“Despite an hours-long search Tuesday night, the bag, containing a fake bomb complete with wires, a detonator and a clock, made it onto an Amsterdam-bound flight”
Its interesting to think about screening in the light of the book I’m reading at the moment – Reckoning with Risk, Learning to live with Uncertainty. The book describes the problems with screening programs, and how the communication of probabilities can be made such that the risks involved in such programs can be understood by those deciding to participate.
Firstly there is an interesting statistic that none of the physicians questioned take part in screening programs, while they encourage their patients to participate. Then there is also significant evidence that the screening programs for both breast cancer and prostate cancer cause as many deaths as lives they save, giving them a net-benefit in terms of life expectancy of zero.
Now, applying the logic from the book onto a screening program for bombs on planes – what do we have? You can easily split probabilities down into four sections, true negatives, false negatives, false positives and true positives. Running any kind of screening process forces some kind of a balance between false positives and false negatives. In general (and certainly in this case) false negatives are considerably more costly than false positives, so the results will err toward saying ‘its a bomb’ when its not, rather than toward saying ‘its not a bomb’ when it is.
The interesting bit comes when you look at the actual number of bombs (true positives) compared to the number of false positives….I’m guessing that any airport is unlikely to find more than, say, one real bomb per week. Out of a passenger roll of, say 400 people per plane on 100 planes per day in 7 days (say 250,000 pieces of luggage). So if the false positive rate is, say, 1 in 1000 they’ll be looking at 251 pieces of luggage, of which 1 contains a real bomb.
What is the cost of this kind of investigation? What is the benefit? Of course, if the screening program were not ‘guaranteed’ to create false negatives it would be a lot easier to actually get bombs onto planes, so there would likely be more of them (or would there? is there a political solution to this kind of problem?)
2 comments