Reddit reviews Statistical Mechanics

The thermal energy and entropy components can be fairly large, however if you're looking at quantities that involve differencing the free energies I've found a lot of the times they cancel fairly well (for reactions in particular). Note this is anecdotal, you should still redo them but just bear in mind those results you're building on may not be so different to what you get from doing it properly. I remember finding this doc really useful. Gaussian whitepapers are usually pretty good in my experience.
Solvation is a whole other pain (can completely change a reaction free energy surface) and I would recommend that at any point that you can compare a computed free energy to an experimental value you do so. I just had to scrap a few calculations because I didn't bother to look up the correct solvation free energy for hydroxide.

Edit: I learned from This book which will almost definitely be in your university library.

Ken Dill has the easiest to follow stat mech book I have encountered. McQuarrie has lots of good problems to work through. David Chandler is the shortest, and simultaneously most brilliant and difficult work on the subject I have read. His brief review of thermodynamics in the first couple chapters is fantastic if you only have a day or two to get back on the horse.

McQuarrie's Stat. Mech. text would likely serve you well...