Report by the Chairman to Community

Following the pattern of the last two years, here is a short informal
report on behalf of AGP panel members – as ever, official news and
policy statements will come from STFC itself. My apologies that this
report is a little later than last year’s.

KEY POINTS

* The 2012 round was not as badly oversubscribed as 2011 (which
required mitigation through transfer of funds into the round), but
still fiercely competitive, and we were not able to fund all of the
excellent science proposed

* Levels of Investigator time awarded (FEC) are worryingly low, but we
cannot change this without losing considerable RA support

* STFC will be reviewing the Consolidated Grants scheme in 2013

* The next deadline is Feb 13th 2013

BACKGROUND

This was the second year of the new scheme. The community is getting
used to the new process, so we deliberately made no big changes in
policy or procedure. Further down this report, there are a few words
on how the panel feels the new scheme has been going.

GRANTS RECEIVED

This round was not as large as 2011, but still intensely competitive.
We received 27 applications (cf 35 last year). These 27 applications
contained 196 distinct projects, including 170 RA and Technician
posts, and requesting a total of £63M, a factor of roughly 2.5 more
than STFC could afford. As last year, the overbidding factor compared
to existing baseline had a large range, roughly from 1 to 4. So no
sign yet of overbidding calming down!

HOW IT WORKS

We allocated projects to one or more of the sub-panels – Astronomy
Observation (AO), Astronomy Theory (AT), Solar Science (SS), and
Planetary Science (PL). As in previous years, every application had
multiple reviewer reports, and an “Introducer”, who writes an initial
report before the panel meetings, and co-ordinates additional
“Assessor’s Questions”. The sub-panels then met in two sessions –
AO/AT together, then SS/PL together – lasting a total of 5 days. The
AGP chair and deputy chair attended all the meetings. Before the
meetings, all sub-panel members provided initial scores, but only the
chairs saw these. At the meetings, each project was discussed in turn,
and then given a final score and ranking around the table. As you
might expect, when it became obvious that a project was near the
bottom or near the top, discussion was accelerated, so that we could
spend most of the time near the middle – but every project was
considered fully.

A subset of the AGP then met as the “merging panel” for a further two
days. Selected projects were cross-read by members from the other pair
of panels in order to agree matching points in the two initial lists.
This process resulted in a small but definite re-scaling of the
rankings.

The final output from the AGP was a recommended ranked ordering of
projects, which STFC could work down through once they knew their
budget situation. To help us arrive at that ranked list we used, as in
previous years, a scoring system based on the categories mandated
STFC-wide, and explained in the official guidelines. Each category was
scored 0-5, with each possible grade defined by a standard wording
such as “competitive with the best science funded worldwide”. This
temporary scoring system is only used for internal discussion; the
actual result is an agreed ranking order for projects, and
recommendations on overall award components for a group. (Some things
people request cut across projects).

It is important to realise that AGP judged investigator-time-only
cases on exactly the same basis as RA cases – is the science
excellent? Is it clear what the investigator will actually do?

BUDGET and FINANCE ISSUES

The budget situation was similar to the year before. The AGP does not
actually hold a budget; we produce a recommendation based on a ranked
list, and STFC work down this until they run out of money. However, we
are traditionally advised on roughly what STFC expects to have
available, which at the moment is approximately £9M per year of new
commitment, so usually we know very roughly but not precisely where
the funding cut-off will be.

STFC staff have to rely on a frankly inadequate RCUK-wide grants
system, and so do additional manual calculations to actually deduce
the financial consequences of each grant. Calculating the effect of
changes is therefore non-trivial. Nonetheless we achieved our aim this
year of notifying applicants of the outcome in November.

RESULTS

In total we recommended 84 posts, including RAs and Technicians.
Compared to existing holdings, groups obtained roughly level funding
on average. This is better than the 2011 round, where most groups
suffered a 30% cut. The difference is partly because this round had
less demand volume, and partly because groups proposing in 2011 did so
previously at a historical maximum in grant generosity, whereas the
2012 applicants last got funding in a round that was already tougher
than the year before. Based on our absolute quality scores, the
fraction of “world class” proposals we are able to fund has stayed
about the same.

We also recommended 20.9 FTEs of Investigator time. We followed our
published guidelines of recommending 20% FTE on average for a major
involvement in a project, and smaller amounts for secondary
involvements. There were also some successful “FEC only” requests. The
net result was that the recommended Investigator time was on average
24.8% of the recommended RA+Tech time.

