aamc's research evaluation project and rand report release
TRANSCRIPT
AAMC’s Research Evaluation Project and RAND Report Release
Webinar
Ann Bonham, PhD Chief Scientific Officer July 8, 2013
AAMC Research Evaluation
Initiative
Health Equity Research Expert Panel
New & Shifting Research Models Expert Panel
Foundational & Basic Science Research Expert Panel
Housekeeping
• We welcome you to type in your questions and comments at any time during the webinar, we will save time at the end to address as many of them as we can.
• We also welcome you to submit questions on this initiative at any time via [email protected]
• This webcast will be will also be recorded, and posted on our website, along with a collection of FAQs. If we don’t get to your question at the end of the presentation, we will try to address it online.
Measuring Research A guide to research evaluation frameworks and tools
Sue Guthrie, Steven Wooding and Jonathan Grant
Webinar, July 2013
7
Outline of presentation
• Background
– Who we are
– Policy background
– Purpose of report
• Frameworks and tools reviewed
• Understanding the challenges and trade-offs in research evaluation
• Developing a customised research evaluation framework
• Next steps and discussion
8
RAND Europe is an independent not-for-
profit public policy research institute
• Not a university or management
consultancy — but with capabilities of both
• Part of the global RAND Corporation
• Work across the breadth and depth of
government
• Strongly held values of quality and
objectivity
• Provider of evidence
• Track record in research evaluation
“help improve policy and decisionmaking
through research and analysis”
9
The project team
Jonathan is a Principal Research Fellow and former president at
RAND Europe with a key interest in research evaluation. He has
conducted a wide range of studies for clients in the UK, Middle East
and Australia.
Steven Wooding is a Research Leader at RAND Europe with an
extensive portfolio of work in research evaluation and science
policy. He has worked for clients in the UK, Canada, USA and
Australia.
Sue Guthrie is a Senior Analyst at RAND Europe working primarily
in the fields of research evaluation and science policy. She had
conducted research for clients in the UK, Europe, Canada and the
Middle East.
10
A brief history of research evaluation (1)
• 1960/70s (USA) – Project Hindsight, Comroe and Dripps, etc.
• 1980s (UK) – Research Assessment Exercise (RAE) and performance
management of research in universities
• 2002 (UK) – Joint funding bodies’ review of research assessment
– “the need to fully recognise all aspects of excellence in research (such as
pure intellectual quality, value added to professional practice, applicability,
and impact within and beyond the research community)”
• 2004-07 (Australia) – Research Quality Framework
• 2006 (USA) – National Institutes of Health (NIH) Clinical and Translational
Science Awards
• 2006 (UK) – Government announces its intention to replace the RAE after
2008 with an assessment system based on metrics
11
A brief history of research evaluation (2)
• 2009 (USA) – American Recovery and Reinvestment Act (ARRA)
• 2010 (USA) – Science and Technology for America's Reinvestment: Measuring
the Effect of Research on Innovation, Competitiveness and Science (STAR
METRICS)
• 2009-11 (UK) – Higher education funding bodies undertake impact pilot and
announce decision to include impact in the Research Excellence Framework
(REF) 2014
– “The assessment of impact will be based on expert review of case studies
[that] include any social, economic or cultural impact or benefit beyond
academia.”
• 2012 (Australia) – Excellence in Innovation (EIA) Trial
– “Universities can generate compelling case studies of impact across the
whole range of disciplines and impact areas”
12
Why research evaluation matters
• Need to demonstrate accountability for the investment of public
funds in research
– demand for good governance, particularly in climate of fiscal
austerity
• Shift in emphasis from purely summative evaluations to more
formative and comprehensive evaluations
– cover wider outputs from research
• Interest in and demand for the evaluation of research is increasing
internationally
13
What did AAMC want from the project?
