I would like to make it perfectly clear that I hate, with a passion, talk shows. Why these people need to use television as a medium to air their laundry is beyond me. I know lots of people want to be on the TV, but I wouldn't want to share anything that personal, and generally anything that embarrassing/pathetic, with the world at large.
But that's not the aim of today's rant. No, today we focus on that Jeremy Kyle staple, the lie-detector (aka polygraph) test. There is much argument on the reliability of these tests, generally somewhere between 60% and 97% is quoted (however a great deal of this evidence is unreliable).
The physiological effects recorded by the polygraph are not unique to deception. Surely this should set off alarm bells?
Well no, these people on what seems to be a daily basis put major decisions in the hands of a machine which is at best 97% accurate. Kyle is guilty, as others are, of making this seem almost infallible.
What I do in a situation like this is turn the statistics around. 97% accurate means 3% inaccurate. For every 100 tests done, 3 (on average) will give the incorrect result (remember that this is the best case scenario!).
If you were a jury member would that be accurate enough for you to convict? If you were a judge would you accept that if the average accuracy of the evidence before you was 97% then for every 100 people you convicted 3 would have been denied their freedom for no reason at all. I hope not.
Innocent people will unfortunately be imprisoned under any system of law and order we have, as a species, yet envisioned. But for me the number of innocents convicted would have to be, as a minimum, somewhere around 1 in 10,000 (99.99% accuracy). If the evidence was any less convincing than that I would, personally, have to call that reasonable doubt.
But Kyle commits a worse sin. An inconclusive test is just that, it is impossible to draw a conclusion from it. Yet the daytime demigod, facing four possible thieves who had undergone the polygraph, three of which got an all clear and one of whom got an 'inconclusive', picked on that 'inconclusive' person as the culprit. An inconclusive result, even given the fact that the others pass the truth test, does not mean that person has failed to tell the truth, it means that the test has failed to discriminate.
Of course this would require Kyle, the juror/judge, the people on the show, etc. to have a sound basic knowledge of mathematics and scientific reasoning. Which given the fact we live in the Jeremy Kyle generation is sadly unlikely.