Causal Analytics Toolkit (CAT) Demonstrations

with PhD statistician Louis Anthony (Tony) Cox

The George Washington University Regulatory Studies Center hosted two demonstrations of the Causal Analytics Toolkit (CAT) developed by Tony Cox

What is CAT? A new Excel-based tool for examining the causal relationship between exposure to substances and health effects.

Government policies are often supported by evidence of statistical associations between events (such as changes in environmental exposure to a substance or changes in economic conditions) and endpoints (such as health effects or gross domestic product), but these associations usually rely on untested (and often potentially biased, mistaken, or uncertain) modeling assumptions. As a result, they may not describe a true cause and effect relationship. Policy decisions based on unreliable models are not only at risk of failing to achieve desired results, but may divert scarce resources from more productive uses.

While objective and rigorous methods of causal analysis can help to formulate and test causal hypotheses and to estimate causal impacts of changes in certain conditions on desired outcomes, they are not readily accessible to users who are not specialists in causal modeling and modern computational methods and software. This may include many in the risk analysis, environmental, and policy analysis communities who might want to carry out appropriately sophisticated causal analyses of observed patterns, but who are currently limited by software packages and methods that only calculate measures of association, such as regression models, odds ratios, and confidence intervals.

The Causal Analytics Toolkit (CAT) provides powerful, easy-to-use software to help overcome these limitations. It makes state-of-the-art causal analytics available to anyone who has Microsoft Excel.