Multiple Determinants and Effect Size

Effect sizes such as correlation coefficients and standardized mean differences gain meaning in the context of how the relevant variables are measured and the causal model in which the variables are embedded. Whether a particular effect size is judged to be small or large hinges on how much measurement error is present, whether the relevant target variables share causal variables, and the redundancy of those shared influences. Effect sizes are attenuated by measurement error and are larger when target variables share many rather than few causal determinants, especially when the causal determinants are correlated.