Objective. If shared decision making (SDM) is to be part of quality assessment, it is necessary to have good measures of SDM. The purpose of this study is to compare the psychometric performance of 3 short patient-reported measures of SDM. Methods. Patients who met with a specialist to discuss possible surgery for hip or knee osteoarthritis (hips/knees), lumbar herniated disc, or lumbar spinal stenosis (backs) were surveyed shortly after the visit and again 6 months later. Some of the patients saw a patient decision aid (PDA) prior to the meeting. The 3 SDM measures were the SDM Process_4 (SDMP) survey, CollaboRATE, and SURE scale. The follow-up survey included measures of decision regret, satisfaction, and decision quality. Results. Patients in the sample (N = 649) had a mean age of 63.3 years, 51% were female, 60% were college educated, and there were more hip/knee patients than back patients (69% v. 31%). Forty-nine percent had surgery. For hips/knees, the SDMP and SURE scores were significantly associated with viewing all of the PDA compared with those who did not (P < 0.001), but not for CollaboRATE (P = 0.35). For backs, none of the scores were significantly associated with viewing all the PDA. All 3 scores were significantly associated with less regret and higher satisfaction (P < 0.001) for hips/knees. For backs, only SURE and CollaboRATE were significantly associated with less regret, and only SDMP was significantly associated with higher satisfaction. For hips/knees and backs, the SDMP and SURE scales were significantly associated with an informed patient-centered decision (P < 0.001), but this relationship was not significant for CollaboRATE (hips/knees: P = 0.24; backs: P = 0.25). Discussion. Each measure has some evidence of validity. SURE and SDMP better discriminate the use of PDAs and have higher decision quality.