Evaluative priming (EP) studies using face primes revealed priming effects based on the evaluative meaning of some features, such as emotional expression and group membership. Ingroup and happy faces are associated typically with positive, whereas outgroup and fearful faces are linked to negative valence. These features can influence evaluations independently or interactively. Interactive effects, with happy ingroup and fearful outgroup faces perceived as more positive than their counterparts (happy outgroup, fearful ingroup), support the social message account (Paulus & Wentura, 2014). However, evidence for both independent and interactive processing in published EP studies remains puzzling. In the presented studies, we investigated potential causes: First, we examined whether increasing prime task relevance, typically absent in standard EP, would promote interactive processing. Introducing a secondary task requiring attention to the primes still yielded independent effects of emotion and group. Second, we tested whether a temporal delay between group and emotion cues (using neutral-to-emotional dynamic expressions) would allow the delayed feature (emotion) to be processed alongside the pre-processed feature (group). The results again supported independent processing. Lastly, increasing the prime duration from 100 ms to 200 ms (and thereby SOA from 200 to 300 ms) provided weak evidence for an interaction: Happy ingroup faces were evaluated more positively than happy outgroup faces and fearful ingroup faces more negatively than fearful outgroup faces. These findings suggest longer presentation durations (and/or increased SOA) might facilitate interactive processing, highlighting the need for further investigation into how temporal characteristics of facial primes shape emotion and ethnicity evaluations.