[extropy-chat] aargh

[Damien Broderick]

At 11:23 AM 12/18/2004 -0600, I wrote:

>Using pre-screened `star' Ss, Ertel claims a robust success rate averaging 
>70% where 50% +/- is expected by chance.

Sorry, I hadn't been awake long when I wrote that. Ertel's set-up contains 
balls numbered 1 through 5, so the chance likelihood is 0.2, not 0.5. Which 
is why his Ss' results are so impressive.

Here's the abstract of a recent poster (I hope the superscripts etc make it 
through the emailer):

Psi test feats achieved alone at home:
Do they disappear under lab control? [1]
                 Suitbert Ertel [2]

                 Abstract
Extraordinary hit rates from multiple choice tests, obtained by 
participants alone in their homes, are ambiguous. On the one hand, their 
feats might reflect psi power manifesting itself better under informal home 
than under lab conditions. Yet hit excesses obtained without control might 
also be due to negligent or fraudulent conduct. One way out of this dilemma 
is to let participants complete psi tests at home and to invite high 
scorers thereafter to do additional runs under lab control. This strategy 
has been endorsed using N = 238 (sample I) and N = 47 (sample II) of 
student participants. Sample I (female 84%) completed the ball drawing 
test, version I. Table tennis balls are drawn from an opaque bag on which 
numbers 1 to 5 are written, each number on ten balls. Participants guess 
and draw numbers blind and record their guessed and drawn numbers (hit 
expectancy 20%). One test unit consists of six or eight runs comprising 60 
trials each (total 360 or 480 trials). Participants shake the bag prior to 
each trial and put drawn balls back into the bag. Sixteen high scoring 
participants of sample I were also tested, using the same test, under lab 
control. Sample II (female 73%) completed the ball drawing test, version 
II. This test resembles test I except that green or red dots are sprinkled 
over the balls, participants guess numbers (five targets) and colours (two 
targets), the combined expectancy being 10%. Thirteen high scorers of 
sample II were also tested under lab control using a pearls drawing test 
where they draw one of five colours (no numbers, expectancy 20%). 
Hypotheses: 1. Hit rates of high scorers in home tests decline (due to less 
psi-conducive conditions under control and regression towards the mean). 2. 
Hit rates of high scorers under control score still significantly above 
chance (due to genuine psi which was also effective at home). Both 
hypotheses were confirmed with sample I and replicated with sample II. 
Sample I obtained an average ES = 0.369 (sd=.126) under home condition and 
an average ES = .122 (sd = .207) under control (ES = Z/root(Ntrials), the 
difference is significant (t (corr.smpls.) = -3.02, df = 15, p = .004, 
one-tailed). The significance of hit rates under control is p = <10-15 by 
Chi2 = 721.7, df = 16 (Chi2  = (Z)(Z) ). Sample II obtained an average ES = 
.275  (sd= .162) under home condition and an average ES = .098 (sd = .153) 
under control, the difference is significant (t = -3.00, df = 12, p 
=  .006, one-tailed). The significance of hit rates under control is p = 
<10-15 by Chi2 = 223.9, df = 13. Surprisingly, three participants obtained 
significantly higher hit rates under control compared with their home 
performance. The issue of fraud and bias loses relevance in view of such 
finding. It is recommended to introduce the “first-home-then-lab-test” 
strategy in parapsychological research on a broader scale. Once this 
strategy were generally applied, the widespread lamenting of psi 
researchers and their critics about tiny and elusive experimental psi 
effects might come to an end.


[1] Based on a poster presentation at the 47th Convention of the 
Parapsychological Association, August 5-8, 2004 Vienna, Austria.
[2] Georg-Elias-Müller-Institut für Psychologie, Göttingen, Germany.
Email: <mailto:sertel at uni-goettingen.de>sertel at uni-goettingen.de. Website: 
<http://www.suitbertertel.net/>www.SuitbertErtel.net