Having joined the ASCL and accepted a new role, I visited their website and reviewed their data for school improvement resources, starting with ‘Making Effective Use of RAISEonline.’ Here, I found a plethora of RAISEonline and Performance table resources produced by ASCL members David Blow and Peter Kent. Seeing that both were presenting at the Leadership of Data Conference it submitted my CPD form.
First, ‘Making sense of all the data,’ (David Blow) then ‘Introduction to using RAISEonline for self evaluation’ (Peter Kent) and finally ‘Understanding the progress and VA measures’ workshop, again with (David Blow). These experienced colleagues had clearly been there, done it, got the T-shirt and done it again and again and again. Why should you find the time to listen to these colleagues speak?
First, they spoke from tried and tested experience. Second, they spoke with sincerity and real clarity. Finally, they spoke with a real collegiate ethos and integrity. Today I witnessed real leadership in action. I do not intend to rewrite their presentations, though I aim to share a few of takeaway messages.
Making sense of all the data
Knowing the performance of your school is a detectives story.’
There is no one set of magic data sets. Use the data to investigate. Equally, it is as important to avoid making the wrong conclusions.
Next David discussed Threshold and average data sets. For this David used a simple Mona Lisa explanation to encourage colleagues to be wary of data generalisation, I added the central threshold image to extend his metaphor.
The message was that threshold measures can be dominated by small groups of pupils around the threshold and that average measure dominated by the outliers, amplified by small groups n=10 (scattergrams are very useful).
Most measures have sub-dividion by KS2 prior attainment – which when taken with national figures gives a simple form of value –added
David put some conviction into the point that the DFE ‘get school data.’ He reinforced notational grades and expected Levels of Progress (Level 3 = D, Level 4 = C, Level 5 = B) and also stressed the importance of recognising there is variation with level eg 3abc. There was also a cursory point that although expected LoP sounds like a Value-Added measure, it is very closely correlated with raw attainment (e.g. %5A*-C incl En+Ma) because a greater percentage of pupils with higher prior attainment made expected progress in English and mathematics compared to those with lower prior attainment. (80.3% of pupils who achieved level 5 in Key Stage 2 English went on to make expected progress by achieving at least a grade B in GCSE English compared to 49.4% judged to be at level 2 at Key Stage 2 and made expected progress to grade E or above in GCSE English.
Given this information, the attainment vs prior attainment should focus on how did pupils perform relative to prior attainment in the higher ability range, (prior attainments 5), middle ability range(prior attainments 4) and lower ability range (prior attainments 3). Incidentally something that is relatively easy to do within SISRA.
Transition Matrices
Basically, a powerful subject value-added (VA) grid can be calculated by combining DfE transition matrices (TMs) for subjects with the existing pupil-level data already produced for each school by DfE for the exam result data-checking exercise each October. As a result you will need the files from the ASCL site and access to the DfE Checking site and files.
When the figures have been transferred from RAISE to the ASCL file, you can compare your school figures against the national figures for both for English and Maths, and by sub-level or whole Level, rather than just looking at the threshold figures. I think I have that right.
David offered one last tip, always start with cumulative assessments starting with the A*s.
Modelled VA
David encouraged delegates to get to grips with Confidence Intervals (CI) – as an indicator to how significant your SIGs where.
Second to be aware that negative VA in languages is often impacted upon by native speakers, who make significantly more progress. Do you know which students have a ‘language’ background?
Introduction to using RAISEonline for self evaluation – Peter Kent
‘Put your tanks on their lawn.’
Equivalent to a “King’s Pawn Game” in which you advance the king’s pawn two squares, occupying the centre square, attacking d5, allows the development of White’s king’s bishop and queen. Analyse and know your data. Prepare and present your key findings and most importantly your response / actions you are, or planning to undertake. Peter took the delegates on a ‘detective’ journey through an anonymised RAISE report and accompanying ASCL self evaluation workbook. Peter’s session was more seminar than keynote and was a welcomed break from data, data, data.
Both presenters made a real effort to recognise the contributions of the delegates.
‘Understanding the progress and VA measures’
This was a final breakout session. At the time is was brilliant, I do not think I have concentrated that hard in a while. I k to reflecting on this session at a later point, as for now, I am off to see if I can get the transition matrices spreadsheets talking to one another properly. I do believe I am quickly becoming a data nerd wannabe.
Interested in SLT data and use Twitter, let’s see if we can promote some collaboration with the hashtag #sltdata
David Blow is Head of The Ashcombe School, Dorking. Drawing on his mathematical and statistical background, he has been using data for many years, both at post 16 and pre 16, both within school and across Surrey. David is part of the ASCL Data Group.