Revised REF rankings tables

Revised REF rankings tables

In my analysis of the Impact of Impact I showed that due to the different scatters in the three different subprofiles in REF 2014 (Output, Impact and Environment) the effective weights in Grade Point Averages were not the nominal weights (0.65, 0.2, and 0.15) as you might naively expect.  This was especially true in Physics where Impact had the largest effective weight (equal with Sociology).  It is interesting to think how the rankings would have been if each subprofile had had the same scatter. Statisticians would do this by calculating the standardised variables (i.e. normalising by the scatter before combining).  So, I’ve recalculated the rankings using standardised variables and weighting by nominal weights.

The ranking table for Physics is below.

So you can see from this that e.g. Sussex is 27th in the standardised table but was 33rd in the original table and so has “lost” 6 positions due to variation in scatter in the different metrics (i.e. primarily because Impact had a larger scatter).   The big losers in this sense were Portsmouth (-10 positions), Birmingham (-8 positions), King’s College (-7), QMUL and Sussex (-6 position).

Winners were Glasgow (+10), Surrey (+9), Leeds (+8) and Manchester (+7).

If you want to look at the rankings for any other UoA, they are all included in this spread sheet. ImpactImpactRankings

One final comment.  It is clear to me from my analyses and discussions with others that the  outcomes in Physics in particular from the REF in absolute terms were generally very good across the board.  The relative differences between different departments are actually quite small.  However, it is these small relative differences which drive the league tables.  The league tables are therefore not a very useful measure and should be treated with extreme caution. As we plan at Sussex for our next REF I am going to focus on what we can do to continue to demonstrate our excellent research what we can do to improve on this but try and ignore the randomness of the league tables.

UoA # Original GPA Original Rank Ranking with standardized GPAs weighted Change in ranking (Standardised – Original) Institution
9 3.33 2 1 -1 University of Oxford
9 3.33 4 2 -2 Universities of Edinburgh and St Andrews
9 3.35 1 3 2 University of Strathclyde
9 3.3 5 4 -1 Cardiff University
9 3.33 3 5 2 University of Nottingham
9 3.24 10 6 -4 University of Southampton
9 3.22 11 7 -4 University of Warwick
9 3.27 8 8 0 University of Durham
9 3.28 7 9 2 University of Cambridge
9 3.27 9 10 1 Imperial College London
9 3.16 13 11 -2 Heriot-Watt University
9 3.13 15 12 -3 University of Sheffield
9 3.28 6 13 7 University of Manchester
9 3.05 24 14 -10 University of Portsmouth
9 3.1 19 15 -4 Queen’s University Belfast
9 3.07 22 16 -6 Queen Mary University of London
9 3.05 25 17 -8 University of Birmingham
9 3.16 14 18 4 University of Bath
9 3.12 17 19 2 University of Exeter
9 3.16 12 20 8 University of Leeds
9 3.12 16 21 5 University College London
9 3.02 29 22 -7 King’s College London
9 3.06 23 23 0 University of Liverpool
9 3.04 26 24 -2 Lancaster University
9 3.02 27 25 -2 Liverpool John Moores University
9 3.07 21 26 5 University of Bristol
9 2.93 33 27 -6 University of Sussex
9 3.1 18 28 10 University of Glasgow
9 3.1 20 29 9 University of Surrey
9 3 31 30 -1 Royal Holloway, University of London
9 3.02 28 31 3 University of Kent
9 2.93 32 32 0 University of York
9 2.89 34 33 -1 Swansea University
9 2.67 36 34 -2 Keele University
9 3 30 35 5 University of Leicester
9 2.7 35 36 1 University of Hertfordshire
9 2.6 37 37 0 University of Central Lancashire
9 2.53 38 38 0 Loughborough University
9 2.43 39 39 0 University of Huddersfield
9 2.42 40 40 0 Aberystwyth University
Advertisements

About sebboyd

By day...Professor Seb Oliver, Professor of Astrophysics and Director of Research and Knowledge Exchange for School of Mathematical and Physical Sciences. By night...Seb Boyd, father of 3, husband of one and dabbler in board-game design
This entry was posted in Uncategorized. Bookmark the permalink.

One Response to Revised REF rankings tables

  1. Pingback: The Impact of Impact | Seb Boyd

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s