Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Causal Inference in the Presence of Latent Variables and Selection Peter Spirtes, Christopher Meek, and Thomas Richardson
 

Summary: Causal Inference in the Presence of Latent Variables and Selection
Bias 1
by
Peter Spirtes, Christopher Meek, and Thomas Richardson
1 We wish to thank Clark Glymour and Greg Cooper for many helpful conversations. This research
was supported in part by ONR contract Grant #: N00014­93­1­0568

2
I. Introduction
Whenever the use of non­experimental data for discovering causal relations or
predicting the outcomes of experiments or interventions is contemplated, two difficulties
are routinely faced. One is the problem of latent variables, or confounders: factors
influencing two or more measured variables may not themselves have been measured or
recorded. The other is the problem of sample selection bias: values of the variables or
features under study may themselves influence whether a unit is included in the data
sample.
Latent variables produce an association between measured variables that is not due to
the influence of any measured variable on any other. It is well known that where
unrecognized latent common causes occur, regression methods, for example, no matter
whether linear or nonlinear, give incorrect estimates of influence. When two or more

  

Source: Andrews, Peter B. - Department of Mathematical Sciences, Carnegie Mellon University

 

Collections: Mathematics