Traditional efforts to delineate the clinical heterogeneity of schizophrenia have been unsuccessful because of the absence of a valid, stable, and meaningful subtyping scheme. A clinically-informed nosology supported by multivariate statistical classification methods may provide a better approach for classifying schizophrenia. The goals of the current study were to 1) use multivariate classification methods to validate a clinical subtyping scheme based on the profile of negative symptoms; and 2) following validation to contrast the statistically-derived subgroups to ascertain distinguishing demographic, clinical, cognitive, and functional characteristics. In the current study, 706 people with schizophrenia completed measures of positive and negative symptoms, premorbid adjustment, cognition, and psychosocial functioning. Latent class analysis served to identify the number of negative symptom subgroups in schizophrenia. Next, statistical classification methods—Bayes Theorem and the Base Rate Classification Technique—were used to assign participants into the identified subgroups. Subgroups were compared on external validation variables not used in the classification process via logistic regression and discriminant function analysis. Latent class analysis supported a three-class model of schizophrenia that included deficit, persistent, and transient negative symptom subgroups. Posthoc comparisons showed that demographic characteristics, positive symptoms, premorbid adjustment, and cognitive profiles can distinguish the schizophrenia subgroups with moderate accuracy. The deficit subgroup had the greatest impairments in psychosocial functioning and quality of life variables. Findings suggest that schizophrenia encapsulates qualitatively distinct negative symptom subgroups that differ in their demographic, symptomatic, neuropsychological, and functional profiles. Schizophrenia heterogeneity reflects a combination of non-arbitrary subgroups and severity-based differences in negative symptoms.