Dear all,
I am involved in the analysis of a diagnostic studies with more than 500
subjects.
Due to the large number of subjects, it makes sense to study also
sensitivity and
specificity in subgroups, e.g. males and females, old and young etc. We
can define
about 10 reasonable subgroups. My suggestion is now to present the
results as
a scatterplot, where each (labeled) point presents sensitivity and
specificity of
a subgroup. This way it is easy to identify subgroups where it is
difficult to make a
diagnose (lower left corner) or very easy (upper right corner).
I am pretty sure that such plots have been suggested in the literature
and I assume that there
exists a name for this type of plot. I just do not know it.
Hence I would highly appreciate any hints to an appropriate reference.
Kind regards
Werner Vach
--
Werner Vach
Forskningsenheden for Statistik - Syddansk Universitet
Department of Statistics - University of Southern Denmark
Mail address: Campusvej 55, 5230 Odense M, Denmark
Visitors' address: Sdr. Boulevard 23a, 5000 Odense C, Denmark
phone: +45 65 50 33 83 fax: +45 65 95 77 66
email: [log in to unmask]
|