periments to study human decision-making. Due to the increased computational power available to researchers and the possibility of implementing online experiments with mobile applications, there is a wide selection of experimental designs for these tasks. Design optimization (DO) methods can help researchers select appropriate experimental designs and guide decisions through an objective analysis of the experiment structure. In this work, we apply DO methods to three different contexts of MABT-based experiments and evaluate their effects on the data collection process and subsequent analysis. In particular, we utilise DO methods to select the optimal time to administer an MABT-based experiment to a subject based on their adherence history and self-reported intensity of symptoms. For this purpose, we employ control chart approaches to identify unusual values of the symptoms and then combine DO methods to include information on the subject’s adherence. Our approach results in better data quality than a naive approach for both simulated and real subjects, especially for cohorts substantially impacted by adherence. Moreover, we investigate the optimal reward probabilities of a two-armed bandit task in both a static and an online experiment. We utilise an inhomogeneous MCMC algorithm to analyse the design variable space and select an optimal design: our results show that this method identifies designs that improve parameter recovery in a simulated setting, up to twice that of a random experimental design selection. We conclude that DO methods are effective in guiding experimental design choices. This work can be used to inform decisions on experimental designs for decision-making studies and contributes to the investigation of the effectiveness of DO methods in this area. These methods have the potential to reduce the patient burden in decision-making experiments, which is critical from both an ethical and data quality perspective.