Objectification occurs when a person is perceived and/or treated like an object. With the present work, we overview the available measures of objectification and present a series of studies aimed at investigating the validity of the task of inverted body recognition proposed by Bernard and colleagues (2012), which might potentially be a useful cognitive measure of objectification. We conducted three studies. Study 1 (N = 101) is a direct replication of Bernard et al.'s study: participants were presented with the same photos of sexualized male and female targets used in the original research. Study 2a (N = 100) is a conceptual replication: we used different images of scantily dressed male and female models. Finally, in Study 2b (N = 100), we investigated a boundary condition by presenting to participants photos of the same models as in Study 2a, but fully dressed and non-sexualized. Using mixed-effects models for completely-crossed classified data structures, we investigated the relationship between the inversion effect and the stimulus' asymmetry, sexualization and attractiveness, and the perceivers' self-objectification, sexism, and automatic woman-human association. Study 1 replicated the original results, showing a stronger inversion effect for male photos. However, no difference between male and female stimuli emerged in either Study 2a or 2b. Moreover, the impact of the other variables on the inversion effect was highly unstable across the studies. These aspects together indicate that the inversion effect depends on the specific set of stimuli and limits the generalizability of results collected using this method.