Assume that those who would plan terror attacks have at least as much information as we have. They know that “it remains unclear whether the AIT would have detected the weapon used in the December 2009 incident“; that “only 2 percent of all passengers would have been picked at random to go through” these questionably-effectual scanners; and that “random” means people who look like they will readily comply or people who might look good naked. Also assume that the planners are more familiar with probability, statistics, risk and cost-benefit analyses than most Americans. A terror attack doesn’t cost a lot of money; with lots of Saudi cash supporting Al Qaeda, the limiting resource is probably people willing to blow themselves up.
Because terrorism’s weapon is terror, a failed attack is not a failure. The schmoe with the shoe bomb might frighten people and change their behavior almost as much if he can’t get the fuse lit as if he can.

So you’re the guy in a cave deciding whether to send the next hapless underwear bomber to try to take down a plane. Does the chance that your operative might (if he looks sufficiently submissive or hot to the behavioral scientist manning the checkpoint) be one of the three people on the plane to be selected for scanning, and that if he is selected the scanner might detect the explosives in his panties, affect your planning one whit?

I’ll leave it to the security experts like Bruce Schneier to tell us whether scope-or-grope will work to deter terrorist attacks. In my view, though, it shouldn’t.