Quickly, because I am exhausted and cranky and I'm sitting alone on my boat when I'm supposed to be in sunny Florida on vacation:
The Lazy Database Admin's Research Truisms
1. Garbage In, Garbage Out
Anyone can get data from a search. Anyone. It's just not always (if ever) going to mean what you think it means, because the *real* issue is learning to ask the correct questions in the correct ways to extract data that means something. When you begin research, be as specific as possible. "Gotta learn about the middle ages." Uhm. You do know that topic is an entire lifetime of study, right? So how's about we narrow that little. What aspect concerns your story? Castle life? What aspect of castle life? Moats? Privies? Rushes on the floor? Did people bath? If all you need is to know whether you can have your heroine lounging in a bathtub in the tower so the hero can walk in on her - you've just saved yourself an incredible amount of work and increased your likelihood of finding a useful answer.
2. Never Rewrite a Search When You Can Steal Someone Else's
It's likely you aren't the first to want to know what you want to know. Use that. Pay attention to the people around you. Your chapter mates, if you belong to a writing group. Your critique partners. Your online communities. In just the three groups I've mentioned, I can get detailed genetic information on how color works in cats. Or I can get stupidly granular detail about Civil War Naval ships and battles. Or I can get anything I never wanted to know about running a horse farm. Granted. This may all be useless to my stories, but suppose (as once happened) I wanted to know how to swear in Romanian. True story. I asked my international cat fanciers association. Within two hours, I have three people bringing me identical lists of words and their translations. Never underestimate your connections.
3. Verify, Verify, Verify
Running your search and getting an answer once is nice, but if you cannot reliably reproduce your results, you got nothing. So. Gotta have a secondary source. Doesn't count without backup. Whatever it is you're searching for, you haven't found it until you have found it at least twice in two different searches. Eh. Okay. My analogy breaks down. For me to verify a search on a database, I have to rewrite my query a couple of ways and make certain I'm getting the data I need rather than what I expect. Subtle difference. When talking about research, I'm only saying you can't stop with finding one instance of the answer you're looking for. You have to verify it with a second source. More, if the tidbit of info you've discovered has any controversy associated with it.
There's more, I'm sure, but here's the other dba truism: I need a drink.