Nexus Statistical Neighbour Methodology

Nexus now offers users the ability to include statistical neighbours as one of the available comparators in the system. This development was delivered in April 2024. 

The new comparator applies across both the Insight and Nova modules and is selectable as an additional option in the same options menus where the user previously was able to select the DfE Region as an option (not including the now optional KS1)

There are a number of factors that need to be noted in terms of this statistical neighbour comparison, explained below. The illustrations included are not real and are simplified for ease of reference. 

Mathematical methodology - Pupil Level

This model was selected for Statistical Neighbours in Nexus based on the feedback of LA when presented with two different options and also represents a value for money consideration.

In this model very large Local Authorities have more influence over the statistical neighbour averages than in smaller Local authorities. This is because they have more pupils who undertake the assessment and the number of pupils achieving a standard (or not) is therefore also greater.

The illustration below highlights - in purple - how this works in the aggregate though - in reality - every pupil's results are individually included to get to these results, with column F being looking at the number of pupils passing in each LA (column E), summing this and dividing it across the overall cohort. 

Differences to some other data sources

The note about the mathematical methodology above points to the selected method used in Nexus highlights the selected method of reporting, however it should be noted that some other methods exist. This means that Nexus may be different to some other sources such as LAIT with regard to the Statistical Neighbour comparison. 

LAIT uses an average of averages approach to producing this comparison. In LAIT that means that the average score for each individual LA is then averaged across all of the LA who are in the statistical neighbour group. This has the effect of ignoring the size of larger and smaller LA and their relative weightings based on numbers of pupils in the results set, instead the LA simply counts once in terms of it's average score being averaged into the statistical neighbour figure. 

Neither methodology is more correct than the other, they are just different and Nexus users should understand the basis on which they are being used.

The average of averages methodology represents each LA as an equal (irrespective of how many pupils achieve the target or took the assessment) rather than allowing the results of a bigger LA to be a determining factor of the overall average amongst a number of other smaller LA. 

The illustration below highlights - in red - how this works, with column D simply being the average of the values in column C. 

This illustrates the differences between the two different mathematical models. 

Statistical neighbour changes over time

Statistical neighbours are changed over time by the DfE as the characteristics of areas also change. For this reason, the Nexus statistical neighbour comparison reflects the results for the Local Authorities that were included in the list for the users LA at that time (i.e. the selected reporting year)

This could mean - where historical reporting is available - the comparison could be using different LA from one year to the next, though these changes are relatively rare so this should not often occur. 

In the event that DfE again change the Statistical Neighbours, this is planned to be reflected in Nexus. 

Vulnerable Groups (CIN, CLA, SEN) analysis

All users who use vulnerable group analysis (CLA, CIN or SEN2) will only see all analyses once all of the required data sets have been imported into Nexus.

Perspective Lite (not available)

This comparison is not available in Perspective Lite as it has little meaning to schools as a concept.