The factorial structure of the Inventory of Callous Unemotional Traits (ICU) is currently under dispute. The present study aims to test the factorial structure of a Greek adaptation of the ICU by considering item keying variance and examining alternative theoretical and empirically derived models. Additionally, it aims to investigate the nomological network of the ICU subscales, after controlling for item keying variance. The sample consisted of 1536 Greek-Cypriot adolescents, who completed a battery of questionnaires, including the ICU. Results showed that the consideration of item keying variance improved the overall fit of all the examined models and led to significant changes in the predictive validity of the subscales, while method factors presented distinct patterns of associations with external variables. Overall, results suggest that ICU is contaminated by item keying variance, which can be filtered out to provide clinically useful insight into the factorial structure of the ICU.