The present study integrates cognitive and affective aspects of media processing in order to make an argument for reexamining the current cold cognition perspective in multimedia research in favor of a more integrative perspective. The Cognitive-Affective-Theory-of-Learning-with-Media (CATLM) assumes that students need to become motivated to make full use of their cognitive resources. Therefore, and even though seductive details (sds) are additional interesting but unnecessary pieces of information that do not conform with the coherence principle, their possible motivational role should not be dismissed. Using a 2 x 3-experimental design, participants (N = 123) were asked to learn about biology with multimedia instruction that manipulated modality (text vs. narration) and presence of seductive details (no-sds vs. textual-sds vs. narrated-sds). Results of variance analyses show a modality effect. In addition, moderated mediation analyses with the moderator modality and mediator situational interest confirm the affective mediation assumption with the following two conditional effects. A direct detrimental effect of seductive details on learning performance under the text-condition and an indirect compensatory effect under the narration-condition were shown.