The crop factor of 1.6 on 30D means that the Field of View (FoV) of a 5D is 1.6 times that of a 20D; so for a 5D to have the same FoV to a 20D set at 50mm, it will have to be set at 80mm.
My question is: suppose a 20D and a 5D are set at the same distance from the object and are adjusted to have the same FoV, what will the difference in the Depth of Fields (DoFs), if there is any, between the two cameras be?

The 5D will have shallower DOF. I can't remember how much. Just over 1 stop, I think.

That is correct, I believe, unless BJL is wrong ) . I have to confess I've never actually tested this because the situation is complicated by the different performance of different lenses at different f/stops, as well as the different resolving capabilities of different format sensors.

I understand that theoretically, if one had a 12mp camera like the 20D with pixel quality equal to the 5D's, and if one had a 50mm lens which at f5 performed as well as the 80/1.8 (for example) at f8, then the DoF multiplier would be 1.6.

But these theoretical ideals rarely exist, so it's all approximation.

Sorry! It's not the 80/1.8 but the 85/1.8, so there's another complication.

There are depth of field calculators all over the web.

The difference is generally not that large when you're look at it with a calculator, but differences in the appearance of out of focus areas, etc.. seem to change more, and differences due to the different lenses used also exaggerate the visual difference.

e.x. for normal lenses

35mm film - 50mm - f/2.8 - 8ft - 1.3ft in focus
APS - 31mm - f/2.8 - 8ft - 2.93ft in focus (lens chosen as allowing most similar FoV & working distance)

Sigma has a 31, when I was using my 10D they didn't have it yet, so you ended up with a 28mm or 35mm lens for your normal.

APS - 28mm - f/2.8 - 8ft - 3.65ft in focus

But what the calculator doesn't mention is how out of focus a given object is X distance out of the depth of field.

That's where the differences in visual appearance come from IMO.

edit: #s are slightly off because a 10D/20D/30D is not exactly the same size as APS film.

I own a Rebel XT and 5D, and researched this topic a bit myself recently. I used the DoF calculator at

www.outsight.com/hyperfocal.html

to find that, for example, using my 85mm f/1.8 lens wide open on the Rebel XT focused at 30 feet (360 inches) gives me a DoF of 30.5 inches. Using the 135mm f/2.0 lens on the 5D will give me almost identical angle of view, and the DoF when focused at 30 feet is 30 inches using an aperture of f/2.8. Of course, I can then open up to f/2.0 and get an even shallower DoF (21.4 inches) or stay at f/2.8 and keep the extra resolution.

Unfortunately, I can't really comment on the difference in background bokeh because my informal testing was at my son's soccer game. Shooting any type of sports you just never get exactly the same shot twice. But I was very pleased with both camera/lens combinations for that use, and for the prints I'll make of that - likely no larger than 6x9 - the Rebel XT is sufficient, considerably lighter (2.1 lb versus 3.6 lb), and less conspicuous. So in the future I'm sticking with the "eight is enough" Rebel XT solution in that circumstance.

I must add that I'm very impressed with the results from the innocuous-looking little Rebel XT and 85mm f/1.8. Both are inexpensive little jewels of performance at $1,240 total versus $3,900 for the 5D with 135mm f/2.