At the request of Science Board, we made a careful analysis of results
by science area, geographical area, and facility use. There were no
obvious trends, but we did note that some unsupported or soon to be
unsupported facilities are still popular with researchers! We also
looked at the effect of group size. There is no consistent statistical
effect, but of course at the small group end, simple statistics means
that there is a greater chance of receiving no funding.

This year there were five New Applicant proposals, of which three went
forward to full review and two were funded. Two others were considered
under urgency and one was funded. There were two consortium proposals
and they were both partly successful.

HIGH PERFORMANCE COMPUTING

There was considerable confusion this round over HPC requests,
basically because the DiRAC initiative, which should for the medium
term future satisfy the majority of community HPC requirements, was
only announced after people submitted their grants. AGP reviewed
requests on a case-by-case basis, but tended to be more supportive of
year-1 requests. In the future, the guidance STFC has given AGP is
that use of DiRAC-II should be the expected route for HPC activities,
but AGP has still not received a clear description of the funding
model, so DiRAC time may or may not require explicit funding.
Applicants are strongly encouraged to contact the Office for the
latest guidance.

NEXT ROUND

The deadline for the 2013 round is Feb 13th, and of course many of you
are already beavering away towards this. Our schedule is similar to
last year – during April applicants will be replying to reviewer’s
questions; in July and August there will be a chance to respond to
Panel questions; the panel meetings will be in August and September;
somewhere around October the provisional outcome will be discussed by
Science Board; and the results should be announced by November. The
general pattern of sub-panels and meetings will be the same as the
last two years. The categories against which proposals are judged will
be the same as in previous years.

THE FEC ISSUE

Even more than last year, we are keenly aware that the levels of
Investigator time we are able to award is not fulfilling the original
aims of the FEC scheme. On average we award 0.25FTE of investigator
time per FTE of RA time, and for most astronomers this is their only
source of “FEC”. If we had awarded the 0.5FTE level needed to replace
QR, the fall in RA numbers would have been even more catastrophic –
Investigators cost twice as much as RAs. Worse, many research active
staff have not received any support. We must all make very clear to
our University administrations that in the current climate, FEC
support is NOT an indicator of whether a staff member is research
active.

REVIEW OF CONSOLIDATED GRANT SCHEME

We understand that during the coming months Science Board will review
the entire Consolidated Grants scheme. Do be on standby to make
constructive comments. For your information, the points that we have
already made to Science Board were as follows:

(i) Management of expectation. Previously, rolling grants were
oversubscribed by a factor 1.3-1.5, and standard grants by a factor of
6-7. Of course, a large fraction of standard grant applications would
be repeated every year with modifications, so their real
oversubscription was perhaps more like a factor 2-3. The first year of
the new scheme was oversubscribed by a factor 2.5, and this year by a
factor 2.1. So the net management of expectation has shown little
change, but in fact the ex-rolling grant groups have been caused to
overbid by more.

(ii) Research concentration. Large groups have declined by about the
same rate as small, groups on average. Some small groups have failed
to win anything. However, some have won resource, and some new groups
have appeared. There is no obvious net drift either towards or away
from major groups. Of course, it may be simply too early to tell.

(ii) Use of flexibility powers. This will only become clear by the
second cycle, but there is some anecdotal evidence. Large groups are
able juggle money at the margins, but even in these groups this has
limited effect. The behaviour is likely to be very different from PP
groups, because grants are facility exploitation focused and
individual PI-centred, rather than group-project focused. Applicants
take their proposed science plans very seriously; they are not seen as
a kind of convenient fiction to win money.

(iii) Simplification. From the AGP point of view, we have saved
administrative costs – by reducing panel size, by removing applicant
presentations, and by reducing the number of meeting days. However the
burden on panel members has increased. We have maintained the
principle of rigorous peer review of specific proposed science, rather
than, for example having a “light-touch” review of past group
performance. It seems clear that any further simplication/cost saving
can only be obtained by sacrificing that rigorous peer review of
specific science cases.

(iv) Co-ordinated research versus individual projects. The result of
the new scheme has been not so much an expanded rolling-grant like
system, but more like a giant standards grants round, but with the
opportunity to try again next year lost.

Andy Lawrence

Edinburgh January 2013