In a message dated 4/28/01 2:59:29 PM Pacific Daylight Time,
[EMAIL PROTECTED] writes:
<< The earlier responders make some good points but......I have seen plotted
regression lines when the rsquare was 0.005, scatterplots where two
"populations were separated by a line that makes a southern gerrrymander
envious, where clusters had fewer than 3 members, etc. etc. The whole thing
would be funny but these journal articles are used to make policy, affect
legislation, etc. there is hell to pay if a chemist misreads a spectrum or a
geologist confuses east from west. My feelingis that most egregious stuff
should be recognized by a "comment" in the journal. Sending in a comment to
a
journal is also a good learning experience for the student in that she have
to
be really sure it is a blooper and that the blooper makes a difference in the
conclusions.
>>
Excellent suggestion provided the journal is willing to print the comments
and admit that the article may have errors. This also means the journal has
to admit that they missed something when reviewing the article.
Just a thought
Dr. Robert C. Knodt
4949 Samish Way, #31
Bellingham, WA 98226
[EMAIL PROTECTED]
The Law of Gravity says, "No fair jumping up without coming down."
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
