the 7 frame test used 'ordinary Joes' as 'testers'. This is a bit like asking a bunch of novice violinists to give their impressions of how playing a Stradivarius differs from a lesser instrument, i.e. you would be lucky to get any sense out of it. Any test riding on the road needs to be carefully designed in order that random factors (such as how many cars go past you) don't dominate the results. A draggy climb at a fixed workrate seems like a good experiment and the route they used in the 7-frame test wasn't.
More recently BQ did a 3-frame, 3-rider test using more experienced cyclists, and the (blind-tested) bikes were weighted internally to be the same mass. Of the three riders, two expressed a clear preference for the lighter-built frames and went faster on them (at a given pulse rate on a draggy climb). The third rider couldn't tell the difference but still went faster.
I think it is fair to say that different pedal strokes suit different frames better than others, and that (on the road) the crunch test is whether you are still pedalling nice circles or not after a day in the saddle.
My personal theory is that a frame without enough flex (lets call it 'flexless' although this is a slight misnomer) doesn't 'talk back to you' enough and thus when you get tired you are more likely to pedal badly. But that is just a theory.
In terms of strength affecting how tubing behaves below the yield point; this can fairly easily be demonstrated. If you take two pieces of steel of the same geometry but different strengths, and suspend them so that they can resonate when struck (like a bell), you can hear a clear difference, even though neither is getting anywhere near yield stress when struck and remains undeformed. This may translate into the efficiency of the frameset (as a spring) when pedalling.
cheers