• Threads: 124
  • Posts: 721
Joined: Jan 20, 2010
November 22nd, 2010 at 8:55:26 AM permalink
relevant article:

I am unpersuaded by the Cornell University experiment to correctly guess the location of an image because the sample size (1,000 volunteers) is too small. Yes, it's amusing that the subjects fared slightly better when the images were pornographic (49.8% success rate for non-porn vs. 53.1% success for porn) but tossing a coin could yield similar results.

However, the following anecdote caught my attention:

"The [1931] study was led by [Duke University] psychologist Joseph Banks Rhine and revolved around the Zener deck, a special set of cards featuring five different symbols. The test itself is straightforward: A card is drawn from the deck and the subject is asked to guess the symbol. While most of Rhine’s subjects performed in the neighborhood of random chance – they guessed about 20 percent of the cards correctly – [Duke undergraduate] Adam Linzmayer averaged nearly fifty percent during his initial sessions. Furthermore, these “guesses” led to several uncanny streaks, such as when he correctly guessed nine cards in a row. The odds of this happening by chance are about one in two million. Linzmayer did it three times... At first, Rhine simply assumed that Linzmayer was cheating, and that he’d had found some way to peek at the cards. This hunch was quickly disproven, however, as Linzmayer’s guesses far exceeded random chance even when facing away from the deck. And so, over the next few weeks, the scientist continued testing the student, repeating the same tedious task under slightly different conditions. (During a three-day binge before finals, Linzmayer guessed more than 15,000 cards, with a 40 percent accuracy rate.) According to Rhine, the most astonishing demonstration of Linzmayer’s talent occurred in his car, after he’d taken the student for a relaxing drive. It was here, in this setting of “easy informality” – the idling engine had a soothing effect – that Linzmayer correctly guessed fifteen cards in a row."

A 40 percent accuracy rate from 15,000 cards? Not bad! Nevertheless, Linzmayer's magic streak didn't last, and in the long run (2 years and 50,000 cards) he averaged 22.4 percent: about what you'd expect with random luck.
  • Jump to: