A statistical technique for testing and estimating causal relations using a combination of statistical data and qualitative causal assumptions.