Monday, May 30, 2011
In the book Debating Psychic Experience, "moderate skeptic" Chris French repeats a skeptical fallacy common among professional "skeptics", namely: that the Milton/Wiseman 1999 study failed to replicate Bem and Honorton's 1994 meta-analysis. In his own words: "Although Bem and Honorton’s (1994) original meta-analysis of 11 ganzfeld studies appeared to provide strong evidence of a replicable anomalous cognition effect, Milton and Wiseman’s analysis did not (although it should be noted that some commentators have argued that this is because many of the more recent studies were process-oriented rather than proof-oriented; Bem, Palmer, & Broughton, 2001; Storm, 2000; Storm & Ertel, 2001). (p. 56. Emphasis in blue added)".
This clearly shows that French is not a "moderate skeptic" at all, but a strong anti-psi believer (otherwise, how the hell can we explain that he continues to repeat an objection against parapsychology that is demostrably false? Is it an example of "moderate skepticism"?). Critical thinking, rationality, honesty and objectivity demand to correct our own opinions when the evidence shows they're false or misleading.
French is clever to affect the position of being "moderate" in order to avoid accusations of dogmatism. Affecting to play the impartial observer (making concessions against skeptics and making positive and sympathetic comments regarding to parapsychologists) has the rhetorical adventage of looking as a non-committed individual whose only purpose is to find the truth, whatever it leads. The actual purpose is to get a high credibility in the eyes of naive or ignorant readers.
But you can see French's actual (non-apparent) position in the way in which deals with the most important facts regarding the case for the existence of psi. If these facts are misrepresented, you can be sure the individual in question is motivated by an agenda.
In reply to French's misleading statement, Dean Radin wrote: "But is it really true that the Milton/Wiseman (M/W) meta-analysis failed to replicate the Bem/Honorton (B/H) outcome? The answer is no, it is not true. As I have pointed out (Radin, 2006, p. 118), and later confirmed by statistician Jessica Utts in a conference presentation attended by both Richard Wiseman and Ray Hyman, when the M/W database is evaluated using the same method as B/H (i.e., as a simple hit/miss statistic) it results in a significantly positive outcome. The reason the M/W meta-analysis purportedly failed is because the authors used an unweighted statistic that did not take into account each study’s sample size. If they had performed the correct analysis, M/W would have reached a conclusion that was diametrically opposed to the “failure” trumpeted in the title of their paper. Unfortunately, the skeptical mythos has uncritically adopted the wrong conclusion, and as such this may become an instance where myth is more comfortable than reality, and so the fictional story sticks." (p.114)
Why didn't French mention these problems in the Milton/Wiseman's analysis? Simple: Because mentioning that flaw would cast doubts in the conclusion of that study and therefore it couldn't be used anymore to favor the skeptical case against psi.
In order to give credibility and plausibility to his own skeptical position against the scientific replicability of psi, French is forced to intentionally conceals or disregards the key facts that would destroy his position.
In fact, the rhetorical trick is even more effective when French adds an "although" apparently in favor of parapsychologists (he says: "although it should be noted that some commentators have argued that this is because many of the more recent studies were process-oriented rather than proof-oriented) But note that this "althought" is intended as a red herring in order to look (in front of naive readers) as objective and impartial: the crucial, key and essentially relevant facts about the unweighted statistic used by Milton/Wiseman study that did not take into account each study’s sample size (and hence, which were responsible for the misleading conclusion) are never mentioned.
Explaining and expanding Radin's point, Chris Carter (in his updated review of Wiseman's research), comments in more detail the technical flaws of the Milton & Wiseman study: "The 30 studies that Milton and Wiseman considered ranged in size from 4 trials to 100, but they used a statistical method that simply ignored sample size (N). For instance, say we have 3 studies, two with N = 8, 2 hits (25%), and a third with N = 60, 21 hits (35%). If we ignore sample size, then the unweighted average percentage of hits is only 28%; but the combined average of all the hits is just under 33%. This, in simplest terms, is the mistake they made.
Had they simply added up the hits and misses and then performed a simple one-tailed t-test, they would have found results significant at the 5% level. Had they performed the exact binomial test, the results would have been significant at less than the 4% level, with odds against chance of 26 to 1. Statistician Jessica Utts pointed this out at a meeting Dean Radin held in Vancouver in 2007, in which he invited parapsychologists and skeptics to come together and present to other interested (invited) scientists. Richard Wiseman was present at this meeting, and was able to offer no justification for his botched statistics.
And this was not the only problem with the study. Milton and Wiseman did not include a large and highly successful study by Kathy Dalton (1997) due to an arbitrary cut-off date, even though it was published almost two years before Milton and Wiseman’s paper; had been widely discussed among parapsychologists; was part of a doctoral dissertation at Julie Milton’s university; and was presented at a conference chaired by Wiseman two years before Milton and Wiseman published their paper.
