BEagle: So if two nearby VORs deviate by several degrees (I've seen 3-5 degrees difference at least), then their local variation differs by that much?
Not quite, variation occurs constantly, but the VOR isn't "adjusted" except when it's out of alignment by a significant amount. So, two "nearby" VORs, one installed in 1950, is off in one direction by 2 degrees, the other VOR, installed in 1990 is off in the opposite direction by 2 degrees. Both are within acceptable differences, however between them the difference is 4 degrees. If they were both installed at the same time, they'd probably both be off by the same amount. Unless, of course, one was "adjusted" because it was .1 degree over the limit while the other was still just barely within the limit.
There is one other aspect to why there are differences between the "stated" angle and the "measured" angle, and that has to do with the surface of the earth being a sphere, and when you flatten it to make a map, the lines aren't straight, but when a map is drawn, the lines are straight. This can mean that when you measure an angle on a map, you get a slightly different value than if you were to put a compass on the ground and measure it.