Example of item selection procedure for final version of MFQ
Note. Figure shows how much is gained, in terms of correlations with internal and external validity criteria scales, as the number of items in each subscale is increased from 1–4. The panels show analyses of all possible item combinations in the Harm Relevance subscale, plotting their correlations with the Harm judgments subscale (lower right quadrant), Harm sacredness scale (upper right), IRI-empathy subscale (lower left), and corrected item-total correlations with all three criterion scales. Letters refer to items in the Harm relevance subscale: H=Harmed, E=Emotionally, C=Cruel, W=Weak (all items can be found in the Appendix except Harmed, which was dropped because of this analysis).