• Put research evaluation on the agenda
– help institutions align mission and incentives
– move away from the ranking business
– understand what impacts are, including clarity around
definitions for quality and impact
• Consider how AAMC members
might develop research
evaluation frameworks
14
Aims of the study
• Act as a ‘how-to guide’ to
evaluating research
– Understand the challenges
and trade-offs in evaluating
research
– Provide examples of
frameworks and tools used
for evaluating research
internationally
• Promote discussion of this
topic of increasing importance
Report available at: http://www.rand.org/pubs/monographs/MG1217.html
15
Our approach
• Review of existing frameworks and tools for the evaluation of research
– Detailed review of 6 frameworks
– Brief review of a further 8 frameworks
– Review of 10 tools
• Analysis of the characteristics of tools and frameworks using a factor analysis approach
– What are the trade-offs?
– What can be learnt?
• Developed decision tree to aid development of customised research evaluation frameworks
16
Outline of presentation
• Background
• Frameworks and tools reviewed
– An example framework – Excellence in Research for Australia (ERA)
– An example tool – case studies
• Understanding the challenges and trade-offs in research evaluation
• Developing a customised research evaluation framework
• Next steps and discussion
17
We reviewed six frameworks …
• Research Excellence Framework (REF), UK – assesses performance
of UK universities to determine funding allocation
• STAR METRICS, US – uses data mining and other low burden
methods to account for federal R&D spending
• Excellence in Research for Australia (ERA), AU – uses bibliometrics,
and other quantitative indicators, to map R&D output
• Canadian Academy of Health Science (CAHS), CA – aims to provide
consistency and comparability while retaining flexibility
• National Institute of Health Research (NIHR) Dashboard, UK –
provides performance management information at various levels of
aggregation
• Productive Interactions, EU – flexible approach, aims to help
institutions learn and improve their performance against their own
goals
18
… and ten tools
• Bibliometrics
• Surveys
• Logic models
• Case studies
• Economic analysis
• Peer review
• Data mining
• Interviews
• Data visualisation
• Site visits
• Document review
19
An example framework: ERA Origin and rationale: Perceived need to include assessment of quality in block funding allocation (previously volume only). Advocacy purpose - demonstrate quality of Australian research
Scope: Assesses quality, volume, application of research (impact), and measures of esteem for all Australian Universities at disciplinary level
Application to date: First round in 2010, broadly successful. Next round 2012, with minor changes. Intended for funding allocation, but not used for this as yet
Measurement: Indicator approach, uses those appropriate at disciplinary level. Dashboard provided for review by expert panel
Analysis: Broadly positive reception. Meets aims, and burden not too great. Limitation is the availability of appropriate indicators
Wider applicability: Should be widely applicable, criticism limited in Australian context. Implementation appears to have been fairly straightforward
20
SWOT analysis for ERA
Strengths • Acceptable to research community in
Australia
• Burden on participants is moderate
• Indicator driven
• Produces a single performance
indicator, which can be used for
ranking
• Multi-disciplinary
Weaknesses • Indicator driven
• Still moderated through peer review,
reducing objectivity
• Not comprehensive – academic focus
• Summative
• Burden relative to return is high (not
yet used for funding allocation)
• Requires some central expertise
(bibliometric expertise on panel)
Opportunities • Potential to add new indicators
Threats • No funding implications
• Politics informed its development
• Government and public appetite to
include impact is limited in Australia
(limits potential for development)
Broken Link designed by Stephen JB Thomas from The Noun Project
21
An example tool: Case studies
What is it?
