Terence Hines - Pseudoscience and the Paranormal

Some years ago several of my students and I used a more sensitive method to examine claims for psi. This method uses reaction time as the major dependent variable, or measure. To date, the vast majority of studies of psi have used accuracy as their dependent measure, as can be seen from the studies described above or in any review of modern psi research (e.g., Akers 1984; Morris 1978, 1982; Palmer 1978, 1982). In cognitive psychology, which is the experimental study of human learning, memory, and higher mental processes, accuracy is rarely used as the sole dependent measure. The dependent measure of choice is reaction time. This is because reaction time is a much more sensitive indicator of cognitive processes and processing than is accuracy. In other words, phenomena that are easily shown using a reaction time measure are not revealed by an accuracy measure.

An example will make the difference between the two measures clear. In a lexical decision task, subjects see, in each of several hundred trials, a string of letters. The string either is a real word—such as queen —or is not a real word—such as rarden . The subjects’ task is simple: Press one button if the letter string is a word, and another button if the letter string is not a word. Reaction time and accuracy are recorded. Consider the results in two situations. In the first, the subject sees the letter string kind on one trial, makes a response, sees queen on the next trial, and again responds. In the second situation, the letter string in the first trial is king and in the next, queen. The reaction time to respond that queen is a word is faster by about 50 to 100 milliseconds in the second situation (Shoben 1982). Thus, having processed one word speeds up subsequent processing of words that are associated with it in meaning. This is a highly replicable result and has served as the basis for literally hundreds of published studies of reading, semantic memory effects of brain damage on cognition, and so forth. The important point here is that accuracy measures do not reveal the effect of having just processed an associated word (Shoben 1982). This is a common result in studies of cognition—accuracy measures are not precise enough to demonstrate the often small, but very important, effects studied.

The implications of this analysis for psi should be obvious. Perhaps the repeated failure to find evidence for psi is due to the almost universal use of a dependent measure that is just too crude to find the looked-for effects. With this in mind, my students and I (Hines and Dennison 1988; Hines, Lang, and Seroussi 1987) conducted a series of studies of psi using reaction time as a dependent measure.

One study (Hines and Dennison 1988) followed naturally from Schmidt’s work described above. On each trial a computer generated, randomly, a 0 or a 1. In the ESP condition, this digit was stored in the computer’s memory. The computer then generated another random 0 or 1 and displayed that digit on a television monitor in front of the subject. The subject had to decide, as quickly as possible, if the digit on the screen was the same as the digit stored in the computer’s memory or not. Reaction time, measured from the onset of the digit on the monitor to the time the subject responded, was recorded, as was the accuracy of the subject’s responses.

Even if ESP is a real phenomenon, the accuracy measure would be unlikely to demonstrate it. Previous studies have failed to provide any adequate evidence for ESP when accuracy measures have been used. In line with this prediction, our subjects performed no better than chance when the accuracy measure was used. The real test, however, was the much more sensitive reaction time measure. If ESP exists, subjects should respond more quickly when they are correct than when they are incorrect. Whatever the source of the hypothesized extrasensory information, if it were getting to the brain, where it would have to be processed to have any effect at all on behavior, it would speed up the process of making a correct decision. However, we found no difference between correct and incorrect reaction times that even approached statistical significance, nor did providing feedback to the subjects regarding their accuracy on each trial change these results. Reaction times of correct and incorrect responses remained the same.

We also examined a precognition condition, again with and without feedback to subjects on their accuracy. In the precognition condition, the computer generated a random 0 or 1 and displayed it on the monitor. The subject then had to decide if the digit the computer would generate after he or she responded would be the same as or different from the one on the screen. After the subject responded, the computer generated another 0 or 1. As in the clairvoyance condition, subjects were no more accurate than expected by chance, nor were they any faster when their responses—predictions, in this condition—were correct than when they were incorrect. Again, the presence or absence of feedback to the subject did not alter this pattern.

In another study, Hines, Lang, and Seroussi (1987) adapted the lexical decision task described above to the study of ESP. There were two conditions. In both the control and the ESP conditions, a subject made a series of lexical decisions on letter strings presented on the right side of a television monitor controlled by a computer. In the control condition, the lone subject had no companion. In the ESP condition, there was a second subject who made decisions about letter strings presented on the left side of the television monitor. The two subjects sat on opposite sides of a divider that prevented either from seeing the other or the other side of the monitor. In the ESP condition the subject who sat on the left side of the divider—the agent or “sender”—saw a letter string, and 400 milliseconds later the receiver—the subject on the right side of the screen—saw a letter string. Both subjects had to make lexical decisions on the letter strings they saw. On some of the trials in the ESP condition, the letter string the agent saw was the same one the receiver saw, 400 milliseconds later. In the standard lexical decision task, if the same subject sees the same letter string twice in a row, reaction time on the second presentation is greatly reduced. This is due to the fact that the location in memory that stores the concept, if the string is a word, is activated by the first presentation and when the same word occurs again, that location is much more able or ready to be activated again. This activation decays after a few seconds. If, then, there is extrasensory communication between individuals even of a very low level, it should show up in the reaction times of the receiver when the agent had just processed, 400 milliseconds previously, the same letter string that the receiver was processing. We found no hint of any such effect in our data.

Failure to find an effect is not conclusive evidence that the effect is not real. However, the fact that several of our studies using a highly sensitive reaction time measure have consistently failed to show any evidence for psi seems to be convincing prima facie evidence against the reality of the phenomena.

Psi phenomena of one sort or another have been systematically studied for more than one hundred years. This century of investigation shows a common, repeating pattern. A new phenomenon or method is found that “finally shows that psi is real,” the skeptic is told. However, upon careful examination, the claim collapses. But by the time one claim has been carefully investigated and found faulty, another phenomenon or method has come along that, “once and for all, shows psi to be real.” And on it goes. First there was spiritualism, then card-guessing experiments, then individual psychics like Geller, then the ganzfeld studies, and now the random events studies of Schmidt and others.