powered by NetLogo

WHAT IS IT?

The purpose of this initial model is to provide an idea of how we might implement an ABM to ask questions about schooling circuits and parental choice. Currently it simulates a random population of parents in a catchment with four schools for only a single year of school allocation. This is more a proof-of-concept model than one useful for investigating interesting questions.

HOW IT WORKS

Four school exist in this imaginary catchment (represented as white squares with their number). All school have the same number of places available to prospective pupils (set by the 'SchoolSize' slider). Schools differ in their quality, determined by their 'GCSE-score' (analagous to the percentage of students with GCSE grades A* - C in the previous year). GCSE-score is assigned to each school randomly using a normal distribution with mean 50 and standard deviation 20. GCSE-scores of the four schools are shown in the 'GCSE-scores' bar chart.

A population of parents (represented by the blue people icons) is initially created (the number of which is specified by the 'Families' slider). Parent locations within the catchment are assigned randomly. Each parent can rank each of the schools for their preference (rank1 is highest, rank4 is lowest). The numebr of parents ranking each school in each position (i.e. ranks 1 - 4) are shown by the bar charts School1 - School4 on the right side of the display.

Currently, each parent can have one of two strategies for ranking schools:

i) Quality Strategy: schools are ranked solely on their quality with 'good' schools (i.e. higher GCSE-score) most preferred
ii) Distance Strategy: schools are ranked solely on the distance to the school, with closest schools preferred

The approximate proportion of parents with the Quality Strategy is specified using the 'Quality-Strategy' slider. Strategies are assigned to parents at random.

The model simulates a single round (year) of pupil place allocation. The closest parents to a school which they have ranked highest are allocated a place at that school. If there are more parents that have ranked the school highest, the closest X parents are allocated places, where X is the size of the school. Once a school is full no more allocation takes place. For school with places unfilled after all the highest ranks have been allocated, they then allocate the closest parents who ranked the school second highest. If places remain unfilled, 3rd rankings are allocated and finally fourth rankings until all schools are full. If there are more parents than total number of places available in the schools some parent will remain unallocated.

Parents change colour on allocation. Parents become green if they have successfully been accepted at the school which they ranked 1st, yellow is they are allocated to the school they ranked 2nd, orange to the school they ranked 3rd and red to the school they ranked 4th. Parents become grey if they remain unallocated.

The proportion of parents successful in gaining at place at their 1st ranked school is shown in the 'Total Success' box. Success rates for the two different strategies are shown in the 'Quality Strategy Success' and 'Distance Strategy Success' boxes. The distribution of the places parents were allocated compared to their ranking of the school they were allocated at are shown by the 'Quality-Strategy-Outcomes' and 'Distance-Strategy-Outcomes' bar charts.

HOW TO USE IT

The model is populated by clicking the 'Populate' button.

Places are allocated by clicking the 'Allocate' button.

Change values of 'Families', 'SchoolSize' and 'Quality-Strategy' to explore outcomes.

THINGS TO NOTICE

When Quality-Strategy = 1.0 spatially homogenous groups of success rates occur and Total Success is always 0.25

When Quality-Strategy = 0.5 spatially homogenous group of successful parents occurs in the area around the best school. Other school sub-catchements are have a mix of successful and unsuccessful parents. The mix of successful with unsuccessful parents varies by school quality (note yellow parents mixed with green around second best school vs. red with green around worst school). Depending on values of SchoolSize and Families Total Success is around 0.68

When Quality-Strategy = 0.0 the vast majority of parents are successful (Total Success around 0.98) because most parents can be allocated to their closest school. Some parents on the boundaries between school sub-catchments can be unsuccessful.

EXTENDING THE MODEL

We could:
- add more strategy types
- add propability distributions to parents choice of strategy
- have parents choose their own stategy choice of strategy driven by their about how far within a particular school's sub-catchment they live (linked to forthcoming paper)
- add parent attributes such as income, ethnicity etc.
- consider multiple allocations through time (maybe link school quality to some parent attributes)
- allow parents to move house between years
- include siblings
- consider multiple neighbouring boroughs (catchments)
- link spatial representation to actual maps of East London

- loads more....

CREDITS AND REFERENCES

Model by James Millington, 01/06/11
Online at: http://landscapemodelling.net/NetLogo4/SchoolChoice_v1.html