Causal model

In metaphysics and statistics, a causal model (also called a structural causal model) is a conceptual model that represents the causal mechanisms of a system. Causal models often employ formal causal notation, such as structural equation modeling or causal directed acyclic graphs (DAGs), to describe relationships among variables and to guide inference.

By clarifying which variables should be included, excluded, or controlled for, causal models can improve the design of empirical studies and the interpretation of results. They can also enable researchers to answer some causal questions using observational data, reducing the need for interventional studies such as randomized controlled trials.

In cases where randomized experiments are impractical or unethical—for example, when studying the effects of environmental exposures or social determinants of health—causal models provide a framework for drawing valid conclusions from non-experimental data.

Causal models can help with the question of external validity (whether results from one study apply to unstudied populations). Causal models can allow data from multiple studies to be merged (in certain circumstances) to answer questions that cannot be answered by any individual data set.

Causal models have found applications in signal processing, epidemiology, machine learning, cultural studies, and urbanism, and they can describe both linear and nonlinear processes.