A set of mathematical and statistical models, sometimes incorrectly referred to a causal moiling, that take the form of multiple regression based on an organized set of multiple linear regression equations with multiple outcome variables. It can be viewed as a special case of structural equation modeling, but it is not a measurment model. These techniques can be used in an exploratory or confirmatory fashion. In the past, path analysis has been referred to as ‘causal modeling’. Such a label is incorrect. Iit addresses data that are correlational, not causal.
See Latent factor/score/variable, Path diagram, Structural equation modeling (SEM)