Presenter: Gabriella Conti, University College London

Title: Workforce quality in early years interventions: Evidence from a large-scale home visiting programme (joint with Sarah Cattan and Christine Farquharson)

Abstract:
Small-scale ‘model’ programs targeting disadvantaged populations have been shown effective at improving children’s well-being in a long-lasting fashion. However, scaling up successful interventions is challenging. Impacts of scaled-up programmes are often found to be weaker and more heterogeneous than those of small-scale programs, but little is known about the factors that drive their effectiveness. 

In this paper, we investigate a crucial but under-researched factor: the quality of the workforce delivering the early years programmes. We do so in the context of the Family Nurse Partnership (FNP), a large-scale home-visiting programme in England targeting first-time teenage mothers, which has been shown to benefit children’s cognitive development throughout primary school. For identification, we exploit a unique feature of the FNP assignment process whereby family nurses are randomly allocated to clients, conditional on a small set of mother characteristics known at recruitment. 

After presenting evidence supporting this feature of the assignment process, we estimate distributions of family nurse effectiveness on the cognitive, socio-emotional and health outcomes of the child. We find evidence of substantial variation in family nurse quality. A one-standard deviation (SD) increase in family nurse quality leads to a 0.25 SD increase in child's cognition, and to a 0.29 SD increase in child's socio-emotional development at age 2. We also show that family nurse quality is correlated across some, but not all dimensions of child development. 

We then exploit the rich data we have on family nurses, their supervisors and the wider context to predict family nurse effectiveness. We find that family nurse effectiveness is difficult to predict: despite the very rich information we have available, we are unable to explain more than 15% of it. In particular, traditional ‘hiring’ measures (demographics, education, training, previous experience) only explain 5% of the variation in effectiveness, while detailed indicators of programme implementation are more predictive of effectiveness than externally assessed process quality. 

Organisers: Timo Hener and Michael Koch

The seminar is on-site and will not be streamed via Zoom