Numerous theories posit a positive relation between perceiving emotion expressed in the face of a stranger (emotion perception) and feeling or cognitively understanding the emotion of that person (affective and cognitive empathy, respectively). However, when relating individual differences in emotion perception with individual differences in affective or cognitive empathy, effect sizes are contradictory, but often not significantly different from zero. Based on 4 studies (study ns range from 97 to 486 persons; ntotal = 958) that differ from one another on many design and sample characteristics, applying advanced modeling techniques to control for measurement error, we estimate relations between affective empathy, cognitive empathy, and emotion perception. Relations are tested separately for each of the 6 basic emotions (an emotion-specific model) as well as across all emotions (an emotion-general model). Reflecting the literature, effect sizes and statistical significance with an emotion-general model vary across the individual studies (rs range from −.001 to .24 for emotion perception with affective empathy and −.01 to .39 for emotion perception with cognitive empathy), with a meta-analysis of these results indicating emotion perception is weakly related with affective (r = .13, p = .003) and cognitive empathy (r = .13, p = .05). Relations are not strengthened in an emotion-specific model. We argue that the weak effect sizes and inconsistency across studies reflects a neglected distinction of measurement approach—specifically, empathy is assessed as typical behavior and emotion perception is assessed as maximal effort—and conclude with considerations regarding the measurement of each construct.