Please wait while the transcript is being prepared...
0:00
Welcome to the second
talk in this series,
Expert Knowledge
Elicitation with SHELF.
This one is entitled 'One
Distribution, One Expert'.
0:13
The course as a whole
- as you know -
has four talks, and this
one is the second one.
The first was 'Uncertainty
and probability',
which we've already done and the
next two are going to be
'Multiple experts' and
'Skills and resources'.
But let's just
recap where we were
after the first one.
0:32
In the first part we dealt
with the uncertainty
and probability,
we saw that expert
knowledge elicitation is
a valuable tool in quantifying
knowledge about
uncertain parameters,
which will help in
decision-making
in many application areas,
such as, for instance,
the pharmaceutical
and healthcare field.
We gave a couple of examples
of that in the first talk.
We also learnt that
expert probabilities
are elicited as
subjective judgements,
which like the many
subjective judgements
that scientists make,
need to be made
carefully, considering
all the available
evidence and using
a recognised and well-designed
protocol such as SHELF.
SHELF is the Sheffield
Elicitation Framework
and it's the subject of
this series of talks.
1:16
In this talk we'll
deal with identifying
challenges for subjective
probability judgements.
These are psychological
heuristics that can lead us to
biased judgements,
and we'll see how
the SHELF protocol
addresses those challenges.
It uses a carefully
constructed sequence of
questions designed to minimise
the potential for bias.
1:38
The outline of this talk
will be, first of
all, the challenges,
heuristics and biases,
and then we'll
take a first look at SHELF,
about eliciting a distribution,
the basic approach
that we use, and
the templates and slide sets
that shelf has
available for use.
