PRIMARY HEALTH CARE STUDY
GROUP
Royal Statistical Society,
Discrete choice experiments
and Raasch modelling
TERRY FLYNN
(
Best-Worst Scaling - A new type of
discrete choice modelling
In eliciting patient preferences,
statements like "Quality of care is far
more highly valued than waiting
time" can be neither refuted nor supported
using traditional discrete choice
modelling. Best-worst scaling overcomes
such problems of estimating
attribute importance. Its advantages will be
illustrated with some
preliminary simulation results showing its relative
strengths.
KERRY HOOD
(
First excursion into discrete choice
experiments: a steep learning curve
This talk will cover the first
experiences of an applied statistician in DCE.
It will cover the perceptions
of initial simplicity through to some of the
complexities of designing an
experiment. The application is to the assessment
of the components of a
general practice consultation which patients rated most
highly within a trial
of shared decision making training for
professionals.
MASSIMO PENTA (Universite catholique de
Introduction to the Rasch measurement
model
Attributes of an individual, such as
pain or ability are generally
assessed through the individual's response to
item questionnaires.
However, the raw count of succeeded items is not
proportional to the
underlying individual attribute. The Rasch model
formulates the
requirements for quantitative comparisons of attributes and
allows the
measures to be expressed on a linear scale.