[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
29188: RE: 29176: Leiderman re: Lancet study: good science? (fwd)
From: patrick richard <rich0303@hotmail.com>
Dear readers,
As a result of these comments, I took a quick look at the study itself.
Although random sampling is considered a gold standard in terms of study design
( GPS had to be used because of issues associated with random sampling in
developing countries-- see author's explanation), the study has serious flaws
and limitations in terms of data availability and statistical analyses. The
author has acknowledged some of them.
But, I do not think the author has ever established an association between the
temporary government and the rates of crimes in port-au-princes ( that was not
the goal of the study either). It would require more advanced statistical
techniques ( such as difference-in-difference or other techniques) than those
used here to compare the rate of crimes ( or the growth rate) after and before
the temporary government. This was simply a description of the rate of crimes,
which can be associated with so many factors not controlled for by the random
sampling. So any conclusion that the the rate of crime is due to the political
system in place is unfounded. There are also other measurement issues that I do
not want to address here.
Patrick Richard
N.B My comments are more concerned about the scientific value of the study as
opposed to defending a specific regime or not.