Self-efficacy is an important determinant of health behavior that can be targeted for intervention. Little effort has been given to the development of valid measures for self-efficacy with medication taking for adherence research. The purpose of this study was to determine the criterion validity of the Long-Term Medication Behavior Self-Efficacy Scale (LTMBSES). Individual patient data from 6 existing adherence studies in transplant, hyperlipidemia, and AIDS/HIV patients (n = 1021) were pooled. Validity was determined by assessing the relation between the LTMBSES score and medication adherence--both self-reported and electronically monitored. A weak relationship was found between the LTMBSES score and adherence, which can possibly be attributed to a ceiling effect, caused by a too homogeneous population and/or a failure of the scale to challenge patients. Generalized Estimating Equations revealed that the total average self-efficacy score predicted reported medication adherence (p < .0001). The Receiver Operating Characteristic curve revealed the area under the curve was 0.67, indicating a significant (p < .0001), but poor predictive capability. Evidence for criterion validity of the Long-Term Medication Behavior Self-Efficacy Scale is not yet convincing. Future research should focus on: (1) validation in a population with a more heterogenous level of adherence, and (2) making the scale more challenging by referring to "always taking the medication without exception."