As Tim alluded, tribology can get tricky. First order things get swamped by something that you forgot to account for, but more often people worry about second order things and forget to factor a really important item.
A few things to think about in the Motorman quiz:
The usual--but not exclusive--reason for a large journal is to increase load carrying capability with a given material combination. In general, this is not the way to reduce friction unless something else is being changed as well.
--local contact area (and resulting stress) after shaft deformation. The small shaft may carry the load, but if it deflects very far it will reduce operating clearance. A long, small diameter shaft that you attempt to run at close tolerances is a potential binding problem. Rarely an issue for a bellcrank installation using the sloppy fits of most ready-made parts!
--deflection and hugely non-precision parts result in the actual area in bearing being far less than the first simple view will use in a calculation. Hence, you get horrible correlation from these flawed calculations and actual results. This makes people think of this type of engineering work as "black magic." This one is knowable and should not be part of the “magic” grouping.
--when a material cold flows, the precision goes out the window, and the contact area is "adjusted" but just for the geometry/position when it sees the load. (Imagine a Teflon bearing that got pull-tested with the controls at neutral every time.) So now the area is even more unpredictable and will continue to change.
--Contact stress for galling or material transfer. Shouldn't be an issue for most things in a model airplane. But I wouldn't use K&S aluminum tubing for guides on solid steel leadouts, either. The tiny diameter and contact just at the rim of the tube where the lines may come out raked makes for a very small contact area. There are tables for galling stresses for various material combinations. It would be more likely to see this on an aluminum collet for an engine drive hub, especially if it was put together squeaky clean with no oil at all.
--Be aware that tables showing coefficients of friction make assumptions about surface quality and wear-in. If you do not duplicate these assumptions in your actual parts then your results will not seem textbook and you will be mumbling about black magic again. For a precision machine, you might need to burnish mating parts during assembly, or operate it thru a certain number of cycles to wear off any surface asperities to get a stable friction number. Kind of like we did with old technology engines. Wear-in for plastic parts is a whole different deal, and I would not assume friction will go down any time soon. So if the bellcrank/pushrod joint is too tight now, it might stay that way right up to the point you are fed up with the plane and either crash or get rid of it.
I just finished reworking a Brodak white nylon(?) bellcrank so that I was comfortable using it. The molded contact diameters were super loose to allow it to work since the diameters are not even close to round, and they have mold draft/shrinkage in them too. The wobble was so bad that I would have needed huge cutouts in an I-beam spar for a mid-sized plane. Not good. So I drilled out the crank and bushed it with K&S thin wall tubing, pressed in. The nylon(?) bushing was not going to help me because the hole in it was not a close fit to the 1/8" music wire shaft I planned to use instead of the 6-32 screw. So I replaced it with a metal bushing. Now have close tolerances and precision (well, relatively!) fits but have to provide lubrication because these materials are not self lubricating. This combination, in new condition, has low drag using a 3/8" (?) diameter journal.