Alistair,
This was never a problem in my area of research when all analyses were
carried out on a 'mainframe' computer with command language or syntax
files driving spss or other similar package and output was on a line
printer. In those days there was an automatic paper trail createdfrom the
raw data to the final analysis - the date and page number appeared on
every page of the continuous printer paper roll.
Quality assurance is far more of a problem now that researchers like to
use graphical interfaces with mouse controlled dialogue boxes on PC
screens and printed output that is often undated and un-labelled. I
prefer to continue with the use of syntax files and the creation of a
paper trail but many researchers dont and I believe there are analyses
finding their way into publications which cannot be replicated and are
simply wrong.
Please will you let me know what else you receive on this topic.
David
David Boniface
Statistics Group
University of Hertfordshire
College Lane
Hatfield
Herts. AL10 9AB
U.K.
telephone: +44 (0)1707 284366
FAX: +44 (0)1707 284799
e-mail: [log in to unmask]
web:
http://www.herts.ac.uk/business/groups/stats.htm
http://www.herts.ac.uk/business/centres/sscu.htm
|