Statistical Tests of Percent Differences Among Execution Actors
Differences in percentage agreement* were tested across all four groups for actual sadism during executions (people should suffer while being put to death). Differences were also tested among only those groups favoring some forms of execution. These people also served on the mock death penalty jury. Differences were tested between Soft Abstract Supporters and Hard Abstract Supporters of executions.
When all four groups were compared, the differences in percentages were statistically significant. (A Sig. value of less than .05 means that differences between groups are 'significant.') This is largely due to extremely high levels of actual sadism among Executioners. The effect size, as indicated by the Phi coefficient (.35) and Eta Squared (.11),** is slightly above medium. Eta Squared=.09 is considered a medium effect size.
For the three pro-execution groups, percentage differences are statistically significant. This is due to high actual sadism among Executioners. The effect size, as indicated by the Phi coefficient (.31) and Eta Squared (.09), is medium. Eta Squared=.09 is considered a medium effect size.
When Soft Abstract Supporters and Hard Abstract Supporters of executions were tested separately, the difference was statistically significant. However, the effect size, as indicated by the Phi coefficient (.11) and Eta Squared (.01), is small.
*Everyone in the study was given a "score" on an index that measured this belief system. For each opinion, a person could strongly disagree (1), somewhat disagree (2), neither agree/disagree (3), somewhat agree (4), or strongly agree (5). If, on average, a person's score was in the "agree" range (greater than 3.0), that person was classified as "agreeing" with this belief system. If that person's score was 3.0 or less, that person was treated as "neutral/disagreeing" with that belief system. In the process, data were collaped. A score of 3.01 was treated the same as a score of 5.00. However, this made the information easier to present and understand.
**The use of Eta Squared is premised on the reasonable assumption that the typology of execution actors (Abolitionists, Soft Abstract Supporters, Hard Abstract Supporters, and Executioners) is ordered. Regarding execution actions, willingness to participate in an execution increases in a linear fashion from Abolitionists (who reject all methods of execution) to Soft Abstract Supporters (willing to sentence a person to death but not attend the execution), Hard Abstract Supporters (willing to attend the execution but not kill the person) and finally Executioners (who are willing to kill the person they sentenced to death).