ObjectivesIn recent years, social cognition has received growing interest in the international psychiatric and neurologic literature. Social cognition impairments are described in many different conditions and are associated with a poor functional outcome. Consequently, an accurate and valid assessment of social cognition abilities is necessary in clinical practice, so as to better understand individual functioning and define corresponding therapeutic interventions. The aim of the present study was to provide further elements of validation to the Bordeaux Social Cognition Assessment Protocol (Protocole d’Evaluation de la Cognition Sociale de Bordeaux: PECS-B) in the general population and in one with schizophrenia. MethodsA total of 131 healthy controls and 101 participants with schizophrenia or schizoaffective disorders between 18 and 60 years old were included. Sociodemographic variables (i.e., age, education level, sex), social cognition (i.e., emotional fluency, facial emotion recognition, theory of mind, emotional awareness and alexithymia with the PECS-B), neurocognition (i.e., processing speed, episodic verbal memory, short-term memory, working memory, selective and sustained attention, spontaneous and reactive flexibility), anxiety and depressive mood were assessed in both samples. ResultsResults show a good discriminative power for the PECS-B within the same population. Only “Attribution of Intention” and “Faux-Pas” tasks show ceiling effects in both samples. Structural validity is satisfactory in the general population sample and suggests that the social cognition structure consists of four factors: facial emotion recognition (1), emotional lexicon (2), explicit processing of sophisticated emotional information (3) and theory of mind (4). Structure validity is also satisfactory in the schizophrenia sample, after removal of the variable “Emotional fluency-Percentage of sophisticated words”. In this case, the social cognition structure consists of two factors: general social cognition (1) and explicit processing of sophisticated emotional information (2). Results show a good divergent validity in both populations, between close constructs such as neurocognition and anxiety/depression. Nevertheless, correlations between social cognition and neurocognition are more frequent in the schizophrenia sample. Internal consistency is satisfying in both samples. Finally, results reveal some effects of sociodemographic variables (i.e., age, education level and sex) both in the general population and schizophrenia samples. Norms are also presented for adults aged between 18 and 60. ConclusionStructural validity, divergent validity and internal consistency of the PECS-B are satisfactory in the general population and the schizophrenia one. The PECS-B presents with good psychometric qualities that permit its use for the assessment of adults’ social cognition in clinical practice as much as research.