This recent contribution by Nicholas and Gordon can be found here from the EOS AGU Journal of July the 12th. This below the flowchart:
The paper is really a good guide. However, it does not tell all the truth. What moves me as a reviewer (not considering that I accepted duties as associate editor or editor) is the curiosity to know a piece of exciting research a little before the rest of the guys. On the other hand, the expectations, are very often, not maintained by the incoming papers, and very few reward the efforts. However, this allows you to know what people thinks is important today, as opposed to what you (me) think it is, and is all experience gained. This can be summarized as: reviewing keep you close to active research.
A second aspect is that rarely you are completely expert of the subject treated. Most of the details of a paper from Authors you did not frequent before (on a topic you supposedly know enough), refers to paper and methods that you do not completely possess. This add sweat to your work, since you need to search and look to other papers to review one. The positive of this that you increase your knowledge. And the efforts are usually paid back with self-consciousness of what you know of a subject.
With time experience and erudition grow, and therefore the task or reviewing becomes easier, because you start to recognize the presence or absence of the structural patterns that makes of a paper a good paper. Outstanding paper are obviously matter or the science they contain.
Now, going back to my reviews.