The pull coefficient of 0.5 or 0.6 is certainly related to the mass the vehicle can pull and the weight on the wheel, the formula they use to get the coefficient is P/W or drawbar pull/wheel load, this proves that the coefficient is based on the mass the wheel can pull and not the force needed to pull it.
And the reasons for that have been explained. I'll try one more time.
The testing regime is operating on steady state propulsion. The wheel is not pulling the mass of the vehicle. The mass of the vehicle is not offering any resistance. On a flat, level surface once it is moving it will stay moving unless additional force is applied. That's basic physics, yes? All the wheel has to do is overcome the rolling resistance of the surface.
So why are they applying a load? Because once it starts climbing a slope (which is what they are applying their conclusions on the rover's behaviour to), the weight of the vehicle
is resisting the forward motion. The lunar gravity is providing an acceleration opposing the motion of the vehicle. As the incline increases, the portion of the rover's weight it now
does have to pull as a resisting load increases. That's what those ratios are. The portion of its weight now acting to resist its motion is determined by cosine (90 - incline in degrees). So at 25 degrees, the rover has to effectively exert a force of cos(90-25) = 0.42 x its weight all the time just to keep moving forward at a constant speed. Not its mass, its weight.
It is only when dealing with a motor trying to pull a load up an incline that drawbar pull becomes a limiting force that has to exceed the portion of the weight of the vehicle acting to resist the forward motion in order to allow the vehicle to pull the load. That's why I can shove a 3,000 lb car along a flat road but can't push it up an incline without assistance.