By way of agreeing with what Reg posted about capacitance indirectly
adding to the loss, consider that for any TEM line (twin-lead and coax
being two examples), the impedance, Zo, is sqrt(L/C), and the
propagation delay, tau, is sqrt(LC) [neglecting the very small
contribution of R and G for practical lines at HF and above]. From
these two, you can see that C=tau/Zo. If the velocity factor is unity,
then tau for a foot of line is one foot divided by the speed of light,
about 1.017 nanoseconds. If Zo is 50 ohms, then C for that line would
be 20.33pF/foot. If you have line which you know to be 50 ohms and
31.0pF/foot, then you know the v.f. is 20.33/31.0 = 0.656, and by my
other recent posting in this thread, you know that its attenuation will
be about 1/0.656 = 1.52 times as many dB/unit length as the same line
with air dielectric (which would be 50 ohms times 1.52 = 76 ohms).

(The interrelation of tau, Z, C, L, line physical length and velocity
factor suggests that you can determine Z, for example, by measuring C
and v.f. accurately. Some line configurations let you accurately
measure conductor diameters as well. You end up with lots of ways to
determine a set of line parameters.)

But note that a 50 ohm air dielectric coax using the same outer
conductor diameter would have a larger inner conductor, but MORE loss
than the 76 ohm air dielectric line because of the higher capacitance.
Quantitatively, it will have about 1.1 times the dB/unit length loss
compared with the 76 ohm line...so the difference in loss between air
inslated 50 ohm line and solid polyethylene dielectric 50 ohm line
(same OD) will be a ratio of 1.1:1.52, or 1:1.38. Going from 50 ohm
line insulated with solid pe to 50 ohm line of the same OD with air
insulation will cut the dB loss by about 27%. Going from solid to
foamed pe will get you about half that much. There's a bigger effect
going from a solid pe 75 ohm line to an air dielectric 75 ohm line,
cutting the dB loss by over 42% (assuming I didn't screw up the calcs
too badly).

Cheers,
Tom