A key claim in the study of emotions is that anger makes people less responsive to risks, whereas fear makes people more responsive. Although risk is a fundamental concern in the area of military conflict, no studies have directly tested whether anger and fear moderate the impact of risk on public support for war. We test this key claim with casualty risks as our case. Across five experiments (N = 4,559), utilizing well-established treatment material to vary casualty risk and induce emotions, we replicate the central finding that higher casualty risk decreases support for war. Emotions, however, do not moderate the effect of risk. These findings, combined with limitations in existing research, raise debate about the empirical robustness of the prominent emotion–risk interaction as well as widely used emotion inductions.