Backwards (pseudo)science

Most people have probably heard claims of hidden messages in popular songs that are revealed when the sound is played in reverse. I suspect also that most see these things as nothing more than amusing coincidences.

Today I came across the website of a man, David John Oates, who has created an entire business around this idea. He claims things as outlandish as that children learn to speak in reverse when babbling before they learn to speak forward, and that the sound of a lie, when played backwards, will reveal the truth. He ties all this to the hypothesis that human speech contains two components, a conscious forward part and a subconscious reverse part. He claims that this supposedly revolutionary idea is useful for lie detection in criminal trials, selection of the best business strategy, and a bunch of other things, and offers to listen to a 5 minute tape backwards for $125 in order to pick out any subliminal messages.

One piece of "evidence" he uses to back up his theory is a study of recordings of subjects made to read identical passages, which were analyzed to determine if the phrases appearing in reverse were the same for all subjects. Also, the number of reversals for people reading an emotional and neutral passage were compared. The data analysis for this experiment is a brilliant example of bad science. In particular, the authors reason that if the reversals were coincidental, they would be the same for all subjects reading the same phrase forward. Any scientist knows this is a ridiculous assertion--coincidental results show MORE variability than ones that are indicative of some underlying phenomenon. In addition, the authors expected to find more hidden messages in the speech of the people reading the emotional passages, while in fact they found the opposite. Most importantly, any use of statistical tests of significance is totally omitted from the analysis. As there were only two subjects reading each passage, using such tests would inevitably reveal the results to be insignificant.

I seriously wonder who would pay $125 or more to have someone listen for these messages in speech, but seeing the non-trivial popularity of psychics maybe one shouldn't be so surprised.