Deconstructing the side-effect effect: Separate inferences of intentions to harm and to help based on social information
The classic side-effect-effect (SEE) entails judging harmful side-effects more intentional than helpful side-effects. We suggest this effect reflects two distinct components, with people drawing on different social information to disambiguate motives to harm from ambiguous motives, versus motives to help from ambiguous motives. We tested this model by altering social information in several ways. Harmful intentions were impacted by manipulations reducing the diagnosticity of the CEO’s intentions by introducing external pressures (e.g., a gunman demanding they start a harmful program). This pattern suggests that perceptions of harmful intentions reflect inferences about harmful motives in the absence of external pressures. Conversely, helpful intentions were impacted by manipulations of the CEO’s stated intentions. Changing the CEO’s statement to be more pro-environmental flips perceptions from unintentional to probably intentional, eliminating the classic SEE difference between helping and harming intentions. Conversely, verbal justifications do not impact perceptions of harmful intent. Thus, judging harmful actions as intentional depends on unjustified trade-offs—in the absence of justifiable reasons to cause harm, harmful acts speak louder than words. Conversely, perceptions of helpful intent depend on verbal justifications clashing with normative expectations of positive self-presentation. Together, these results support the social information hypothesis, as changing relevant diagnostic social information about the CEO’s intentions changed people’s judgments, even to the point of eliminating the SEE.