Statistics reach beyond number-crunching

News Editor
March 26, 2004

Jessica Utts is looking for proof of the paranormal.
She doesn't turn to photographs of ghosts or communications with dead relatives for evidence -- she uses statistics.

"Using the standards applied to any other area of science [that uses statistics], it is concluded that psychic functioning has been well established," Utts wrote after serving as a statistics consultant in a classified government investigation of psychic phenomena.

Utts, a professor in the University of California at Davis' department of statistics, spoke yesterday afternoon in a seminar organized by Pitt's statistics department.

"Statistics really can't tell why" things happen, in any field, Utts said. By performing a probability test, she can use statistics to tell her whether the potentially psychic subjects are beating the odds.

In the experiments conducted by the government, which were declassified in 1995, hundreds of photographs were divided into packets of four dissimilar photos. A packet was randomly selected by a research assistant, meaning each packet had an equal and separate chance of being chosen, and a single photo from the packet was then selected.

In a separate room, the test subject was asked to draw the selected target -- in this case, the photo that had been selected. In this way, researchers tested each subject's clairvoyance, or ability to receive information from another place. The subject would receive no information, except that the target was a photo -- and, in some experiments, that it was a photo.

Since the subject's drawing could not be objectively analyzed using statistics, experimenters then gave the drawing to a judge. The judge would receive the original selected packet, including the selected photograph, and choose which photo had been selected, based on the subject's drawing.

According to random probability -- if the subjects' drawings were of no use to the judges, and the judges were randomly selecting the target -- the judge would select the correct photo about 25 percent of the time. According to the experiments, however, judges selected the correct photo more often than random probability would allow.

By comparing the predicted probability and the actual probability, Utts concluded that "there was something going on in the data."

"People are either totally convinced that it's real or totally convinced that it's not real," Utts said, adding that "it's so easy to lose credibility in this field."

Utts' use of statistics in the realm of extrasensory perception -- which is sometimes called psi, to remove the implication of sensory involvement -- illustrates one of many applications of statistics, according to Pitt statistics Professor Nancy Pfenning.

"I think there are a lot of issues in students' lives that can be understood through statistics," Pfenning said.

Students can find statistics more interesting, she said, "if they pay attention to all the types of problems that can be studied in statistics, instead of getting bogged down with crunching numbers."

Pfenning also pointed out that knowledge about statistics can help students understand which studies to trust, and which to consider with skepticism.

"If you're too naive, you might believe everything you read in the paper," she said.

Addressing Utts' research, Pfenning said that the experiment "deserves to be taken seriously."

"Statistics is not a black-and-white kind of subject, the way math is," she said. "We can't expect Jessica Utts to stand up there and prove [the existence of the paranormal].