Stimulus-category confounds reduce, but do not explain, different roles of intent across moral domains: Valid statistical inference helps explain ‘inconsistent’ findings

Research overview

The text below is the source article’s original abstract under a verified reuse license. cc by

The reduced role of intent when judging impure, vs. harmful, acts provides evidence that our moral mind/brain may be composed of distinct, domain-specific systems. However, a study that manipulated and statistically adjusted for the target (i.e., self vs. other) of the act reported a significant, but notably smaller, intent × domain effect (Chakroff et al., 2013). Recent studies that either manipulated or controlled for the context, severity, weirdness, and target of the act reported nonsignificant effects (Kupfer et al., 2020; Parkinson & Byrne, 2017). We tested two competing hypotheses to explain these inconsistent findings: the intent × domain effect is a methodological artifact of the stimuli employed, or the effect is merely reduced in magnitude when controlling for potential stimulus-category confounds. To test these hypotheses, we conducted a close replication of Kupfer et al. We found a small, significant intent × domain effect, supporting the reduced-magnitude hypothesis. We find that the “inconsistent” findings can be explained by the reduced magnitude of the effect when employing carefully controlled stimuli, and by whether researchers test support for the null hypothesis, conduct sample size simulations, and perform model checking and appropriate robustness checks when modeling ordinal (i.e., Likert) items as interval data. Indeed, Kupfer et al.’s original null result becomes significant when an ordinal mixed model is fitted. Our findings are consistent with perspectives that view the moral mind/brain as constituted by multiple moral-cognitive modules. However, they are also consistent with a single moral- cognitive system with domain-dependent modulation. While carefully controlled experimental work and valid statistical inference help the replicability, robustness, and reproducibility of psychological science, delineating and decomposing the moral mind/brain will also require reforms to theory development.

Related topics