Not all psi experiments are exclusively or even primarily conducted to simply prove the reality of psi. Most, in fact, have as their main objective to learn about the nature of psi and what other variables it relates to. One such question that has commanded a great deal of attention over the years is what kinds of people score best on psi tests, particulary tests of ESP. This question is often addressed by correlating scores on ESP tests with scores on paper-and-pencil tests of personality. One of the most common personality traits studued by this method is extraversion, which refers to how much a person is sociable and outward looking (extravert) or withdrawn and inward looking (introvert).
Honorton, Ferrari, and social psychologist Daryl Bem (1990) reported a meta-analysis of 60 extraversion-ESP experiments involving 17 different experimenters and 2963 subjects. The overall p-value was 4 x 10^-6, confirming what had previously seemed to be the case from less formal analyses, that extraverts score better on ESP tests than introverts. A blocking analysis showed that, once again, the effect was restricted to subjects tested individually. The authors also divided the studies into two other categories: those used a force-choice (FC) methodology (like card-guessing) where the number of response alternatives is limited, and those that used a free-response (FR) methodology where the targets are more complex and subjects are encouraged to report any imagery that comes to mind. Although the ESP-extraversion relationship was significant for both, it was stronger for FR.
A controversy has arisen over the interpretation of the effect for the FC studies. Through further blocking, Honorton discovered that this FC effect was entirely attributed to those studies where subjects had filled out the extraversion scale after they had learned their scores from the previously completed ESP-test. He hypothesized that this feedback influenced how they filled out the extraversion scale; ie, high-scoring subjects on the ESP test were thereby induced to fill out the scale differently than were the lowscoring subjects. If true, this means that the extraversion-ESP relationship is artifactual for FC experiments.
This conclution has been challenged by other parapsychologists, most notably K Ramakrishna Rao. He cites a series of four experiments conducted by H Kanthamani and Rao (1972) that were essentially identical except for the ordering factor. In the one experiment where the extraversion scale was given before the ESP test, the support for the extraversion-ESP hypothesis was significant and just as strong as in the other studies, where the order was the opposite. More recently, Shanti Krishna and Rao (1991) collected new data where the order of testing and the kind of ESP feedback subjects received were systematically manipulated. They found no evidence of bias. There also seem to be disagreements about how some studies should be classified. This controversy is by no means over, and it will be intresting to see how it turns out. There is no debate about the validity of the extraversion-ESP relationship for FR studies, where presence of the artifact is not related to the strength of the effect. Also there was no correlation with study quality.