This is an excellent thread. As a physicist, my thoughts/experiences are:
(1) Different speaker cables can affect bass performance a little, because bass drops off slightly if the cable is more resistive (most of the resistance is in the speaker, of course). Typically a speaker cable run is so short that it makes little or no discernible difference. On the other hand, with a 100 foot cable run most people would be able to hear the difference in bass between different manufacturer's cables.
(2) A looped cable, or a coiled cable, will make a difference to low frequency response as a loop or coil is inductive. If you don't lay your cables straight, that will make more of a difference in scientific terms than any internal 'inductance per unit length' characteristic of a cable. If conducting a DBT, all cables should be the same length and laid straight. A speaker cable should never be tested by passing the signal through the whole coil.
(3) With a digital signal, the cable quality should make no difference whatsoever until it reaches a point where there is a high error rate. That would need several miles of cable. If you think about a DSL internet connection (typically 2-3 miles of narrow-gauge copper wire between you and the nearest router, at 500kbps), you can see immediately that this point is correct. Obviously there is some error connection with internet transmissions, in the same way that there is some error connection on CDs and SACDs, so ultimately the user may not even know about errors unless they are so bad that the system does not work at all.
(4) With all cables, including digital, the terminations matter much more than the cable itself. A poorly terminated cable can lead to internal reflections (just like a piece of glass, or a mirror - have you ever stood close to a mirror and looked diagonally into it - you can count 5, maybe more, faint 'ghost' reflections because of internal reflections in the glass). Typically an expensive cable will be well terminated. A $20 cable should also be well terminated, however.
(5) Cable binding posts/banana plugs can have a surface layer of corrosion / dirt. (Even gold ones - the gold is not usually 24 carat.) If you change your cables over for a test, that will improve your electrical connection (or could make it worse) - either by scratching off some corrosion / dirt (improvement), or putting the new cable on over a place that has been more exposed to the air (worse).
That is why most people will notice an 'immediate improvement' if they upgrade their cables. They would also notice an 'immediate improvement' if they replaced their cables with a new version of the same one they already had.
(6) A very old cable can be oxidised internally, and can also have worn/broken strands where it has been bent in the past. 'Upgrading' to a new cable would have the same effect on the listener as in (5) above. It is not really because the cable is better.
(7) No speaker cables are shielded (OK, OK, some high end ones are sheielded, and braided/twisted cables have a shielded effect.) Although a speaker cable may pick up some RF interference, which may perhaps affect your amplifier's output stages by a microscopic amount, the point is that all unshielded unbraided speaker cables of the same length will be almost equal in this respect. The price of the cable will NOT affect how much RF interference it picks up.
(8) Cable quality can make a small difference to interconnects (where shielding from RF interference is important). Nevertheless it would be better to used balanced interconnects than to spend more than $20 on an unbalanced cable. Even with a $20 interconnect, the good quality termination of the cable is the single most important aspect (see above).
(9) Cable quality (or more specifically, termination quality) is very relevant to video cables. Internal reflections within the video cable can be of a speed/time delay that makes 1/2 pixel or 1 pixel difference, if the cable is several feet long. I recently bought a $100 VGA cable from Bettercables.com and it made a real difference to my projector-based system - previously with a standard cable I had several internal reflections leading to 'ghost images' of pixels, and the new cable reduced ghosting to about 25% of what was there before. Anybody who comes to my apartment can see that with their own eyes.
This time-delay effect mentioned here is not relevant at audio frequencies, which are about 1000 times slower than the frequencies of the signals carried by a video cable.
(10) Green ink can make no difference to the audio quality of a CD. Cleaning a CD (probably a pre-requisite before you apply the $50 green ink) can make an audible difference.