OBJECTIVE Shoulder subluxation after cerebrovascular accident has been associated with chronic pain, orthopedic complications, peripheral nerve damage, and autonomic dysfunction. Clinical diagnosis and gradation of subluxation is problematic due to the lack of precision in frequently used measures. The objective of this study was to assess the validity of clinical techniques used to assess inferior subluxation of the hemiplegic shoulder by comparing these techniques with radiographic measurement. METHODS In 20 male subjects with hemiplegia, the presence, type, and degree of subluxation was assessed with three clinical measures: palpation, arm length discrepancy, and thermoplastic jig measurement. Anterior-posterior X rays of the hemiplegic shoulder were taken after clinical examination. RESULTS Spearman rank correlation coefficients between the X rays and the three clinical measures were relatively low. Palpation had the highest correlation (rs = .76), followed by arm length discrepancy (rs = .46), and thermoplastic jig measurement (rs = .42). CONCLUSIONS These findings provide cautious optimism about using these clinical measures to identify subluxation. Although detection was best with palpation, the likely inability to determine clinical overcorrecting of subluxation makes use of palpation alone suspect. Improved techniques of arm length measurement may provide a solution to this problem. These findings further necessitate that improved procedures for clinical assessment of subluxation be developed.