Before I start ... broadly researching the period can give me plot points I would not have thought of, and leaves fewer pits to write myself out of.
I have a book where the author's entire premise was based on an assumption about English law that was not correct. Apparently the flaw was brought up by an editor because you can see the holes and patches where a fast fix was done.
Well, if the family consisted of morons, I can see how it could have happened...if your net is set up through MSN and it asks you to select a username and email, some might not understand that other accounts can be created. But that's neither here nor there.Like a lot of others-- I research the topic. Then, I write, putting XX in the manuscript where I need to research a specific detail. Then, I go back and research those details.
In the 1990's, I read a novel by a best-selling author where the entire plot revolved around a mistaken assumption regarding email. (She assumed that each computer could only have one email account, like each house only has one mailing address. So, a dozen people were using the same computer and could all read each other's mail.)
Before, during, after-- research is my favorite way to work on my stuff without working, yet I still feel like I'm working.
On a related note, how far are you willing to go in pursuit of authenticity for your work? Would you observe a surgical procedure if possible? View an autopsy? Skydive? Or would interviewing someone who had done what you're writing about be good enough for you?