Climbers over the years have moved from the bowline to the figure 8 as the preferred knot because of documented cases of unsecured bowlines untying. For survival applications, I think it is more efficient and effective to teach one knot and then show how its variations can join two knots together, can form a loop at either the end or the middle of the rope, etc.

One important variable is the type of rope involved. In non-synthetic ropes, the bowline is apparently more effective, but who uses hemp or manila anymore? A very common rope is yellow poly(propylene?/ethylene). Whatever type of plastic comprises this rope, it is pure garbage, and it holds most conventional knots rather poorly. Also, are we using knots to lash up a shelter or to cobble up some sort of emergency rappel? Knots good for one application are death traps for another. This information is just as important as the knots themselves.