This is the ninth in a series of lecture notes that, tied together in a manual, you might call “practical regression.” The purpose of these notes is to complement the theoretical content of most statistics texts practices based on nearly three decades of experience of the author, along with over a hundred years experience of colleagues who have offered advice tips. As the title of “practical regression” implies, these notes are a guide to regression in practice. This note addresses the issue of endogeneity, explaining how an explanatory variable may be endogenous (and thus its coefficient may be biased) if the cause is in doubt. Through an extensive learning curve example medicine, note introduces the concept of instrumental variables (IV) provides an intuitive explanation for why the tools to solve the problem of causality, and how to estimate IV and two stage least squares regressions. The paper describes the statistical evidence for the validity of the instruments.
Source: Kellogg School of Management
Release: June 11, 2012. Prod #: KEL643-PDF-ENG
Discrete dependent variables regression practical solution cases