Ipsos Encyclopedia - Causal Modelling / Path Analysis

Causal modeling is a way of testing causal relations between variables.

Definition

Causal modeling is a way of testing causal relations between variables. In market research the models typically have one clear target variable, such as loyalty, brand desire or purchase intent, and below that different layers of variables which can all have an effect. A good example is a Structural Equation Model (SEM), which tests if the causal model specified fits with the observed correlations between all variables.

A special case of causal modeling is path analysis, where different multivariate analysis (such as regression, manova, ancova) are put in one model. When conditions of multivariate normal distribution are met the simple decomposition rule is valid that the total effect of one variable on another is the sum of all direct and indirect effects via all possible paths. An indirect effect is the product of all direct effects on the path.

Path analysis can also include factor analysis models and latent variables.

Access the full Encyclopedia

New Services