I teach what to do based on the prevailing circumstances and try to get the student to think about the entire issue, not just memorize a procedure but actually think about the fact you can lose your engine anytime and anywhere when you are flying an airplane. The geography in the US varies so much I would say that the main criteria to consider is whether you have any options besides roads and airstrips at all. Landing on a busy highway is usually fairly low in the matrix of options, but if you have rocks and water on all sides, then maybe it's worth a try. In day VFR it's best to land in a field beside the highway, if there is one, because there are power lines and cars on the road itself. If the highway is one that is not very busy and it looks to be clear of power lines as some rural divided highways do, then go ahead and have that. Night is a whole other ball game, I would never land on a road at night because you stand a much better chance of wrapping the airplane up in the power lines. I would opt for dirt near a lighted area in that case.

One of the things I see is often missed is how to calculate gliding distance. The lift to drag ratio for any airplane can easily be used ot figure out how far you are going to go before hitting dirt. For example, in an airplane that has a 10 to 1 lift-to-drag ratio, and loses an engine at 10k ft. above ground, for every thousand feet of downward travel you get ten times that in forward travel. So, 100k ft. forward travel divided by 6,100 ft per nm gives about 16 miles of glide. This is a quick way to estimate what your options really are.