NetNews Usenet Archive 1992 #26/NN_1992_26.iso/spool/sci/skeptic/19454

Text File | 1992-11-12 | 12.5 KB | 274 lines

Newsgroups: sci.skeptic
Path: sparky!uunet!charon.amdahl.com!pacbell.com!sgiblab!darwin.sura.net!Sirius.dfn.de!gwdu03.gwdg.de!ibm.gwdg.de!SERTEL
From: [e-mail]
Subject: Mars effect
Message-ID: <[e-mail]>
Sender: [e-mail] (USENET News System)
Nntp-Posting-Host: ibm.gwdg.de
Organization: GWDG, Goettingen
Date: Thu, 12 Nov 92 17:44:26 MEZ
Lines: 262



       BACKSTAGE NEWS ABOUT THE EUROSKEPTICS PROCEEDINGS

I would like to invite interested readers of this post to give
comments and Dr. Nienhuys to answer a number of questions at the
end of my message:


------- (1/5) Proceedings just published ------------------------

The Proceedings of the 3rd Euroskeptics Congress, edited by J.W.
Nienhuys, have just been published. They contain three critical
papers about the Gauquelin Mars effect presented in Amsterdam
1991: (1) by de Jager & Jager (J&J), (2) by Jongbloet (J), and
(3) by Koppeschaar (K). Francoise Schneider-Gauquelin (S) and
myself (E) had the opportunity to give some comments at the
conference itself and to submit detailed written rebuttals of
the three critical papers to the Proceedings.

I received the drafts of J&J's paper (Jan 24, 1992) and submitted
a rebuttal referring to it (Feb 18, 1992). As to K, this author
sent me first and second preprint versions of his paper to which
I gave lengthy critical comments privately (13 pages) and he
made amendments. I renounced public comments to K's (and J's)
paper, K's final version was done late (May 1992) and J's paper
was less pertinent. (Rebuttal E (my own) is independent of S's).

--------- (2/5) Early suspicion ----------------------------------

In one of his letters (April 23), Dr. Nienhuys, the Proceedings
editor, wrote to me:

  "About the contributions to the Proceedings on the Mars Effect,
  these will amount to about 75 pages of haggling about
  statistics, and I will try to summarize the arguments. I will
  comment on the fact that the papers say something different from
  the impression created at the time of the conference... I think
  a general reader has more benefit from the opinion of an
  `outsider' than from one of the protagonists himself".

I became somewhat suspicious about those "75 pages", so I asked
Dr. Nienhuys in my reply (April 28):

  "Those 75 pages of haggling about statistics", did I read them?
  Even though you and I myself would probably not like that I give
  [in the Proceedings] comments to these pages it might be wise to
  send me the pages that I don't know of yet so that I have a
  chance to give [before publication] my comments to the
  respective author. I suggest this could also be done regarding
  your summary."

I did not receive any pages of those papers neither did I
receive Dr. Nienhuys' "summary".

---------- (3/5) Startling discoveries in the Proceedings -------------

In J&J's paper published in the Proceedings I discovered two long
additions filling together one full page.  The first addition
(p. 135) refers to the "eminence effect". In my rebuttal I
criticize J&J for not considering in their paper my eminence
argument of 1988:

  "Had the authors considered my account of ...  the role of
  emnence for planetary correlation they would probably not have
  concluded that the Mars effect lacked replication".

J&J's insertion in their paper is an attempt to undo the
criticized neglect. I won't comment here on what they added but
only state that readers are misinformed by J&J's new paragraph.
They will conclude that my critizing J&J for neglecting the
eminence issue is not justified.  In J&J's published paper the
eminence issue is indeed NOT neglected (that's what readers will
find) due to the authors' addition (that's what they aren't
told). The critique that I brought forward against J&J and which
they apparently took serious (otherwise they would hardly
have dared the risk of that addition) has thus been turned, for
careful readers, towards me.

As to J&J's second behind time addition, its effect is even
worse. In my rebuttal I write:

  "Since Gauquelin claimed that a Mars effect would only show up
  with sports figures of higher fame, he was entitled not to
  include in his analyses mediocre samples... . As a matter of
  fact, my independent assessment of the eminence of all sports
  champions based on objective citation counts in 21 appropriate
  sources showed...[:] unpublished Gauquelin sports figures were
  indeed much less famous on the average than published sports
  figures".

J&J inserted in their paper, again behind time, lengthy
comments on the issue of eminence differences between published
and unpublished Gauquelin samples. The readers' critical
judgement will again turn against me: "Ertel again ignores J&J's
account of this issue pretending that J&J did not discuss it".

Apart from this misinformation, I found one grave distortion of
facts in J&J's inserted paragraph.  In my 1988 paper I decribed
the Gauquelin athletes sample (N=4,391) as consisting of 2,888
published and 1,503 unpublished cases (p. 64 ff). Published
samples (N=2,888) were reported as consisting of three units:
2,088 athletes, published in the Gauquelin first data book
(Series A, 1970), 450 athletes published later in Series D6
(1979), and 350 athletes published in Series D10 (1982) (p.59)
(see Figure 1).

