A Study of Experimenter Effects in Psi Research
One of the great challenges for twenty-first century science is our ability to grapple with differences: in worldview, in belief systems, and in ways of engaging reality. Although science offers a standard approach to understanding the world and our place in it, many assumptions that are made may or may not be justified. One of them is the issue of objectivity. Is it true that researchers can approach their work without some form of bias? And do the methods of randomized, double-blind procedures (the gold standard of good science) effectively rule out these biases? What if investigators hold different assumptions about the plausibility and potential outcomes of their studies? How might these impact the findings and what else might be learned about the nature of reality?
Over the past decade, IONS researchers have been addressing this issue of experimenter effects in the context of research on psychic or psi phenomena. I, a proponent who has obtained successful results in my research, have teamed up with Richard Wiseman, a skeptic, who has not found evidence for psi effects in his studies. This research has involved jointly conducted experiments exploring the possible existence of a commonly reported phenomenon: the "sense of being stared at."
Surveys suggest that between 70 and 90 percent of people have experienced an uneasy feeling of being stared at, turning around, and discovering that somebody was looking at them. Research into this phenomenon has a long and distinguished history, with initial papers on the topic being published around the turn of the last century by two pioneers of modern psychology: E. B. Titchener (1898) and J. E. Coover (1913). The first experimental investigation into the phenomenon was conducted by Coover at Stanford University. It involved an experimenter sitting behind participants, either staring directly at their backs or looking away, and then asking them to decide whether they had just been stared at.
What if investigators hold different assumptions about the plausibility and potential outcomes of their studies?
Subsequent work has involved increasingly sophisticated methodological and statistical procedures. For example, researchers have minimized potential experimenter-participant sensory cues by employing one-way mirrors and closed-circuit television systems, and have created a more sensitive dependent measure of participant's arousal by recording their electrodermal activity (EDA), rather than asking for a vocal confirmation.
Present-day procedure places the participant and experimenter in two separate, sensory-isolated rooms. A closed-circuit television system feeds a live image of the participant to a monitor in the experimenter's room, and at randomly determined times, the experimenter either stares at this image with intention to physiologically arouse the participant ("stare" trials) or looks away from the monitor and disengages intention ("no-stare" trials). The participant's EDA is continuously recorded during the experiment, and any significant differences observed in EDA between stare and no-stare are inferred to reflect the existence of psi. A recent meta-analytic review of 15 experiments using these types of procedures revealed a small, but statistically significant, overall effect (Schmidt, Schneider, Utts and Walach, British Journal of Psychology, 2004).
The first two collaborations between Richard and me generated evidence of "experimenter effects"; I obtained positive results in my experiments, while Richard found chance results in his. Such effects are common in parapsychology studies, and many researchers have argued that it is vitally important to establish why they occur, both in terms of assessing past research and attempting to replicate studies in the future.
To better understand the nature of these differences, we designed a third study. In the previous work, the experimenters both interacted with their participants at the start of each session (explaining the purpose of the study, administering questionnaires, etc. ) and carried out the staring/no-staring. The differing outcomes, we theorized, might have been due to either the different ways the experimenters interacted with participants or how they performed their staring. The new study attempted to evaluate these hypotheses. In two of the conditions, either Richard or I met the participant at the start of the session (i. e. , acted as "greeter") and then the same person carried out the stare/no-stare trials (i. e. , acted as "sender"). In the other two conditions, the greeter would switch to the other participant to act as sender. If the previous results were due to the way in which we, the experimenters, interacted with the participants, one would predict a main effect of greeter. If, however, they were due to the way in which the stare/no-stare trials were conducted, one would expect a main effect of sender.
It turned out that the new study failed to replicate the previous findings, yielding no evidence of psi in the data for either experimenter. Such a finding makes it difficult to identify specific reasons for experimenter differences in previous studies, but this is only one attempt. The work does present a model that explores the benefits of conducting collaborative work for resolving disagreements in controversial areas of science and speaks to the need for more research on this fascinating topic.
Marilyn Schlitz, Ph.D. is Director of Research at
the Institute of Noetic Sciences and Senior Scientist at the
Complementary Medicine Research Institute at the California Pacific
Medical Center. She has published numerous articles on psi research and
psychophysiology, cross cultural healing, consciousness studies, and
creativity.
