UNIVERSITY OF HERTFORDSHIRE STATISTICS GROUP SEMINAR SERIES

Sponsored by Department of Statistics, Accounting and
Management Systems, Business School, University of Hertfordshire.

Thursday 4th May 2000, time 1pm, Hertford Campus, Room M211
Roger Jones (University of Hertfordshire)

Title: "Cuscore Control Charts"

Abstract:
"The problem is this. Given present (time t) and past (times t-1,
t-2 ...) measurements y(t), y(t-1), y(t-2) ... on some quality
characteristic of interest, what is the best way of charting the data
so as to be able to detect, say, a step change in the process mean as
soon as possible? Box and Luceno (1997) describe a way of answering
this question that generates some interesting results. In this talk we
look at their method and some of the results that stem from it."

For further information regarding location, parking, etc.,
please contact Neil Spencer (details below)



Wednesday 7th June 2000, time 2:30pm, Hatfield Campus, Room A209
Michael Healy (Visiting Professor, University of Hertfordshire)

Title: "Measuring physical maturity"

Abstract:
"It is easy to see that some children are more physically mature than
others of the same chronological age, yet it is not obvious just how
maturity is to be measured.  It differs from measurements such as
height and weight in that all normal individuals pass from one common
condition of complete immaturity to another of complete maturity, so
that it makes sense to measure maturity on a scale going (for example)
from 0 to 100.

A start can be made  by the use of 'milestones', observable stages in
development which occur in all subjects.  Thus a 13-year old girl who
has started to menstruate can be considered to be more physically
mature than a girl of the same age whose periods have not yet begun.
A whole sequence of events can be used to give more information,
though the assessment of maturity from such a sequence is bound to be
coarse.  However there is no problem in finding many sequences of
events which proceed in parallel - examples are the eruption of the
teeth and defined stages in the appearances on X-rays of the bones of
the hand and wrist.  There is then the statistical problem of
combining the evidence on maturity which is provided by each of the
sequences.

The solution to be discussed consists on ascribing a numerical score
to each stage in each sequence and measuring maturity by the average
score.  Underlying this approach is the assumption that the sequences
are all reflecting a single maturity dimension, and so the scores are
allocated so as to minimise the disagreement between parallel
sequences summed over a standardising sample.  Crucially, the
minimisation must be constrained to avoid the trivial solution which
allots the same score to every stage of every sequence.  It will be
shown that the choice of constraint is not straightforward."


****************************************
Dr Neil H Spencer,
Senior Lecturer in Statistics,
Department of Statistics, Accounting and Management Systems,
University of Hertfordshire,
Hertford Campus, Mangrove Road,
Hertford, SG13 8QF,
ENGLAND.
Telephone: +44 1707 285529
Fax: +44 1707 285489
E-mail: [log in to unmask]
****************************************



%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%