Stream Director: Simon Burgess
Funder: Leverhulme / ESRC
Staff: Simon Burgess, Kelvyn Jones, Ron Johnston, George Leckie, Carol Propper
We have continued to build on our research exploring the role of place in individual's outcomes. We have focussed on social renters as a group, and considered their long-term outcomes in terms of income, mental health and employment. These are an important group to study in their own right, but also offer one route for mitigating the problem of endogeneity as they have more limited choices on where to live. Results to date confirm our earlier findings that even within this group, variations in neighbourhood deprivation do not appear to have long run impacts. We have also considered employment outcomes in detail. We show that after taking account of the state of the local labour market, some individuals living in small neighbourhoods with very many unemployed people have lower chances of being employed. This effect is only present at extremes: in very high unemployment areas, and people with low propensity to be employed. Interestingly, the effect of unemployment in the local area appears to be stronger for the unskilled than for the skilled and for professionals.
Funder: Leverhulme
Staff: Simon Burgess, Richard Harris, Ron Johnston, Deborah Wilson, Ruth Lupton (CASE)
Our research on ethnic segregation in schools has established the intensity of that segregation for each ethnic group, at both primary and secondary school, and has shown that school segregation is much greater than residential segregation for those groups in the relevant Local Education Authorities. Exploratory work based on those findings is assessing the extent to which school choice (as indexed by the number and characteristics of students who do not attend the nearest school to their home) and student performance at KS1- KS4 reflect ethnic segregation.
Funder: Leverhulme
Staff: Marisa Ratto (with Richard Thomas and David Ulph (HMRC))
Peer groups can be important in setting social norms, which influence behaviour. The first project analysed tax compliance as a social norm and considered its implications for the deterrent effects of investigations. More recently, we have started work on social interactions and individuals' changing behaviour on tax law (non) compliance.
Funder: ESRC
Staff: Simon Burgess, Ron Johnston, Tomas Key, Carol Propper, Deborah Wilson
The national PLASC database allows us to model the evolution of a child's educational value added. The dataset provides all the school level quality variables and a few personal characteristics. Because of the location data, we can also measure the characteristics of the ward and Census OA for each child. Thus we can measure for each child the initial quality (in terms of Key Stage 1 and Key Stage 2 scores) of the school year cohort, the quality of the primary schools they went to, and the attainment of the children in the same residential area, plus other measures of the nature of the local area.
Funder: ESRC
Staff: Paul Greg, Lindsey Macmillan, Liz Washbrook
Description to follow...
One of the key points at which secondary school peer groups are formed is in the transition from primary to secondary school. The research analyses the implications of these flows for the composition of secondary schools. We look at the fragmentation of primary peer groups into many different secondary schools. We examine the degree of bifurcation in the flows – the extent to which poor children in a primary school go to different schools than do non-poor children.
We use PLASC to establish whether there is any causal relationship between school ethnic composition and segregation, and educational performance. In further, on-going work we have followed two statistical approaches to deal with the selection issue and get at any causal relationship. We have modelled variation in the minority-white test score gap across LEAs, and we have compared minority student test scores across LEAs. This research is on-going but results to date suggest that segregation may have a role to play in influencing test score outcomes.
(With Allen, IoE)
Whilst there are a range of measures of segregation available, none of them are ideal for modelling the correlates of sorting at a local area level. We are developing a regression-based approach to measuring segregation that does this and also allows a more direct link between the measure and a model of sorting.
