« Back to all events

Econometrics Research Workshop: External validity of school choice effects: Evidence from 350 markets

Dates:
  • Mon 04 Dec 2017 15.00 - 16.15
  Add to Calendar 2017-12-04 15:00 2017-12-04 16:15 Europe/Paris Econometrics Research Workshop: External validity of school choice effects: Evidence from 350 markets

Since Friedman (1955), economists have asked if giving parents greater school choice improves children's outcomes, with causal estimates suggesting effects ranging from positive to negative. A distinct literature, focused on external validity, considers the usefulness of causal estimates in this and other areas. Some papers argue that experiments from a given context can guide policy elsewhere, whereas others question whether this is the case. This paper brings these two strands of work together. It asks to what extent the causal effect of access to a better school generalizes across about 350 towns in Romania. It calculates value added for each high school in the country, as well as regression-discontinuity-based estimates of the effect of access to a better school. It then tries to predict causal effects across different towns using OLS and a random forest with five-fold cross-validation.

Seminar Room - Villa La Fonte DD/MM/YYYY
  Seminar Room - Villa La Fonte

Since Friedman (1955), economists have asked if giving parents greater school choice improves children's outcomes, with causal estimates suggesting effects ranging from positive to negative. A distinct literature, focused on external validity, considers the usefulness of causal estimates in this and other areas. Some papers argue that experiments from a given context can guide policy elsewhere, whereas others question whether this is the case. This paper brings these two strands of work together. It asks to what extent the causal effect of access to a better school generalizes across about 350 towns in Romania. It calculates value added for each high school in the country, as well as regression-discontinuity-based estimates of the effect of access to a better school. It then tries to predict causal effects across different towns using OLS and a random forest with five-fold cross-validation.


Location:
Seminar Room - Villa La Fonte

Affiliation:
Department of Economics

Type:
Workshop

Speaker:
Prof. Miguel Urquiola (Columbia University)

Similar events

 

Page last updated on 18 August 2017