This project considers issues that arise when data are generated by some process of which there is imperfect knowledge and one wishes to extract information about some feature of the process from the data it generates. The following issues are addressed.

  What knowledge of a process is in principle obtainable from the data (evidence) it generates? Models of processes provide restrictions which can be sufficient to identify features of a data generating process, that is structural features. We study the nature of the restrictions that are required to identify interesting features and seek to determine minimally restrictive models for particular structural features.

  How can data be processed to give information about identified structural features? We study methods for estimation and inference in the context of models embodying weak identifying restrictions.

  Can models be falsified? There may exist a model that identifies a structural feature which embodies restrictions so weak that the model is non-falsifiable. Conclusions drawn from processing evidence through such a model must be contingent on the veracity of the restrictions embodied in the model. We study the characteristics of non-falsifiable models that identify interesting structural features and how the existence of more than one distinct non-falsifiable model bears on the interpretation of evidence.

The project builds on a stream of research on the subject of identification and inference started in econometrics in the 1920’s and pursued since in economics and other areas of social science.

The project investigators are:

Andrew Chesher Hidehiko Ichimura and Sokbae Lee.

Visitors to the project with interests in this research agenda include:

Charles Manski (Northwestern University),
Joel Horowitz (Northwestern University),
Whitney Newey (MIT) and
James Heckman (University of Chicago).

Publications / other docs