You raise some good points, but...
800 dude wrote:
You should use a treadmill because your pacing will be inconsistent on the track. But the treadmill has to be research-grade, i.e., as hard as concrete.
Most runners (>99%?) do not have ready access to a 'research grade' treadmill. My pacing was not as consistent on a treadmill, but I plot the data for further analysis, and use the pace averaged over the full rep.
Gym treadmills are incredibly cushioned, which makes shoe cushioning redundant and will give results that favor the lighter, more minimal shoe.
That may or may not be true, depending upon a lot of factors; but I didn't use a treadmill.
The (bad) studies suggesting that running barefoot was more efficient than running with shoes were done on commercial treadmills rather than research treadmills.
The efficiency of barefoot running, just like running in a very lightweight flat, is significantly going to be determined on whether the runner has the structural capacity to do so (over any significant difference). One of the caveats I currently have with the Vanish-R is that I'm not sure (yet) I could take it out in a 5k road race without significant risk of injury. (But I saw a commenter over on the RunningWharehouse site say he set his marathon PR in them. Yikes! lol)