There isn't anything substantive here. He says
researchers sometimes make faulty inferences,
that different models can produce different
results, and notes the trickiness of the ol'
correlation v causation issue. Several of the
papers the gives as case studies are misrepresented.
A contradiction: He cites regression models
to show that results from other regression models
"myths," yet simulataneously argues that regression
models are bunk. Frankly, his description of how
he thinks statistical research is done, and of the
intellectual climate in which discussion of such
research occurs, is terribly naive.
A disturbing footnote is the author teaches a course
utilizing regression analysis:
http://crab.rutgers.edu/~goertzel/methods.htm
Cheers,
Chris Auld
Department of Economics
University of Calgary
[EMAIL PROTECTED]
