# statistics colloquium

Event Type:

Colloquium

Speaker:

Richard Hahn

Event Date:

Friday, January 29, 2016 -

3:00pm to 5:00pm

Location:

SMLC 120

Audience:

General PublicFaculty/StaffStudentsAlumni/Friends

Sponsor/s:

Stat group

### Event Description:

Title: Bayesian regularized regression for treatment effect estimation

Abstract:

In this talk, I describe the use of regularization priors in the context of

treatment effect estimation where the number of control variables is large relative

to the number of observations. A re-parametrized simultaneous regression model is

presented which permits prior specifications designed to overcome certain pitfalls

of the more conventional direct parametrization. The new approach is illustrated on

synthetic and empirical data. Finally, I describe the extension of our method to a

nonlinear mean regression model with homogeneous treatment effect.