One place that we often find errors is in the census. Usually one family member gave the information to the census taker, and he or she might not actually know the facts. For instance, a wife who is at home when the census taker arrives may not know the birthplace of her husband’s parents.
It is easier, though, to believe that a “fact” in print is accurate. However, there are many reasons that an incorrect “fact” can creep into a printed work as well.
Consider the following confusion that I came across. A self-published (comb-bound) family history book states the following:
Clarence Gordon Self was born 28 March 1893 in Grant, Oklahoma, to Berry Crenshaw Self (born 14 November 1856 in Cumberland County, Kentucky to William Jennings and Marry C. “Polly” Guinn Self) and Viola Frances Bledsoe Self (born 5 March 1873 to George Chilton and Nancy Sloan Bledsoe in Russell County, Kentucky). Clarence lived with his family in Kentucky from age two to nine, where he spent time with his uncle Jim Self, a lawman of some repute.
The 1900 Census states that Clarence was living with his parents in Creelsboro, Kentucky, and that he was born in Kentucky.
The 1910 Census shows that he was living with his parents in Pittsburg County, Oklahoma, and that he was born in Texas.
In the 1920 Census he is shown living with his widowed mother in Washita County, Oklahoma. His birthplace is unreadable, but appears to have been Kentucky overwritten by Oklahoma.
The 1930 Census shows him married, living in Washita County, Oklahoma, and with a birthplace of Oklahoma.
His WWI draft registration card shows that he was born in Creelsboro, Kentucky.
My first inclination would be to believe that Clarence Self was born in Kentucky – because the census closest to his birth indicated so, and because he listed Creelsboro, Kentucky, on his draft registration card. Since he filled out the registration card, you would think that would be correct. However, the author of the book, a relative, seems to have “insider” information.
When you find conflicting evidence like this, you just have to consider all to be candidates until some conclusive evidence arises. Texas doesn’t seem like a very convincing candidate, but Grant, Oklahoma, and Creelsboro, Kentucky, both seem viable. Until someone comes up with a birth record from one of those places, this fact needs to stay in the “inconclusive” category.