Dr Janet Godolphin
Design and Analysis of Experiments
This is my main area
of research. I am particularly
interested in design selection for experiments involving blocking. In
general the research is motivated by medical, horticultural and
is an acknowledged serious problem in experimental design. I have
investigated causes of disconnectivity for various
types of design involving blocking such as; Latin
square designs, row column designs, crossover designs and blocked factorial designs and
have established that disconnectivity is due to a structure in the
model termed a staircase
partition of the treatments. I
have made some investigation into the choice of design
with particular attention given to the distribution of rank reducing
computation of sums of squares in ANOVA performed by
various statistical packages can result in sums of squares
pertaining to null hypotheses different from those under
consideration. In conjunction with two collaborators, I am working on a
method for computing ANOVA sums of squares which incorporates null
hypotheses in the process.
methods in Statistics
Journals containing my papers include: Applied
Statistics; Computational Statistics and Data Analysis; Quality and
Reliability Engineering International; Utilitas Mathematica.
Please see my publications list.
Invited seminars have been
given at several universities in the U.K. and my work has been
at several international conferences.
Please see conference proceedings.
Reducing Observation Sets
researchers and practitioners have asked that an automatic program
the specification of rank reducing observation sets (see publications) be made generally available. At present
I use a personal interactive program. However, I am cooperating with
colleagues to produce a user-friendly version of this program to be
placed here on the website.
Courses taught in the 2007-2008 academic year include, MS237 Mathematical
Statistics, MS333 Statistical Methods for Business and Finance and
MS343 Mathematics Education.