I'm not a professional when it comes to these things, but in general, I'd keep in mind what someone at El Al once said about our practices: "In America, you search for weapons. In Israel, we search for
terrorists." Their approach seems to have served them well for the most part. As such, I'd go heavier on the profiling side (not just ethnically, either, but on all aspects of a scientifically valid profile), and probably go a little easier on weapons in general. If it means that a terrorist slips by with a weapon, it would also mean that the other passengers are armed as well, and so his effectiveness would be, shall we say, diminished.
But like I said, I'm not an expert.
As such, I'd go heavier on the profiling side (not just ethnically, either, but on all aspects of a scientifically valid profile), and probably go a little easier on weapons in general. If it means that a terrorist slips by with a weapon, it would also mean that the other passengers are armed as well, and so his effectiveness would be, shall we say, diminished. Why go easier on weapons in general? While I think the premises and conclusions are rather questionable, you can at least make an argument that armed passengers could prevent a takeover of the airplane. But that presumes that this is necessarily the goal of the suspect - if Richard Reid had his s*** together, instead of trying to light his shoes with a pack of matches, it's difficult to see how his fellow passengers could have prevented him from blowing up the plane, armed or no. In that case, the best solution would seem to be to prevent him from boarding in the first place.