Impairments in self-assessment in schizophrenia have been shown to have functional and clinical implications. Prior studies have suggested that overconfidence can be associated with poorer cognitive performance in people with schizophrenia, and that reduced awareness of performance may be associated with disability. However, overconfidence is common in healthy individuals as well. This study examines the correlations between performance on a social cognitive test, confidence in performance, effort allocated to the task, and correlates of confidence in patients with schizophrenia and healthy controls (HC). Measures included self-reports of depression, social cognitive ability, and social functioning. A performance-based emotion recognition test assessed social cognitive performance and provided the basis for confidence judgments. Although schizophrenia patients had reduced levels of overall confidence, there was a substantial subset of schizophrenic patients who manifested extreme overconfidence and these people had the poorest performance and reported the least depression. Further, a substantial number of HC over-estimated their performance as well. Patients with schizophrenia, in contrast to HC, did not adjust their effort to match task difficulty. Confidence was minimally related to task performance in patients but was associated with more rapid decisions in HC, across both correct and incorrect responses. Performance on social cognitive measures was minimally related to self-reports of social functioning in both samples. These data suggest global self-assessments are based on multiple factors, with confidence affecting self-assessments in the absence of feedback about performance.