirkhoon house. I then found their obituaries, confirming their relationship to Anna and their sister Mary, my great-grandmother. This puzzle, which began with that mysterious 1880 census entry, needed other resources for the pieces to mesh.

Will these approaches always help you find a missing ancestor or use the census to fill blanks in your family tree? Some census mysteries may remain forever unsolved, but that doesn’t mean you should stop plugging away. Sometimes all it takes is a fresh look: I hadn’t been able to find my ancestor Joel Stow in the 1820 census until I was researching this article – there he was, waiting for me, as Joel Stoe.

On the other hand, I still haven’t been able to find ex-Geronimo captive Jimmy McKinn in the 1910 census. Maybe he got fed up with city life and ran off to the wilderness for a while. In census research, you learn to expect the unexpected.

A version of this article appeared in the May 2007 issue of Family Tree Magazine.

New course: Using U.S. Vital Records

In this 4-week course starting July 19, you’ll learn the basics on how to find and analyze three key records – birth, marriage, and death – for your U.S. ancestors.