Social Profiling of Schools from Postcodes

The idea here is simple: use students' Postcodes to assess the social mix of their school. You can compare schools anywhere in Britain - and it works far better than Free School Meals.

Using Postcodes is preferable to building up simple school catchments from Census Enumeration Districts. If more children come from one ED than another, you have to apply weightings to allow for this. Using Postcode data adjusts automatically for exactly where the children live - even if this changes year by year.

You can use the Postcodes to grab Census data attributes if that's what you want to do. But our work suggests a better approach. This is based on using MOSAIC.

MOSAIC is a social area segmentation tool updated every 18 months. It uses geodemographic data from the Census and from more detailed and more up-to-date sources to classify each Postcode into one of 52 Types. By classifying individual Postcodes (average 16 households), MOSAIC provides a more precise summary of social circumstances than data for Census EDs (average 160 households). The Coventry map shows Postcodes as coloured dots. You wouldn't distinguish 52 different colours, so Postcodes are colour-coded to 12 broad MOSAIC Groups. Even at this crude level, however, you can see the difference between Stivichall (higher income area) in the south, Hillfields (disadvantaged inner city neighbourhood) and Keresley (old mining district) in the north. Coventry was chosen for the simple reason that it is near to SPA's office.

The table Exam Scores by MOSAIC is also by MOSAIC Group - the real thing uses the 52 detailed types. Results have been heavily "fictionalised", but resemble real results obtained for a large LEA nowhere near SPA's office! There is a clear relationship between social composition - as represented by MOSAIC - and exam performance. With results like this for a LEA, you can calculate what you would expect the exam results to be for any individual school, taking into account the social composition of its roll. This leads to a simple ranking of schools (see School Ranking) based on how favourable/unfavourable their social mix. What is more, the ranking focuses only on the relevant part of "social mix" - that affecting school examination performance.

One obvious application is for resource allocation in the context of LMS - providing a fairer distribution of money than using Free School Meals. In past studies, the correlation between exam performance and FSM was -0.62; that with a MOSAIC-based measure -0.81.

Other possibilities include adjusting League Tables of school performance to allow for social composition, or as a starting point for value added work. SPA, is currently piloting work on value added between GCSE Grades and A Level Grades, and we hope to add MOSAIC into that framework too. One of the advantages of a MOSAIC-based technique is that it allows comparison between schools of different types (eg primary/middle/secondary).

SPA Marketing Systems specialises in building computer models to predict shop turnover. This gives us access to highly detailed Postcode, Census and other databases and considerable expertise in social segmentation and analysis. Before moving to Leamington, SPA's MD, Dr Mark Shaw, was at the Centre for Environmental Studies, where he worked on multiple deprivation for the Department of the Environment. SPA pioneered the use of MOSAIC in education in the early 1990s, working with Professor Ivan Reid, Professor of Education at Loughborough University.

SPA offers a commercial service provided to LEAs on an on-going basis. There are two prime requirements for this - the technique has to work better than available alternatives, and it has to be affordable. At this stage, analysis is offered on a project basis only. However, SPA is also developing Windows software systems for use in-house by LEAs, and is considering the Internet as a possible delivery mechanism in the future.

DATA REQUIREMENTS
Project work would require a data file containing one record per student for all students - ideally in all primary and secondary schools (usually excepting special schools).

The content of each student's record might be roughly as follows:

Notes
If data has to be assembled from different sources, SPA could do this work if required. Data may be supplied in almost any standard Windows or DOS format.

Depending on sample sizes, it might be judged sensible to use data for two years. This would also allow us to look at year-on-year stability of the patterns that emerge. >

For maximum value to the Authority, you might wish to provide the Postcodes of ALL students of the schools - though there would then be many for whom no standard performance measure is available. SPA would base initial analysis on students with GCSE or SATs Scores (or similar) but could then apply the results to the whole school roll.

Types of Analysis
Whilst SPA is particularly keen to champion the use of MOSAIC in Education, we are very happy to tailor the analysis to meet specific requirements. LEAs currently using Free School Meals as an indicator may - for example - wish to compare the merits of FSM, MOSAIC and standard Census counts as a basis for resource allocation.

Click here for further information on Analysis Options.

back to top>

SPA Marketing Systems Ltd
Leamington Spa, Warks, CV32 6PT
Tel 01926 334978
Email:info@spamarketing.co.uk or Contact us

Please Note: If you wish to PRINT this page out. Click on the page first then select print