scientific misconduct

[Martin Braun <nombraun@xxxxxxxxxxxxxxx>]

Alain,

instead of replicating experiment and analysis, you go on trying to spread
unsubstantiated allegations. This is scientific misconduct, which will do
damage only to you.

Examples from your latest letter to the list:

First example:
"from the start there was one chance in 25 that we're all talking about a
random pattern.  More if the author was on the lookout for an effect like
this, and (conciously or not) chose this database among others."

The facts:
If you know something about f0 research in speech, as you claimed you do,
you know very well that nowhere in the world is there any collection of
speech f0 data that could be compared, concerning quantity and quality, with
the set of IPO data that I analyzed. Further, as you apparently read the
paper, you know very well that I used the available data set in its
entirety. Knowing this and still suggesting that I could have selected the
data base in order to find a predetermined result is a maneuver that has got
nothing to do with science.

Second example:
"Notes ACDEFG (note: no B) were selected on the grounds that they are more
common than others in western music ...... This boosted p to 0.002.  It's
not clear if this selection was planned after the author looked at the
data."

The facts:
The selection of ACDEFG, for some of the analyses, was based on the
extensive studies on this issue by Miyazaki, which were published in three
papers. A division of the twelve tones, according to prevalence in music,
into two subsets of six tones could NOT have been otherwise. Knowing this
and still suggesting that I could have selected the tones in order to find a
predetermined result is again a maneuver that has got nothing to do with
science.


Conclusion:
My suggestion to you is this. Carry out a technically correct extraction of
speech-target f0 (the team at IPO will certainly help with any details),
carry out a meaningful statistical analysis, write a report, and then
publish it in a journal with a good review process. After that we can have a
discussion.

What you are trying to do at the moment is not doing any good to anybody.

Martin



----- Original Message -----
From: Alain de Cheveigne' <Alain.de.Cheveigne@IRCAM.FR>
To: <AUDITORY@LISTS.MCGILL.CA>
Sent: Sunday, May 13, 2001 2:16 PM
Subject: Re: histograms of F0 in speech contours