Skip to main content

Data Generating Process

Go Search
New Atlas
Atlas, A-Z
Atlas Maps
MPP/MPA Programs
Core Topics
Illustrative Courses
Topic Encyclopedia
Concept Dictionary
Career Tips
Best Practices Project

PPGPortal > Home > Concept Dictionary > D, E > Data Generating Process

Data Generating Process (DGP) 

The unobserved mechanism that yields sample data on outcome variables.

( Paul Grootendorst, Toronto, PPG2010.)



The analyst’s goal is to learn about the actual underlying data generating process, through the examination of sample data. The task of learning about the DGP can be understood as a three-step process:

1.) Write down a model of the outcome variable that hopefully is consistent with the underlying DGP.

2.) Estimate unknown features of this model by: collecting sample data on realizations of the outcome variable and, if necessary, explanatory variables; proposing an estimator the unknown features of the model; and computing the estimates using the sample and the estimator.

3.) Test the adequacy of the model by asking questions about it. Examples of good questions to ask are: do the estimates seem plausible? Are the predictions from the model consistent with the actual values? Is the estimator sufficiently precise to be able to produce useful estimates?




Paul Grootendorst, Toronto, PPG2010.


Important Notices
© University of Toronto 2008
School of Public Policy and Governance