METHODS: Weak Influence


When the 'Effect Size' Is Small

In the Death Penalty Study, researchers asked 4 questions that addressed whether a person considers the death penalty immoral and a violation of human rights. Those answers were combined in an index, the average of the 4 answers. Only a minority of the 1,600 people in the study agreed that the death penalty was immoral (31%). Republicans were less likely to agree that the death penalty is immoral, when compared to non-Republicans (mostly Democrats and independents). 

As shown in the drawing, 34% of those that were Not a Republicans consider the death penalty immoral. Only 22% of Republicans consider it immoral. The 12-point spread is 'statistically significant.' But how much does being a Republican 'influence' a person's belief that the death penalty is immoral? The answer is: not much. Eta Squared is only .02. A social scientist would say that being a Republican accounts for only 2% of the variance in beliefs that the death penalty is immoral. The effect size is small.