The Stereotype Content Model proposes that perceivers evaluate social targets primarily in terms of their warmth (sociability and morality) and competence (skills and assertiveness), while other models (e.g., Koch et al., 2016) suggest alternative dimensions (e.g., socioeconomic status and liberal-conservative beliefs). In my research on this area I address this issue by combining theory and data-driven approaches to study the dimensions that perceivers use spontaneously when evaluating others. To achieve this, I developed text analysis methods to analyze open-ended reports of stereotypes and first impressions of social groups. In adversarial collaborations between proponents of various social cognition models, we have found that perceivers spontaneously prioritize warmth impressions, but that a wide variety of dimensions are used, including status and beliefs. More notably, we show that studying which dimensions are spontaneously salient to perceivers provides novel insights compared to traditional scales that measure only the content of social groups' stereotypes. In a more recent project, I am using these methods to understand whether these social dimensions are also spontaneously applied to objects.