Upload
lidc
View
199
Download
1
Embed Size (px)
Citation preview
Analysis of the Research Excellence Framework 2014 impact case studies
Ian Thornton
In partnership with
King’s analysis
Four strands to extract insights
Topic modelling: Identify hidden thematic structures or topics in corpus of documents
Keyword in context: Identify keywords displayed within surrounding context
Information extraction: Automate extraction of specific words (nouns) such as countries
Qualitative analysis: Read and hand-code samples of case studies
287 relevant case studies, found through two methods
Surprisingly little overlap between case studies identified by the two methods
These 287 case studies were mostly submitted to the life sciences and social sciences panels
The research has impact in 175 countries
With many usual suspects, but also c.40 mentions for Australia and the Netherlands
This research was supported by a very wide range of UK organisations
But the number of mentions did not
correlate at all with investment in development-
relevant research
2013-14£383m of public R&D funding is spent on internationaldevelopment.
Equal to 8.3% of the £4.6bn science budget
% of UK public R&D funding spent on international development
Breakdown by UKCDS members2013-14 Each £ symbol = £8mDFID
£ £ £ £ £ £ £ £ £ £ £ £ £ £ £ £ £ £ £ £ ££ £ £ £ £ £ £ £ £ £ £ £ £ £ £ £ £ £ £ £ £MRC
£ £ £ £ £Other RCs
£Wellcome Trust
£ £ £ £4.6% of Wellcome Trust committed awards are on science for international development
2011-12 2012-13 2013-140%
5%
10%
Given that they spend dramatically different amounts (which itself is changing)
Most research claimed to have impact across multiple sectorsWith lots claiming to have an impact on health outcomes
The majority of case studies claiming to
impact on health were not submitted to life
sciences UoAs
84 case studies cited observed outcomes stemming from the research
Caveats galore, but still (I think) useful insights
• Caveats with the sample
• Caveats with the methods
• Hints of interdisciplinarity – but the UKCDS Board decided not to pursue
What people really want are stories, not analysis
UKCDS’ top 20