See, the problem with random searches of people in airports is that it’s wholly ineffective.  What are the odds of finding someone carrying an explosive in a carry-on that is undetectable by a the x-ray machine using random searches?  Well, if TSA searches 1 in 1000 people that come through security and a terrorist has cleverly disguised the explosive device as a laptop, cell phone, hair brush, camera, music box with a cute little dancing ballerina, or anything other than an easily recognizable tangled mess of wires and explosive putty, it’s practically 0.  It could be as high as me winning the lottery (I don’t buy tickets).

Even if you could search someone or something with a 100% chance of finding hidden objects, the odds of selecting the right person are minimal.  Statistics are full of things like “risks of over-reliance” and “representative samples” and “expected error rates” and other phrases that all boil down to the simple truth: You can’t sample for pinpoint precision. Yeah, that’s bold, italic and underlined.

You can sample with a certain degree of precision when attempting to find out how people across a nation are planning on voting, and sometimes Dewey does win; but polls are only as reliable as the pollsters, the sampling methodology and the number of people polled.   You can often predict the outcome of a two-outcome vote by sampling only 1000 people of any large population.  That is, if the poll shows that one is a clear winning beating the margin of error (+/- 5% at that sample size (and people answer truthfully and actually vote, etc.)).

But to get a degree of certainty that you would need to find a single dangerous device in a population of 1.8M (that’s how many people flew through O’Hare this holiday season, according to some random TV show I was watching while getting a pedicure), you’d have to sample …. math is hard.  But my understanding is that it’s most of them.

Yup, freedom gropes for million people.

Sampling is stupid.

What are we really looking for?  Weapons?  Can’t hijack a plane anymore; armor and locks on those cockpit doors solved that problem.  That was a no-brainer, BTW.  Hurrah for locks and armor on cockpit doors.  Intelligent security features.  Mythbusters has proven that a firearm won’t cause explosive decompression (someone used science!)  So weapons doesn’t mean the plane comes down.

Explosives?  Know what finds explosives?  Dogs.  Well trained dogs are far more amiable than an unwanted hand down your pants.  Know what else does?  Little white cotton swabs and some chemicals.  Yeah baby.  Explosives are smelly and dirty.  You are more likely to find an explosive with a dog or a swab for an explosive than by searching a bag in a 30 second search where you never even opened the other pocket of my backpack, TSA agent.

Now I know no-one is listening to my advice.  But in these days of concern over the economy and big-government, I have to ask, are these new measures cost effective?  Tons more TSA agents standing around chatting with other TSA agents (They’re working?  They look like they are just standing there chatting.)  Machines that can see your junk.

I’m sure a well-trained dog is way cheaper than a backscatter machine.  Heck, I’d pitch in for the white cotton swabs.

So, for those of you in charge of designing security for the TSA, how about a statistics class?  I’ve only had two, and I can figure it out.  And I wasn’t really paying attention in the first class.