The rates of medication non-adherence among Indian patients with schizophrenia are high, and its detection poses problems. Comparisons of suitable measures to detect medication non-adherence in schizophrenia from Indian outpatient settings are scarce. This study compared simple and inexpensive methods of detecting medication non-adherence in schizophrenia among outpatients from a tertiary-care center in North India. A longitudinal comparison of two self-reports, the Morisky Medication Adherence Questionnaire and the Drug Attitude Inventory-10, clinicians' ratings employing the Compliance Rating Scale, and clinic-based pill counts was conducted among 70 outpatients with schizophrenia. The rates and detection of medication non-adherence, associations with determinants of treatment non-adherence, and agreement between measures were examined at intake and after six months of follow-up (n = 53). The self-reports had greater ability to detect medication adherence (specificity 41-65%; positive predictive values 25-51%; negative likelihood ratios 0.86-1.14) and moderate ability to detect medication non-adherence (sensitivity 27-65%; negative predictive values 49-69%; positive likelihood ratios 0.78-1.10). They yielded higher medication non-adherence rates, detected changes in medication non-adherence over time, and were associated with the other measures and the well-known correlates of medication non-adherence. Clinicians' ratings and pill counts had high sensitivity (56-90%) but low specificity (35-49%) to detect medication non-adherence. Self-reports are comparatively better screening options for detecting medication non-adherence among Indian outpatients with schizophrenia. However, the sequential use of different measures could lead to better recognition of medication non-adherence.