Guest Post by Marlie Parker Wasserman

When writers of historical novels talk about their preparation, they usually focus on how many months or years they spend on research and how much of that research they end up using. It’s common to hear novelists say they spend one year on research and use under ten percent. But few novelists discuss one of the common challenges of research—how to deal with conflicting information, specifically regarding life and death years, occupations, and names.

How much time should we spend on accuracy? Novelists who set their stories in the United States in the nineteenth and twentieth centuries most likely rely on census data, genealogical databases, newspapers, and cemetery records for the birth and death dates of real characters who make an appearance as primary or secondary characters.

We start out thinking this will be easy. Then we find one set of dates in one place and another in another place. If we were professional historians writing a monograph, we might dedicate a paragraph or a footnote to recording the inconsistencies and weighing the evidence. As novelists, we know few if any of our readers will check up on us, yet we aim for accuracy.

In my own writing, I start by recording in my notes the variations in fact that I find, and their sources, and then eliminating those dates that seem unlikely—a man born in 1853 who I know fought in the Civil War, or a woman born in 1890 whom I later find married in 1900. After crossing out the low hanging fruit, I may still see conflicting dates in my notes. If the sources all appear reliable, that’s when I gravitate toward the date that works best for the flow of my story.

I struggle too with conflicting information about occupations. According to census data, a character I research may be a “workman” or an “office worker.” Such designations don’t add much life to a tale. The workman may excel as a skilled craftsman or work as an unskilled hauler. The office worker could be a junior clerk or a bank president.

Sometimes, I can look at three different census reports, covering three different decades, and learn of an upward or downward trajectory. More often, I am puzzled. I search newspaper articles, looking for so-and-so’s occupation, with mixed results. Only after I’ve explored a variety of sources and found no conclusive information do I flesh out a character’s occupation according to the needs of my story.

We might think names of real characters would be clearcut. But clerks filling in census data often used longhand that may be faded or unreadable. Or the person being counted, named Ted, might feel like an Ed on the day the census taker shows up. Or the father providing the census date might think of his daughter with her proper name, Eleanor, while she always calls herself Nell. Today, parents select the names of their newborns from a wide range of possibilities, but in the nineteenth century, we find seemingly endless numbers of Franks, Williams, Bridgets, and Thelmas.

If you do an internet or genealogical search for Frank Johnson, I wish you well.

The problem of repeating names leads to problems in readability. In my writing I have had to change the names of some real figures to avoid confusion on the readers’ part. For instance, if three men named Frank—all historic characters—play a role in my story, I tend to let the best known of the three Franks retain his name, while I rename the less well-known. I am not proud of such a choice, but I defend the unequal treatment because my primary goal is clarity for readers.

Why do I worry about accuracy when I am writing fiction? I believe that when we aspire to write historical fiction, we are in effect making a pact with the reader that we will not change facts but will fill in the empty spaces around facts. Even if a reader does not care whether a particular character is born in 1853 or 1855, we as authors care. We want to write with authority about what we know and with imagination about what we don’t know. But while we strive for accuracy, we should acknowledge that accuracy may be elusive.

---