            I---------------------------I-----I-----I-------------------I
     N =    I       2,088               I 450 I 350 I     N=2,088       I
            I---------------------------I-----I-----I-------------------I
correct:    I     published                         I    unpublished    I
            I---------------------------I-----------I-------------------I
wrong (J&J):I    "published"            I     "unpublished"             I
            I---------------------------I-------------------------------I

          Figure 1:  Gauquelin data, correct and wrong (J&J) divisions

In their second behind time insertion J&J write:

  "One might tend to think that the `unpublished data' of
  Gauquelin consist solely of athletes of such low quality that
  the material `therefore' would not show a Mars effect, but that
  is not true. The `unpublished data' contain athletes of all
  categories of eminence, just as Gauquelin's data set of 2,088
  athletes which does not show the Mars effect. A further
  examination of the data even shows that the relative number of
  athletes of the five categories of eminence are practically the
  same in Gauquelin's (2,088) data set and in the 2,303 athletes
  of the `unpublished (ES) data'...  These various aspects
  contradict the suspected significance of the Mars effect...".

J&J's concise "This is not true" sounds as if that statement were
true.  But it is wrong. J&J's error is to mistake the N = 2,088
sample as the only published Gauquelin sample. The remainder of
the published sample (450+350=800) and the unpublished sample
are erroneously pooled. This error is hardly excusable. All
information regarding source of data and eminence counts are on
the file I posted them for analysis and in my 1988 paper minute
descriptions of that information is provided. References to
"2,888 published" and "1,503 unpublished" athletes (not "2,303")
are made nine times and five times, respectively, the samples
are listed in two tables.

Now, let us have a look at how J&J should have compared
the eminence (citation frequencies) of published and unpublished
samples (see Tabel 1):

                            Table 1

         Published       All published         Unpublished
 N     N =      N =           N =        %         N =         %
Cit   2,088   450+350        2,888      100        1,503     100
------------------------------------------------------------------

 0    1,178     153          1,331      46.1        940      62.5
 1      431     226            657      22.7        443      29.5
 2      271     191            462      16.0         73       4.9
 3      115      95            210       7.3         43       2.9
 4       39      57             96       3.3          4       0.3
 5       43      35             78       2.7          0        0
 6       10      23             33       1.1          0        0
 7        1      17             18       0.6          0        0
 8        0       3              3       0.1          0        0
-------------------------------------------------------------------

As can be seen, numbers of citations are considerably less for
unpublished as compared to published athletes. J&J having based
their conclusion on wrong sample divisions said there is no
difference. Their final conclusion based on this wrong
statement must therefore be rejected.

----------- (4/5) Another surprise ------------------------

As Dr. Nienhuys did not send me his own contribution to
the Proceedings I thought there was apparently no reason for
considering my request. His "the general reader has more benefit
from the opinion of an outsider..." had restored my trust in
him. I expected an essentially unbiased summary, his editorial
role of an 'outsider', I figured,  would per se not allow for
bringing in much of his personal views. But what he actually did
publish is a paper replete with support for the critics of the
Mars effect and devaluating comments on the rebuttals. I here
select only one item. JWN writes:

  "In the meantime Ertel and others have explored the eminence
  effect further. Around the middle of 1992 Ertel has offered the
  conjecture (based on empirical data, not theory) that the
  eminence effect does not hold for high ranks. A skeptic might
  interpret this as an outright refutation of the whole idea, and
  compare the investigation of the planetary effect to the
  detection of weasel and camel shapes in passing clouds. But
  there is another interpretation as well.

            Bias as artefact of exploration."

(The latter is the heading of the next paragraph which has
a different topic. The reader can hardly avoid to understand
that the "other" interpretation Nienhuys has in mind is
"Bias...").

My first objection regarding the quoted paragraph is this:
Nienhuys refers to a talk I gave at the First European Meeting
of SSE (Society for Scientific Exploration) in Munich in August.
Nienhuys was present.  Chairing the organization committee I had
invited him as a speaker. He knew that what I had presented in
my talk was not just "lack of a Mars effect for very high ranks"
but a unifying model proposing a curvilinear shape for eminence
relationships. This approach was supported by respective data
(actors, scientists, athletes). The model (shape of the
relationship) is a precise claim totally different from
Nienhuys' derogatory simile with clouds and camels (my paper
will be published in the Journal of Scientific Exploration).

My second objection is as follows: While picking up from my
Munich paper an issue with less bearing for the Proceedings
(What shape has the eminence relationship?) Nienhuys disregarded
another information in my Munich paper which is crucial for
the main issue discussed in Amsterdam: (Does the Mars effect
exist at all?). That is, in Munich I had reported about positive
results with CFEPP (skeptics) data, obtained of late, supporting
the Mars effect. Nienhuys did not refer to that information even
though I had indicated in my Amsterdam rebuttal that it would
soon be available: "...  final agreement should be expected, at
the latest from pending discussions with the French group
(CFEPP)".  In a footnote I provided details about that important
study.

--------- (5/5)  Questions ----------------------------------------

Postponing final conclusions I would like to consider, dear Dr.
Nienhuys, your answers to six questions:

(1) Did you send a copy of my rebuttal to J&J before publication?
(2) If yes, did you accept the revised version of J&J's paper
    *after* having provided them a copy of my paper?
(3) Were you aware of their having changed their paper?
(4) If yes, did you consider informing me about these changes
    and if yes, why did you decide not to inform me about these
    changes?
(5) Why did you decide not to send me your own contribution
    ("Summary") despite my having requested for it?
(6) If you hold that replication trials of the Mars effect with
    independent data are important why didn't you add while
    referring to my Munich report in your "summary" my information
    about Mars effect-supporting results with CFEPP?

I am looking forward to your replies.

I would also appreciate comments from readers of Mars effect
messages in this electronic community.


Suitbert Ertel

Institute of Psychology
Gosslerstrasse 14
3400 Goettingen
Germany