Panel Data and Tobit Analysis on Health Care Dataset

As part of a research project for ECON 438: Econometrics II, we used a german healthcare dataset which is of panel data nature, and explored the research question What are the factors that determine the number of recent doctor or hospital visits of an individual?

We addressed the research question using two methods:

1. Zooming in on a single cross-section and using tobit models.
2. Using panel data models involving fixed, random effects and their numerous variants.

Much attention was paid to whether key assumptions like normality for tobit models or presence of serial auto-correlation for panel data models were violated or not. (Project Report + Source Code).