Flexible enough to capture a wide variety of impacts, including the
unexpected, and can provide full context around a piece of research,
researcher, or impact
When should it be used? Primary limitation is that findings are less generalisable - best used
where examples are needed rather than full coverage. Often used to
provide the contextual information alongside another approach that
provides generalisable information and to understand process variables
in depth
How is it used? Can be used in a variety of ways depending on aims. Key
considerations are the unit of analysis and the sample selection for
multiple case studies. Structure and approach should reflect the purpose
Watch from The Noun Project; Question designed by Martin Delin from The Noun Project
22
Outline of presentation
• Background
• Frameworks and tools reviewed
• Understanding the challenges and trade-offs in research evaluation
– Key findings of factor analysis
• Developing a customised research evaluation framework
• Next steps and discussion
23
Key findings of analysis
• There is no silver bullet
• The framework should be designed based on the purpose of the
evaluation
• Research evaluation tools typically fall into one of two groups
• There is a range of possible units of aggregation
• There are some perennial challenges to research evaluation that
need to be addressed
• Research evaluation approaches need to suit their wider context
• Implementation needs ownership, the right incentives and support
24
There is no silver bullet
Designing a research evaluation framework requires trade-offs:
• Quantitative approaches tend to produce longitudinal data, do not
require judgement or interpretation and are relatively transparent, but
they have a high initial burden
• Formative approaches tend to be comprehensive, evaluating across
a range of areas, and flexible, but they do not produce comparisons
between institutions
• Approaches that have a high central burden tend not to be suitable
for frequent use
• Approaches that have been more fully implemented tend to have a
high level of central ownership
• Frameworks that place a high burden on participants require those
participants to have a high level of expertise (or should provide
capacity building and training to achieve this)
25
The framework should be designed based
on the purpose of the evaluation
• Analysis
• What works in research funding?
• Advocacy
• ‘make the case’ for research funding
• Accountability
• To taxpayer, donors, etc.
• Allocation
• What to fund (institution, field, people …)
Advocacy
Allocation
Analysis
Accounta
bility
26
Research evaluation tools typically fall into
one of two groups
• Group1: formative tools that are
flexible and able to deal with
cross-disciplinary and multi-
disciplinary assessment
• Group 2: summative tools that do
not require judgement or
interpretation, and are
quantitative, scalable,
transparent, comparable and
suitable for high frequency,
longitudinal use.
Gro
up
2 c
hara
cte
ris
tics
Group 1 characteristics
Bibliometrics Data
mining
Economic analysis
Surveys Logic models
Document
review
Peer
review
Site
visits
Interviews Case
studies
Group 2
Group 1
• Multiple methods are required if researchers’ needs span both groups
27
There is a range of possible units of
aggregation
• The units of aggregation
used for collecting,
analysing and reporting
data will depend on:
• target audience(s)
• tool(s) used
• Units of collection,
analysis and reporting
are interdependent
Research
group
Institution
Department or
programme
Field
Research
system
Project
Researcher
ER
A
RE
F
SM
CA
HS
NIH
R
PI
29
Research evaluation approaches need to suit
their wider context
• Acceptability and credibility
• Differences between countries
• Need to ensure framework
does not discriminate
30
Implementation needs ownership, the right
incentives and support
• Where compulsory, the challenge is to obtain support from the
academic and wider community
• Where participation is voluntary, incentives need to be in place to
promote and sustain uptake
• In both cases, participants need to be given the skills necessary for
the process, through simplicity, training or a toolkit
• In all cases, strong central ownership is needed for effective large-
scale implementation
31
Outline of presentation
• Background
• Frameworks and tools reviewed
• Understanding the challenges and trade-offs in research evaluation
• Developing a customised research evaluation framework
– Identified a set of key questions to consider in developing a framework
– Developed a decision tree to support the development of a framework
• Next steps and discussion
34
Outline of presentation
• Background
• Frameworks and tools reviewed
• Understanding the challenges and trade-offs in research evaluation
• Developing a customised research evaluation framework
• Next steps and discussion
– Piloting of set of measures and metrics for AAMC membership
– Discussion?
35
Two expert panels will develop and test set
of metrics and measures for evaluating
research • Two panels have been established of ~7 experts each
– Health Equity and Research
– New and Shifting Research Models
• Aim is to develop evaluation framework(s), metrics and measures
that can be implemented, tested and piloted for wider dissemination
as a potential model for other areas of research
• Focus is on self evaluation, not competitive benchmarking
• RAND Europe working with AAMC to facilitate this process
– Starts in Sept 2013 with two workshop, and will report in Sept 2014
Q&A
• Please submit your questions and comments via the chat feature.
• We also welcome questions and feedback about this initiative anytime at [email protected]
• This webinar will be recorded and posted online with FAQs at www.aamc.org/researchevaluation