ok, so i did some maths using (approximately) the data in my link.
using continental ultra gator skin tires (no numbers are given for diameter or inflation pressure), rolling resistance will adds about 7741 joules to a mile, at 18.6 mph.
using continental grand canyon tires (26" x 2.125") inflated to about 70 psi , rolling resistance adds about 7885 joules to a mile (no speed is given).
that's about a 2% difference. but i just looked and saw the mtb tire tests were done with a 40kg load, and the road tests were done with an 85kg load. whoops! the link suggests that the resistance may increase linearly with load, so let's assume it does. that gives 16755 joules per mile. that implies that mountain bikes would require 2.16 times as much energy as road bikes, based only on rolling resistance.
conclusion: the increase with load isn't fully linear, but i imagine that the difference is still pretty significant. there's also a lot of missing control data for these comparisons. i wouldn't be surprised if edmonds is close.