A Collection of Strange Beliefs, Amusing Deceptions, and Dangerous Delusions

From Abracadabra to Zombies

A Novel Way to Make an Ass of Yourself

Gary Schwartz Rides Again


Robert Todd Carroll

In a recent paper, Gary Schwartz once again tries to validate the work of mediums who claim to get messages from the dead while engaging in what the rest of the world knows as cold reading and bad science. The paper is called "Anomalous Information Reception by Research Mediums Demonstrated Using a Novel Triple-Blind Protocol" and is co-authored by his assistant, Julie Beischel, Ph.D. It was published in Explore: The Journal of Science and Healing (January 2007). After an opening paragraph reminding the reader of how popular the subject of ghosts is in the world of television shows and movies, he then evaluates the quality of his work: "This report illustrates," says Schwartz, "how rigorous and credible research can be conducted in the laboratory...." A fairer assessment would be: "This report illustrates how ridiculous and incredible research can be conducted by an ass and his assistant." Schwartz is right, though, that as far as television is concerned the paranormal is the norm.*

In this experiment, eight mediums are tested in a novel way. Eight students at the University of Arizona, where Schwartz practices his craft, were recruited for the study. They are referred to as "sitters" throughout the paper, though none of them actually sits for a reading. Four of them have experienced a dead parent and four have experienced a dead "peer." The eight mediums aren't named but they were chosen because they performed well for Schwartz (or somebody in some lab) in the past.

Schwartz avoids one common criticism of his past experiments: the mediums do not come in contact with the sitters. But the mediums, he says, feel comfortable with somebody on the other end of the phone line when they do their readings, so he incorporated "proxy sitters," namely himself and his assistant. The proxies ask questions of the mediums during the phone readings. What part of experimenters should not participate in their own experiments because they might bias the process doesn't he understand?

Schwartz writes that the proxies were "blinded to the identity of the sitters and to any information about the discarnates beyond their first names." (The discarnates he refers to are the alleged spirits of the dead parents or peers.) We'll have to trust him on this, though one wonders how they set up the experiment without having some idea of who went with what first name.

According to the paper:

Information about each discarnate and his/her relationship with the associated sitter was collected from the sitter participants [i.e., students] by a research assistant who did not interact with the mediums. Discarnate descriptions were then paired to optimize differences [emphasis added] in age, physical description, personality description, cause of death, and hobbies/activities of the discarnate. Four deceased parents were paired with four deceased peers of the same gender for a total of four pairs of sitters. It is important to note that this procedure (a) maintained rater blindness by pairing discarnates of the same gender, while (b) optimizing the ability of blinded raters to differentiate between two gender-matched readings during scoring.

Did you follow that? The mediums are not going to do eight readings each and the researchers are not going to then give the readings to the students to pick out the ones that they think match their discarnates. They're not going to give the students who lost parents 32 readings for them to select or rank for accuracy for their own parent. They're not going to give the students who experienced the loss of a peer 32 readings for them and ask them to select or rank them for accuracy. You might wonder why the discarnate descriptions were paired to optimize differences, except gender differences. Schwartz says they did so to optimize "the ability of blinded raters to differentiate between two gender-matched readings during scoring." Why would he want to help the raters in this way?

He also thinks that:

Each sitter in a pair acted as a matched control for the other sitter in the pair: each sitter scored the reading intended for him/her as well as the reading of the control sitter while remaining blinded to the origin of the readings.

Interesting concept: a student who has lost a parent is given a pair of readings to rate, one for his parent and one for a "peer" and the one for the peer is a "control." And a student who has experienced the loss of a peer is given a pair of readings to rate, one for a peer and one for somebody's parent; the reading for the parent is to be a "control." He writes that "each sitter scored the reading intended for him/her as well as the reading of the control sitter while remaining blinded to the origin of the readings." You could call it a "control." You could also call it a "gift" or a "lead." I say this because Schwartz tells us that they screened 1,600 students to get to the final eight for their beliefs regarding an afterlife and mediums. They were asked to answer "yes" or "unsure." I think it is safe to say that he did not exclude people who are sympathetic to his mission. The eight finalists probably want him to succeed and would be motivated to help him do that.*

In fact, the eight mediums didn't do eight readings, they only did two each: "one for each sitter in a pair." This might sound like the proxy sitters are standing in for a pair of students, one with a dead parent and one with a dead peer. Even Schwartz is not that demented. There were 16 readings, two each by each of the eight mediums. Each medium read one student who'd lost a parent and one who'd experienced the loss of a peer. Clear enough?

Schwartz writes that "an experimenter blind to details about the sitters or discarnates" transcribed the readings and created a numbered list of individual items for each reading. A scoring system was devised, based on a 6-point scale.

6.Excellent reading, including strong aspects of communication, and with essentially no incorrect information.

5.Good reading with relatively little incorrect information.

4.Good reading with some incorrect information.

3.Mixture of correct and incorrect information, but enough correct information to indicate that communication with the deceased occurred.

2.Some correct information, but not enough to suggest beyond chance that communication occurred.

1.Little correct information or communication.

0. No correct information or communication.

There is something obviously wrong with this scoring system being used by the students. They might well be asked to rate the accuracy of the information provided on a scale of one to six, but they are not trained to make judgments about the quality of the reading. Nor does Schwartz explain how these students are to know that a mixture of correct and incorrect information has enough correct information to indicate that communication with the deceased occurred.

In any case, the average score for the correct readings was 3.6 and the average score for the "controls" was 2.0.

Schwartz concludes: "The present findings provide evidence for anomalous information reception but do not directly address what parapsychological mechanisms are involved in that reception." Maybe these findings provided evidence for anomalous information reception, but they don't provide good or compelling evidence for anything. Why? His results are based on the ratings of eight subjects. If just two of them had been switched, the results would be 2.9 for the controls and 2.6 for the "real" readings. This study is too small to warrant drawing any conclusions about anomalous information transfer. Investigation into how information was received can wait until there is better evidence that there is anything to explain.

The only thing that needs explaining is why the University of Arizona continues to stand by Schwartz, whose work should be an embarrassment to the University.

October 15, 2007

This page was designed by Cristian Popa.