Because it is causing the signal to see a change in the dielectrics it passes through as it travels from a to b, most of the time solder resist/air, when it passes through some silkscreen ink you have solder resist/silk screen ink/air.
I imagine the question is, "how much". A professor will insist
there will be "some". Someone doing the testing may say
"so what, it's in the noise". Numbers matter.
Silk screens are usually placed over the solder resist and that gives some space. If the wavelength becomes comparable to this scale (less than 1mm) then there may be some observable effects. By the way, 30GHz approx corresponds to 1 cm wavelength....