My first thoughts on using Dragon Naturally Speaking 10 (preferred) 
1) The Computer: Once trained, and configured, Dragon informed me that due to limited resources, they turned off some features, choosing speed over accuracy.
I need a faster computer with much more RAM. I’ve known this for a while but after the Nth software program complained about my computer, I feel motivated to get a new desktop. The kids have the fastest computer right now but it sounds like a jet engine – the short-term answer may be to swap machines. I would lose a few days wiping, reinstalling, and configuring but the gain may be worth it.
2) The Headphones: The headphones do matter. I used my husband’s headphones that are earmuffs with a microphone. They are tight on my head and I wear glasses so it quickly became uncomfortable. I bought the version of Dragon NaturallySpeaking that came with headphones. These were much lighter and not as tight on my ears.
3) Time: The software needed to be trained to listen to me, which took time. Expect at least one hour to install and train the software. You need to be crisp in your diction and awake. If you have a groggy voice early in the morning, or a nasal twang due to congestion, consider waiting until later in the day.
My favorite part of training was this dialog box where I chose which text I would repeat to train the software. 
I choose Dave Barry in Cyberspace. This was a little bit of a tongue twister, not on purpose, but it was a medium read. The software also has choices for children and teens. This means your kids can speak letters to Grandma before they can write. Cool, huh?