The purpose of this study was to establish the reliability and construct validity of a new behavioral instrument to measure the presence and intensity of respiratory distress for patients who are unable to self-report about dyspnea. Patient reports about dyspnea were compared to displayed behaviors in three groups of 70 patients (n = 210). Pulmonary rehabilitation patients were assessed with the respiratory distress observation scale (RDOS) after controlled exercise while hypoxemic and subsequently asked to report current dyspnea on a dyspnea visual analog scale (DVAS). Patients with postoperative orthopedic pain were evaluated with the RDOS and asked to report current pain and dyspnea. Healthy volunteers were assessed with the RDOS at rest and asked to report current dyspnea. The internal consistency (alpha) of this seven variable scale is 0.78. A positive correlation between the RDOS and DVAS (p = 0.001) was found in dyspneic patients indicating convergent validity. Significant differences were found when RDOS scores were compared between groups indicating discriminant validity. The instrument is reliable and has convergent and discriminant validity. There are clinical and research applications for this scale for assessment of patients who are unable to self-report about distress from dyspnea.