Elections are all that matter to them. And they would love to do away with that farce.
Originally Posted by John Shriver
We'll never be rid of career politicians until we make politics a less rewarding career.
And change the way we vote. We vote for all the wrong reasons. (character? ha!) The best people for the office get run out because of a lurid affair or something of no real importance while the 'pros' rob us blind.
Originally Posted by michaelbsc
The next time I'm holding out on my vote until they promise more film !!
What gets me totally, is the LAG restrictions they put in place. The amount we are permitted to carry on board is a great deal more than the shoe bomber (who got all this rolling) had on him - 100ml of liquid allowed that will fit in a small 8"x8" (approx) bag. Hmm, if I can fit 5 bottles in that bag that's 500ml of nitro - more than enough to blow a small hole in a plane methinks??
After all, they don't actually test any of the LAG we carry on now do they.
I assume that the airport screening test is a fairly simple one that performs some kind of measurement, and if the measurement exceeds a certain threshold value the result will be called "positive" and if it is below threshold it will be called "negative.
Nothing is perfect, so any type of test has a certain rate of false positives and false negatives. The rate of false positives and false negatives relates to how sensitive and specific the test is. (Actually, for convenience I am going to frame the discussion in terms of true positives and true negatives.) If the test catches all the true positive specimens and correctly labels them as "positive" then the test is highly sensitive. If it catches all the true negative specimens and correctly labels them as "negative" then the test is highly specific.
Generally speaking, a test can be tuned to make it more sensitive (it will always catch a positive) by lowering the decision threshold. (I am assuming that a high value represents a positive.) However, there is a trade off when you do this because the test then becomes less specific.
Conversely, if you set the decision point at a higher level you will never falsely classify something as negative (i.e. the test is specific) but you will miss a lot of the positives (you have lost sensitivity).
Thus, you can see there is a trade off between sensitivity and specificity. One of the tricks to testing is to pick the decision point so it best suits the purpose. Do you want to set the decision point low, so you catch most of the bomb-carrying passengers, but at the same time subject a lot of innocent people patdowns, or do you want to set the decision point high, so that very few innocent people are subjected to patdowns, but you let more of the bomb-carrying passenger through undetected. It presents a dilemma, especially if the quality of the test is not very high.
The quality of the test can be summarized in a graph called a "receiver operator characteristic", also called an ROC curve. This concept was formulated during war time to characterize the quality of radar operators in correctly classifying blips on radar.
Nowadays ROC curves are use a lot in characterizing the quality of clinical tests. An example of a test with a favorable ROC curve is troponin for diagnosing heart attack (myocardial infarction). An example of a test with a rather poor ROC curve is PSA for diagnosing prostate cancer.
One could go into other concepts like positive predictive value, negative predictive value, pretest odds, post test odds, likelyhood ratios, Bayes' theorem, and a lot of other related concepts, but it kind of makes my head hurt, even though I am supposed to know that stuff.
I just thought you might want to know that.
Last edited by alanrockwood; 02-17-2011 at 11:13 PM. Click to view previous post history.