In the social sciences, researchers often are interested in how one characteristic of a person influences another characteristic. In the Death Penalty Study, researchers were interested in how certain belief systems 'influenced' the type of actions people might take regarding the execution of a condemned murderer. Notice that small, medium, and large effect sizes are at the lower end of the scale in the diagram. An example of a large effect size is the 'influence' that a belief (the death penalty is immoral) exerts on self-identifying as an Abolitionist (no method of execution is acceptable). While the effect size is large, it only accounts for 25% of the variance in identification as an Abolitionist or non-Abolitionist. The other 75% is not known or understood. 

Indeed, the classification of small, medium, and large effect sizes is a matter of convention (which is contested by some as arbitrary). The convention is an intuitive way to communicate a range of Eta Squares, a way to summarize the 'strength' or 'power' of one characteristic over another. 

For humanists, this inability to mathematically predict human feelings and actions is good news. One can pursue scientific explanations of social behavior—like supporting or opposing executions—by using sophisticated data collection and data analysis tools. But 'explanations' apply to large groups of people, not individuals. These 'explanations' are always only partial explanations. Predictions are probable, not absolute. Unlike robots, humans arguably exert agentry and free will. 

RETURN>