Tire rolling resistance varies widely based on carcass construction and tread compound, amongst other factors.
True, and rotational inertia varies significantly with tire construction specifics also. Because the question did not contain specific tire models I was answering the question while assuming undefined variables remained the same.
One can't assume two tires with different dimensions will have comparable RR. (Ten years as an engineer at Michelin R&D responsible for drum-based RR testing, but I didn't sleep in a Holiday Inn last night).
I agree, with similar rubber compounds, similar carcass stiffness and similar tread designs, the wider tire will have more rolling resistance. And as a former engineer focused on tire rolling resistance for Michelin, I can see how easy it would be to focus on rolling resistance while minimizing the relative importance of aerodynamics at high speed.
Either way, we're just blathering on about engineering nit-picks. For city driving, tire/wheel rotational inertia trumps all this other silliness.
But it is fun to ponder...
Yes, aerodynamics do not come into play (significantly) at city speeds. That's why I limited my comment to "high speed" driving. And the only reason I brought up aerodynamics which is an (admittedly) small effect is because the two tire/wheel sizes in the initial question appear to have very similar rotational inertia.
As an engineer experienced in measuring rolling resistance, I have to ask, how large of a test drum was considered large enough to reduce the inaccuracies introduced by the drum surface being round, not flat like a road, to an acceptable level? Because I can see how two tires with the same RR on a flat road could have different measured RR if the drum diameter was too small.
Also, did any of your RR test drums have actual sharp, coarse gravel glued to them to simulate the kind of chip seal roads so popular on many secondary roads in the American West? Or is the RR testing primarily limited to carcass construction and not so concerned with rubber compounds and tread design differences?
One last thought to ponder. If I put 2" wide knobbies on my mountain bike, use a stand to elevate the wheels off the ground, put the bike in high gear and use my hands to pedal the crank at a normal 60 rpm riding cadence, the rear tire creates quite a breeze and I must maintain significantly heavy pressure on the crank to maintain 60 rpm's. Conversely, if I do the same thing with 2" wide city slicks mounted, very little breeze is created and very light crank pressure will maintain 60 rpm's. This difference simulates 50% of the aerodynamic efficiency difference between slicks and knobbies (and none of the RR differences) because in actual use there is also a front wheel spinning (and there is friction with the ground).
As a RR test engineer, how were aerodynamic efficiency differences handled. I can see they might be ignored as being insignificant or any small differences could be considered to be "rolled into" RR measurements or a fan and a circular fairing could be closed around the test tire to minimize aerodynamic effects. Or, were you testing different carcass constructions only and the tires all had either the same tread or were slicks?
I would happily sleep in Holiday Inns for a week if I could learn more about tire testing.