Date of Original Version

4-2013

Type

Thesis

Rights Management

All Rights Reserved

Abstract or Description

This thesis presents an alternative method for the detection of MIMIC models. Previous methods (such as factor analysis) suffer from a number of significant aws and limitations, which the new method (a causal search algorithm) doesn't suffer. A new algorithm is introduced, followed by a worked-through example of its application. Discussion focuses on some of the limiting assumptions the algorithm currently requires. Finally, recommendations for future work address improvements of the algorithm, as well as its applicability

Comments

Advisor: Clark Glymour

Department of Philosophy

Embargo Date

2013

Share

COinS