IEU seminar: Meng Le Zhang
Online via Zoom
Title: The potential of NILS for studying peer effects in health: A case study of fertility amongst neighbours using a natural experiment
Abstract:
Our neighbours can potentially influence our behaviour. For instance, poor health behaviours amongst neighbours may normalise and reinforce our own poor health behaviours. This is an example of a peer effect. A shortage of data on peers (e.g. neighbours); homophily and violation of SUTVA are the main obstacles to measuring peer effects.
The Northern Ireland Longitudinal Study (NILS) is a study consisting of 28% of the NI census and roughly 50% of all households. Household information from NILS is linked to a database of all addressable properties in NI. In theory, NILS contains relatively complete information on households and their next-door neighbours.
Using this information, it is possible to study the peer effects of any intervention provided that a valid experimental design exists (inclusive of natural experiments). I present an example pilot study using a well-known natural experiment to test whether Neighbour A's fertility (i.e. number of children in the household) affects Neighbour B's fertility.
I will begin my talk with a showcase of other spatially informed causal designs that I have worked on.
Biography:
My recent papers and works in progress include studies using:
- administrative boundary changes to evaluate a house regeneration project;
- postal survey response rates to measure the impact of Scottish policy on homelessness;
- data anonymisation in online crime maps as a natural experiment on the effects of public statistics on house prices;
- natural experiment(s) in fertility to study peer effects amongst neighbours in Northern Ireland
All welcome Zoom link