What are projections? 

Projected grades are statistically generated predictions of the grade a student is likely to score at GCSE based on their most recent Trust assessment.   They are calculated using the same methodology that examination boards use to assign actual GCSE grades.  It does not matter, for the projection methodology, whether that common assessment is at the beginning of Year 7, or March of Year 11. 

How is this forecast calculated? 

  1. All students in Trust who study a given subject sit the same assessment and their work is marked. 
  1. Place all students in order, based on their mark, and assign them a percentile placing.  (At this point, their marks will no longer be used) 
  1. Look at national data to see what proportions of students, nationally, achieved each grade in a GCSE qualification.  The divisions for English are shown to the right as an example. 
  1. Assign identical proportions of students in the Trust those grades, using their percentile rankings. 

The diagram to the right tells us that, nationally, 3% of students achieved a grade-9, and 65% achieved a grade-4 or above.  In this case, we would assign the top 3% of students in the Trust a grade-9, and we would ensure that 65% of students achieve a grade-4 or above.  

Common sense, and experience, both tell us that students who are top of a class for their whole academic career will achieve top grades.  However, in some high achieving classes you may also achieve top grades if you are the 5th best student, or even the 15th.  In other classes, being top may only make you a strong average.  Projections allow us to give students a rank order in a class with several hundred members.  If you are top of a class of 500 children we can be extremely confident you will achieve a grade-9.  If you are ranked 250 in such a class, so on the 50th percentile, we can also be extremely confident that you will achieve a grade-4. 

Projections rely on children throughout the Trust completing the same assessments, so we can be confident that the marks are directly comparable. 

That seems very simple? 

The above explanation is simplified, but the full methodology uses the same principles of ranking students, and dividing them up based on national proportions to assign grades.  The more detailed version is included at the end of this document. 

What are the benefits of projections? 

These projections will allow us to say with confidence exactly how individual students, departments, cohorts, and academies are performing.  For example:  

  • We could see that a student has been projected a grade-6 from the start of Year 7 to the end of Year 9.  If they dip down to a grade-4 in Year 10, we can quickly see the discrepancy and investigate why this is happening 
  • Within a single Academy we could see Year 8 students have a larger pupil premium gap than Year 10 students.  That Academy could then deploy resources accordingly.   
  • We could see the percentage of a Year 7 cohort that would achieve grade 5 or above in both English and maths.  This would allow us to make a judgement about the five year trends in an Academy, or intervene with borderline students far earlier than ever before.   
  • We could see which Academy has the strongest provision for geography in Year 8.  We could use their expertise to improve provision for students elsewhere in the Trust. 

How accurate are projections? 

One of the key tests of the projection methodology is to take children’s raw marks from their actual GCSE and place them into our own system.  When we do this we find that 72% of mathematics, 90% of English literature and 91% of English language grades are accurate.   No grade is more than one away from accuracy. 

Another key test is how accurate our projections are following Year 11 mocks.  Here, we count a projection as accurate if it is the same as a child’s eventual GCSE grade, or one under their eventual GCSE grade (partly because some improvement in a child’s performance is to be expected; partly because we prefer to give cautious judgements in the run up to GCSEs).  Using this metric, 89% of mathematics projections are accurate, and 54% of English projections are accurate after the final mocks.  At the end of Year 10, 72% of mathematics projections and 50% of English projections are accurate.  These figures are all higher than the equivalent figures for predictions. 

What if a teacher feels a projection is wrong? 

Before we publish or use projections we will provide all departments with a chance to look at a first draft and ‘adjust’ projections.  Departments will be able to be make as many adjustments as they like and there is no limit on the size of these adjustments.  If teachers feel a child will achieve either a higher or lower grade than that they are projected to achieve then we ask them to make the appropriate adjustment.  The adjusted grade is the one that will be used for all purposes. 

When and how do we use projections? 

We generate projections for Years 7-11 for both Trust assessments (Y7-10) and mock/trial examinations (Y11) 

Projections are used to report home for Years 7-10, whether they are directly printed onto the report, or used to calculate a progress-to-target judgement.  We use projections in these reports because one of the aims of reporting is to give families an indication of how well students are likely to perform in terminal examinations, which is typically compared to a student’s target to give an indication of the extent to which that grade constitutes success.  Projections are the most accurate way of providing this information.   Some academies also use current grades for Year 11 students because it is useful to have a current performance grade so close to terminal examinations. 

In addition to their function as the basis for reporting home, projected grades are also analysed within schools and departments to gain an understanding of the progress of both individual students and cohorts are performing and inform action in response.  More information on this process is available from Data/Impact Leads. 

What is the more complicated methodology? 

There are three key changes to the methodology described above: 

  1. There are some extra steps in the methodology that precisely mirror the national methodology for setting grade boundaries.  The extra steps are as follows: 
    • Instead of dividing students up into each grade (9 slices), divide students up into four slices, each with multiple grades in it (grades 1-3; grades 4-6; grades 7-8; grade 9).   
    • This is used to generate grade boundaries for achieving a grade 1, a grade 4, a grade 7, and a grade 9 (look at the mark that the lowest scoring student in each slice achieved).   
    • The remaining grade boundaries are created by equally splitting the grade boundaries that already exist.  For example, if the grade 7 boundary is 64 marks, and the grade 9 boundary is 78 marks, the grade 8 boundary will be set at 71 marks –halfway between the two. 

    The net result of this is that students in each slice may not be evenly distributed.  We may have many, many students getting a grade 4 and relatively few getting a grade 5.  Or many students getting grade 8s and not many getting grade 7s. 
  1. The methodology above assumes that the students sitting an assessment are statistically normal and directly comparable to the national cohort.  This may not always be the case: triple scientists, for example, are likely to be a bright cohort.  Fortunately, we have access to transition matrices that tell us how well students from each starting point are likely to achieve.  We use these to work out the likely result for each specific cohort of students. 
  1. The methodology above assumes that students in the Trust will perform as well as students across the country.  It doesn’t allow for the possibility that the Trust, as a whole, might exceed or fall below the national average.  To ensure accuracy we also project based on the proportion of grades achieved by that subject in the Trust in last year’s GCSE results.  If key thresholds change by more than 10% we will use this alternative set of grades.  At this point, we will communicate to departments what is happening so they can pay careful attention to their adjustments, in case the new cohort are likely to perform differently to past cohorts.