EarlySight

Blog

Is a class performing well or performing badly? How EarlySight draws the distinction.

The top class in a year may actually be under-performing hugely...and the worst class in the year may be performing fantastically. EarlySight lets SLT draw a distinction and identify that gap. How do we achieve that? We look at who is in the room before we calculate anything.

· 5 min read

What your current class data hides
Never take a class’s performance at face value.

A class is a group of individual pupils, not a unit

A class’s sanction count, a set’s grades and a form’s attendance say as much about who is in the room as about what happens in it. Put a group of pupils who struggle everywhere into one class, and it will look like a problem class even when the lessons are going well. The worst-performing Geography class may actually be a group of pupils who are exceeding their expected performance and SLT should recognise this. So before EarlySight calls anything good or bad, it compares pupils with themselves: the same pupils in their other lessons, in their other subjects, or at an earlier point.

Same sanctions, different stories

Two Year 8 science classes, Class 1 and Class 2, each have 24 sanctions this half term. On a raw count they look the same.

EarlySight compares each class with what the same pupils receive in their other lessons, scaled to the lessons they have each week. With four science lessons a week, Class 1 ‘s pupils would be expected to pick up about 22 sanctions, so 24 is in line. In Class 2 you’d expect about 8, so 24 is three times as many - Class 2’s pupils are behaving a lot worse in Science than in their other subjects.

Class 1 is a class of pupils who find school hard everywhere. Class 2 is a class where something about the room, the subject or the mix is worth a conversation. On EarlySight only Class 2 shows up as a concern. The same check works for one pupil. A teacher’s share of a pupil’s sanctions is set against that teacher’s share of their timetable. Ten sanctions from a teacher they see half the week isn’t a pattern. Five from a teacher they see once a week might be.

Weak at Latin, or weaker overall?

Take a Year 9 pupil whose Latin grade puts them in the bottom 20% of the year. On raw grades, that’s a Latin problem and a name on the Latin intervention list. But across all their subjects, they sit in the bottom 5%. Latin is one of their better subjects. The concern is broader than Latin, and it needs a broader response. “Subjects: the same pupils, compared with themselves” sets each pupil’s position in a subject against their average position across all their subjects. It works for sets too: a strong set looks strong on raw grades even when the subject is doing less well for those pupils than their others. Grades are ranked on the school’s own reference scale first, so a 9–1 grade, a BTEC grade and a working-at code can sit side by side. Comparing pupils with themselves can hide low attainment, so the raw position is always shown alongside.

When the room changes

Classes also change during the year. Take 9R from our demo school at the end of last year. Three pupils left it during the year. One moved to 9T in January after a run of incidents, one changed schools in February, and one started at alternative provision after Easter.

9R, end of last year Everyone who was in 9R Pupils in 9R now Pupils counted 31 28 Average attendance 92.6% 94.8% Below 90% attendance 6 (19%) 3 (11%) Behaviour points per pupil 2.5 1.5

Using traditional metrics, 9R is a problem group and the underperformance of those three pupils mean that the numbers will never recover within the Academic Year even though they have now left. Counted by who is in the room now, we get a genuine measure of how the cohort is performing. Trends are where it matters most. Counting everyone who was there at the time, 9R’s behaviour points fell by two-thirds between the autumn and summer terms, from 38 to 13. Count only the 28 pupils in 9R now, in both terms, and they went from 14 to 13. The room changed. The pupils mostly didn’t.

Nobody disappears. Pupils who leave stay in 9R’s history for the time they were there, and the pupil who moved to 9T takes their record with them, so their new tutor sees the autumn as well as the spring. Grade changes work the same way: a group’s movement is the average of each pupil’s own change, never the group’s average now against a different set of pupils then.

What this means in EarlySight

Before EarlySight calls a class good or bad, it asks how the same pupils do elsewhere.

  • A class’s sanctions are compared with what the same pupils receive in their other lessons.
  • Each subject is compared with the same pupils’ other subjects, with the raw position shown alongside.
  • Today’s class and tutor-group figures are built from the pupils in the room now. Pupils who leave stay in the group’s history for the time they were there, and a pupil who changes class takes their record with them.
  • Change is each pupil’s own change. Only pupils here at both points are counted, including in a group’s last 12 months, which follow the pupils in the group, not the group itself.
  • Where the question is how the school did at the time, such as last year’s headline figures, we count everyone on roll then. The SLT week-by-week charts can switch to today’s pupils, and the cohort tables to “Pupils here throughout”.

A raw number mostly tells you who is in the room. EarlySight tries to tell you what happened in it.

We’ll show you EarlySight on a demo school, and talk through what it would look like with your own data. Email to book a demo

← All posts