I remember a Michael Ondaatje reading where he was asked about his research (for Anil’s Ghost). He said he liked to immerse himself in a place or a time period and read all the papers and books from that era. Patrick O’Brian told of spending years at the Admiralty reading the logs and Gazette articles about the various actions in his stories. After a point, though, he was familiar enough with the subject matter that he declared further research was unnecessary. He also was writing a long series about a ten-year period, and he quickly ran out of liner time. His solution was to create what he called an “1812 A and an 1812 B.”
One of the great things about writing in this modern era is the availability of ready research. When I’m writing anything, I always pop back and forth between source material I keep handy. I wrote a story about a lonely man in 1970s Montreal, and through the power of Google was able to find ample information about the city and its character. I never trust one source, though… I use something like Wikipedia as a starting point, and then find supplementary pieces from various books, blogs and newspaper articles.
Another fringe benefit of doing this is that I often will discover some element that can take the story in a new direction. For example, my current novel-in-progress features Hawes in 1948 Los Angeles investigating what may or may not be a crime (I’ll never tell). In doing the research for the period, I have found all kinds of new directions and plot elements that I never could have come up with on my own. As with Hawser, I am doing my best to have my wholly fictional characters interact with very real people along the way. I do my best to avoid some kind of jarring anachronism in speech or attitude or (God forbid) a current event.
By the way, even a great writer like Michael Ondaatje can make a mistake. In Anil’s Ghost, a woman in Tucson spots an armadillo on the road. I grew up in Tucson, and nobody I knew ever saw an Armadillo. They never venture that far west.