I have found that the stepped approach was effective in Photoshop CS and earlier versions but with the new Bicubic Smoother and Bicubic Sharper that it makes no difference in the end result. People don't seem to be able to pick out which photo was adjusted by steps as opposed to the one that was done in one step using the correct Smoother or Sharper Bicubic Resampling technique.
Here is a link to an example reducing the photo size in three different ways. One was stepless using bicubic smoother, one stepless with bicubin sharper and one using enough steps to reduce the photo to the same size as the others using the default resampling technique.
http://www.rfredricksmith.com/threesaves.jpg
Can you tell which was done with which method? For that matter, can you see really much difference at all?
Cheers,
rfs