Participants invited to join the three-week debate that was introduced in Part I (Milton, 1999) contributed 90 messages. These messages came to H. E. already numbered, so that he did not know the authors. H. E. first read the messages for courtesy. There was only one message that gave him pause, but in the spirit of open debate, this message was sent forward unchanged. Most of the problems encountered were practical ones, especially dealing with differences in formatting, with over half of the messages requiring reformatting before being sent out to the participants. Occasionally, there were additional technical problems which slowed down dispersing certain messages, and even assigning new numbers to them, but, in general, the process of H. E. checking the messages and redistributing them (by simply sending them back to the server) worked fairly well.
The edited messages thus came to G. S., numbered in the order in which they were received, except that numbers 2, 20-23, 32, and 45 did not correspond to new responses, and the messages listed here as 16a and 86a had not received a number.
Milton asked G. S. to edit the responses, putting them in intelligible order and keeping all substantive points while making them reasonably concise. G. S. decided on topics and subtopics, then listed under them in chronological order the relevant messages or parts of messages (each preceded by its number in boldface type). Parts of a single message often appear under different headings, either because the writer shifted to a new topic, or because a coherent argument cuts across the classifications.
The exact wording of the message was retained except for deletions and for a few trivial changes, like short insertions [in brackets] for clarity or using full terms instead of abbreviations. After the close of the debate, G. S. forwarded this edited version to those who had participated and restored a few parts that the participants asked her to keep.
Several responses make substantially the same point, and a reader may ask whether they came from one person repeating an argument or from different discussants who agreed with each other. The reason that the transcript seldom gives this information is that Milton required G. S. to omit the discussant's name (except when the message identified the discussant). The transcript is followed by an Appendix that identifies the author of each message.
Almost half the messages clustered around criticisms or defenses of Milton's conclusion (Milton, 1999) that the Milton and Wiseman meta-analysis of 30 recent ganzfeld studies failed to replicate the evidence for psi in earlier work. This is the first topic below, followed by criticisms or defenses of other points in the Milton and Wiseman paper and in Milton's "Discussion paper." Then come the two next most frequent topics: attempts to define the ganzfeld and the desirable direction for future ganzfeld research. Other topics, such as the value of the debate, follow these.
CRITIQUES AND DEFENSES OF MILTON'S CONCLUSION
Misuse of Meta-analysis as a Proof
#5, part 1. Meta-analysis is being misused by aiming specifically toward proof-of-existence of a phenomenon. The title of one of the papers forwarded for this debate begins with the inappropriate question, "Does psi exist?" Although there has been a widespread use of meta-analysis in various fields to argue such questions, its real power and value derive from the ability to usefully and quantitatively summarize large amounts of evidence. Meta-analysis should be used to learn something; it is not suitable for the simple-minded effort to prove something. Indeed, the latter effort is terribly vulnerable to contamination by the desire to prove something.
2. Meta-analysis must, in order to be useful, follow the intelligent dictum, "Concatenate widely and categorize wisely." This means to assemble, or fairly sample, all of the material that is pertinent to the topic, and then do an analysis that allows insight into the contributions of various factors. …