Whether the higher rates of mental hospitalization and involuntary treatment for marginal social groups are due to differential labeling or simply to the occurrence of higher rates of disorder in these groups remains unresolved. I reexamine this issue with data from the National Comorbidity Survey (N = 5,877) that allow comparisons between disturbed individuals living in the community untreated and disturbed persons who have been hospitalized or seen a professional for their mental health problems under pressure or voluntarily. Contrary to labeling theory, members of lower status groups are not consistently overrepresented among those who have been hospitalized or seen a professional against their will. Consistent with self-labeling theory, persons with greater education and those not in poverty are disproportionately present among individuals who sought treatment by choice. Additional analyses show that factors that predict service utilization are important determinants of mental health service use but do not account systematically for status disparities in hospital or outpatient treatment, especially disparities by poverty status. Although I do not confirm a central tenet of labeling theory here, the negative consequences of labeling and stigma continue to be well-supported in the literature. (Contains 4 tables and 10 notes.)