Here we have a case in which Wiseman nullified a positive result by first engaging in “retrospective data selection” - arbitrarily excluding a highly successful study - and then, by botching the statistical analysis of the remaining data."
I ask the objective readers of this blog: Do you think that French is showing a "moderate skepticism" when part of his skeptical case rest on such crucial factual omissions (against parapsychology) concelead with a language of moderation, impartiality and objectivity?
I let you to make your own mind.
This clearly shows that French is not a "moderate skeptic" at all, but a strong anti-psi believer (otherwise, how the hell can we explain that he continues to repeat an objection against parapsychology that is demostrably false? Is it an example of "moderate skepticism"?). Critical thinking, rationality, honesty and objectivity demand to correct our own opinions when the evidence shows they're false or misleading.
French is clever to affect the position of being "moderate" in order to avoid accusations of dogmatism. Affecting to play the impartial observer (making concessions against skeptics and making positive and sympathetic comments regarding to parapsychologists) has the rhetorical adventage of looking as a non-committed individual whose only purpose is to find the truth, whatever it leads. The actual purpose is to get a high credibility in the eyes of naive or ignorant readers.
But you can see French's actual (non-apparent) position in the way in which deals with the most important facts regarding the case for the existence of psi. If these facts are misrepresented, you can be sure the individual in question is motivated by an agenda.
In reply to French's misleading statement, Dean Radin wrote: "But is it really true that the Milton/Wiseman (M/W) meta-analysis failed to replicate the Bem/Honorton (B/H) outcome? The answer is no, it is not true. As I have pointed out (Radin, 2006, p. 118), and later confirmed by statistician Jessica Utts in a conference presentation attended by both Richard Wiseman and Ray Hyman, when the M/W database is evaluated using the same method as B/H (i.e., as a simple hit/miss statistic) it results in a significantly positive outcome. The reason the M/W meta-analysis purportedly failed is because the authors used an unweighted statistic that did not take into account each study’s sample size. If they had performed the correct analysis, M/W would have reached a conclusion that was diametrically opposed to the “failure” trumpeted in the title of their paper. Unfortunately, the skeptical mythos has uncritically adopted the wrong conclusion, and as such this may become an instance where myth is more comfortable than reality, and so the fictional story sticks." (p.114)
Why didn't French mention these problems in the Milton/Wiseman's analysis? Simple: Because mentioning that flaw would cast doubts in the conclusion of that study and therefore it couldn't be used anymore to favor the skeptical case against psi.
In order to give credibility and plausibility to his own skeptical position against the scientific replicability of psi, French is forced to intentionally conceals or disregards the key facts that would destroy his position.
In fact, the rhetorical trick is even more effective when French adds an "although" apparently in favor of parapsychologists (he says: "although it should be noted that some commentators have argued that this is because many of the more recent studies were process-oriented rather than proof-oriented) But note that this "althought" is intended as a red herring in order to look (in front of naive readers) as objective and impartial: the crucial, key and essentially relevant facts about the unweighted statistic used by Milton/Wiseman study that did not take into account each study’s sample size (and hence, which were responsible for the misleading conclusion) are never mentioned.
Explaining and expanding Radin's point, Chris Carter (in his updated review of Wiseman's research), comments in more detail the technical flaws of the Milton & Wiseman study: "The 30 studies that Milton and Wiseman considered ranged in size from 4 trials to 100, but they used a statistical method that simply ignored sample size (N). For instance, say we have 3 studies, two with N = 8, 2 hits (25%), and a third with N = 60, 21 hits (35%). If we ignore sample size, then the unweighted average percentage of hits is only 28%; but the combined average of all the hits is just under 33%. This, in simplest terms, is the mistake they made.
Had they simply added up the hits and misses and then performed a simple one-tailed t-test, they would have found results significant at the 5% level. Had they performed the exact binomial test, the results would have been significant at less than the 4% level, with odds against chance of 26 to 1. Statistician Jessica Utts pointed this out at a meeting Dean Radin held in Vancouver in 2007, in which he invited parapsychologists and skeptics to come together and present to other interested (invited) scientists. Richard Wiseman was present at this meeting, and was able to offer no justification for his botched statistics.
And this was not the only problem with the study. Milton and Wiseman did not include a large and highly successful study by Kathy Dalton (1997) due to an arbitrary cut-off date, even though it was published almost two years before Milton and Wiseman’s paper; had been widely discussed among parapsychologists; was part of a doctoral dissertation at Julie Milton’s university; and was presented at a conference chaired by Wiseman two years before Milton and Wiseman published their paper.
Here we have a case in which Wiseman nullified a positive result by first engaging in “retrospective data selection” - arbitrarily excluding a highly successful study - and then, by botching the statistical analysis of the remaining data."
I ask the objective readers of this blog: Do you think that French is showing a "moderate skepticism" when part of his skeptical case rest on such crucial factual omissions (against parapsychology) concelead with a language of moderation, impartiality and objectivity?
I let you to make your own mind.
No comments:
Post a Comment