Here are some notes based on my experience using the "critique a
statistics article" project.

1.  The students do not know what a primary source is.  Add an explanation
to the document and show them some examples.  Also some comments should
be added to the effect that a "serious" source must be found (e.g. it must
have an author and references).

2.  The students' library and computer skills are in many cases
extremely weak.  Therefore (in retrospect) it seems clear to me that one should:

     Do an in-class demo of the use of the PubMed database.  This should
     be complete: start with some key words, select some articles, and get
     them (via interlibrary loan if need be).  Have the articles available to
     show the students.  They should see the whole process!  I would get three
     articles, because you don't know for sure (in advance) which are primary
     sources.

3.  Emphasize to the students that they may want to initially try some keywords,
and if they have no luck, try looking for materials with a completely different
subject.  Some subjects are not altogether amenable to the project.  For
example, some topics lead one (ultimately) to monster studies of mortality
(for the entire U.S. population); for these it is hard to get at the primary
source.

4.  Perhaps one should require that the study involve getting at least some
medical information from the subjects.  Otherwise you find the students 
looking at some really silly surveys.

5.  Can we ask the Library to conduct sessions to teach the students how
to use the web?  If so, this could be done prior to the classroom demo.

David Jaffe