METHODS: Strong Influence

aaf

When the 'Effect Size' Is Large

Researchers compared Abolitionists and non-Abolitionists regarding the belief that the death penalty is immoral. As with the belief that the death penalty is unfairly applied, a large majority of Abolitionists (72%) consider the death penalty immoral. However, among non-Abolitionists, only a small minority (16%) agree that the death penalty is immoral. There is a 56-point spread between Abolitionists and non-Abolitionists regarding the belief that the death penalty is immoral. Thus, knowing that a person considers the death penalty immoral provides a powerful predictor of whether or not a person is an Abolitionist


Note that as differences between groups in the bar charts increase, 'influence' (effect size) also increases. When Abolitionist status was predicted by the belief that the death penalty is applied unfairly, the 32-point spread here indicated a medium effect size (Eta Squared=.08). When Abolitionist status was predicted by the belief that the death penalty is immoral, the 56-point spread (see above) indicated a large effect size (Eta Squared=.25). 


The bar charts reported in the Findings section will give you a visual sense of how much one characteristic influences another. Eta Squared gives you a precise numeric measure of 'strength' or 'influence' or 'power' that you can compare across different bar charts in the Findings section. 


Statistical significance is reported as a convention here, simply because some readers will expect to see such statistics. More relevant to this study is the Eta Squared for each bar chart that shows how one characteristic of a person (sadism, for example) may 'influence' another characteristic of that person (willingness to personally execute a convicted murderer, for example).


The prediction above is not perfect. Among those who support the use of 1 or more methods of execution in the U.S. today (non-Abolitionists), 16% still consider the death penalty immoral. Likewise, 27% of Abolitionists who oppose all methods of execution still consider the death penalty moral. 


In order for this belief (the death penalty is immoral) to perfectly predict Abolitionist status, 100% of Abolitionists would have to view the death penalty as immoral; 0% of non-Abolitionists would view the death penalty as immoral. Then and only then would the predictor (the death penalty is immoral) account for 100% of the variance in Abolitionist status. Only then would Eta Squared equal 1.00. 


NEXT>