A popular prediction in persuasion research is that decreased ability
to process information increases reliance on peripheral cues and decre
ases reliance on central claims. This article explains why this predic
tion does not necessarily hold when processing capacity is impaired by
high arousal. Three experiments suggest that two types of processes u
nderlie arousal effects on persuasion. Arousal induces selective proce
ssing of cues that are diagnostic at the expense of cues that are nond
iagnostic-the selection effect. Arousal may also dilute the influence
of cues that are capacity demanding-the representation effect. It is t
herefore important to disentangle the diagnosticity of persuasion cues
from their processing demands.