...PVC pipe and vinyl siding in very close proximity to you ladder line will cause some additional loss. These plastics have metal-based “ plasticizers” in them that I suspect cause the extra losses...
Actually, if you look at the physics involved, the feedline losses tend to be
more in the conductor resistance rather than the dielectric characteristics, at
least though the VHF range.
For example, foamed dielectric coax has lower loss than solid polyethylene, but
that is primarily because the conductors have to be larger to maintain the same
characteristic impedance when the dielectric constant of the insulation changes
due to the presence of more air in the mix. TV twinlead has relatively small
conductors compared to typical ladder line, which accounts for most of the
difference in loss, rather than the dielectric surrounding it.
While it is true that putting a J-pole inside a length of PVC pipe shifts the tuning,
in my experience the change is much less for a normal dipole. It isn't the antenna
itself that is detuned, but primarily the velocity factor of the matching stub
(especially if the antenna lays against the inside of the pipe rather than hanging
straight down the middle.) You can test this by inserting the radiator of the J-pole
inside a pipe, measuring how much the SWR changes, then pushing the rest of
the matching stub inside to see the difference.
Open wire line has a relatively high impedance (though it is possible to make
lower impedance lines using multiple wires). That helps to keep the impedance
high along the length, which minimizes the current, reducing resistive losses.
But when open wire line runs at a high SWR, so there are points of relatively
low impedance with high current, the losses can be quite higher than expected.
For example, using VK1OD's transmission line loss calculator here:http://vk1od.net/calc/tl/tllc.php
The losses in 20' of Wireman 551 on 80m (so a short line on a low frequency
where we expect feedline to be low) is about 0.85dB with a load impedance of
10-j200 ohms, and 1.5dB at 5-j200 ohms. In fact, this is a convenient tool to
look at the losses in partial wavelengths of feedlines, which don't always behave
as the textbook loss tables might suggest based on SWR. That's because the
tables assume the loss is averaged over the whole wavelength when operated
at a high SWR, while in practice the lower impedance portions will have the
highest losses, consistent with our observation that loss is more due to conductor
losses rather than dielectric losses.