I've spent the last few days focused on leveling my hunter, and have managed to get him from level 69 to level 72. While I know it makes rational sense to stick to a rotational leveling system between all of my alts in order to maximize their rested xp and minimize their total /played time, there's something inherently satisfying about watching those levels tick by on a single character.
So I may throw logic and reason to the wind and just focus on getting my hunter those last few levels to 80. He's almost done with Borean Tundra at this point and getting ready to move on to Howling Fjord. I did try using the Dungeon Finder tool while leveling, but found that the xp per hour while instancing was significantly lower than what I was running while questing. It may have been a function of the group that I was with, which was somewhat slow (including two wipes), but the difference was around 40%.
In any case, I'll probably try it again once or twice just to see, but right now, the numbers don't look too good. Even if it is slower, if you're planning to use the character that you're leveling for endgame raiding, I'd probably still advocate leveling using the Dungeon Finder system, simply because of the skill you'll develop in grouping and in playing your chosen endgame spec.
Finally, while I expect it will be more difficult to quantify that my rested xp calculations, I expect that there is something more efficient about focusing your leveling on a single character. Switching back and forth between classes must result in some period early on where you are not as effective as you could be at leveling as you are "relearning" to play the class - even if it has just been a week or so since the last time you played that class - it's still not fresh in your mind and your reactions and skills aren't as sharp.
One way to compare might be to level one set of characters using the rotational system and then level another set one at a time and compare the /played times of the two. That would be quite an undertaking though. So we'll see.