I get lots of requests for lists of "best practices"...in e-learning, in the virtual classroom, in instructional design, in classroom presentation. Here's the deal: there's no such thing. A "best practice" is best only in the precise, specific context in which it exists. I don't recall who first offered this analogy, but think of it this way: what works in my marriage won't necessarily work in -- and may even damage -- yours. Even if moved from one situation to another very close one, the odds of transfer being made with practice intact is nil.
In education they call this a problem with "fidelity": one teacher writes a fabulously effective lesson plan and shares it with her friends. They each decide to 'adapt' it in a slightly different way to suit some unique need of their students. It is no longer the practice that was supposedly "best". Of course then, when the end users don't get the desired outcome, they say it's isn't their fault...because after all, they were using "best practices".
So how do we address those who pressure us to produce a list of, or abide by, "best" practices?
[Update: I ran into a great visual example of the problem of fidelity in best practices. Check out the update.]