Path Analysis

Abstract

Let us assume three variables , A, B, and C, to be analyzed. The regression analysis for predicting C from A and B is based on the causal model , with A and B causes and C the result. However, this model is not guaranteed to indicate the true relationships among A, B, and C. The true causal model might be “A causes B which causes C” or “A causes B and C.” Path analysis is a procedure in which users form causal models by themselves and select the model fitted well to a data set. The origins of path analysis can be found in Wright’s (1918, 1960) biometric studies and Haavelmo’s (1943) econometric ones (Kaplan 2000).

The original version of this chapter was revised: Belated corrections have been incorporated. The erratum to this chapter is available at https://doi.org/10.1007/978-981-10-2341-5_17