From:    Jan Willem Nienhuys                    
To:      All                                    Msg #41, Jan-01-93 05:56AM
Subject: Mars effect research (JWN)

Organization: Eindhoven University of Technology, The Netherlands
From: wsadjw@rw7.urc.tue.nl (Jan Willem Nienhuys)
Message-ID: <6775@tuegate.tue.nl>
Reply-To: wsadjw@urc.tue.nl
Newsgroups: sci.skeptic


This is an answer to a post of Ertel of mid-December.
Summary of Ertel's proposal: to test whether a certain group
of athletes displayed a `Mars effect' that is very striking or
just so-so, Ertel has proposed first to shift all athletes'
birthdays by a full year. I objected to this procedure as the
resulting distributions cannnot be counted as mutually independent.

Ertel then proposed that each athlete's birthday be coupled with
the birth time of a randomly chosen (other) athlete.  This method
was used before by Gauquelin and the Comit\'e Para.

Here is my answer. 


Dear Professor Ertel,

I think that random recouplings, of the type Gauquelin used,
would randomize the data sufficiently. I do very well remember
the Para procedure.  Actually I advised Francoise Gauquelin to
insert the passage on page 206 of the Euroskeptics Proceedings,
because I thought it was the best evidence that Michel had 
earnestly and convincingly tried to control for diurnal and 
annual rhythms.  The procedure has only one weak point.  It would
destroy any "synergistic" effects of the combination of diurnal
and annual birth rhythms.  This point you mention as a reason to 
favor your `Ertel preferred shifts'.   But this "synergistic" 
explanation is so far fetched that the burden of proof rests 
squarely on the shoulders of those who propose it. 

I cannot comment on your argument #2, namely that additional
astronomical variations would already provide enough randomization.
I haven't seen the data, nor the computations. The fact that
the 32 year cycle seemed to have mutated into a 31-year cycle 
worries me.  Even so, I still think that your method can't 
yield more than about 10 independent pseudo random samples,
and I think I have made clear that the burden of proof that
it is 50 rather than 10 now rests on you.

And by the way, I am still curious as to the Mars percentage of the 332 "Para
champions".  I can find out for myself (from the files you sent De Jager and
Koppeschaar), if I know their names.  

Jan Willem Nienhuys 
A happy New Year