Individual differences in emotional reactivity emerge early in development and predict important child outcomes. Unfortunately, methods used to assess these often fail to tap dynamic changes in emotion, obscuring nuanced relationships between maladaptive emotional reactivity and early internalizing psychopathology. We therefore explored the utility of an emerging, multimethod approach for examining children's emotional reactivity. Thirty-nine children (22 girls; M[subscript age] = 7.19 years, SD = 0.76) viewed 11 video clips eliciting happiness, sadness, or fear. We used multilevel growth curve modeling to estimate change in children's self-reported and observed emotion across clips of increasing potency. Higher anxious/depressive symptoms predicted steeper trajectories of child-rated happiness and lower happiness and fear toward low-intensity clips. Higher depressive symptoms predicted lower child-rated sadness toward a low-intensity clip. The limited associations between children's symptoms and emotion ratings averaged across clips suggested that modeling child emotion across stimuli of varying intensity yields a valuable index of emotional reactivity.