Hierarchical models are an incredibly useful and widely used approach to repeated measures designs where many observations are nested within participants. In such designs, hierarchical models can optimally account for the structure of the data, thereby reducing estimation bias and improving predictive accuracy compared to simpler alternative models. However, an obvious advantage of hierarchical modeling is rarely exploited: By explicitly separating noise variability from true variability across people, it can help us improve our understanding of individual differences. In this talk, I will highlight recent projects with some of my collaborators (most of whom are graduate students) that have used hierarchical modeling to better understand the nature of individual differences in different cognitive tasks. First, I will give a brief introduction to Bayesian hierarchical modeling (collaborator: Myrthe Veenman). I will then focus on the speed-accuracy trade-off and how multivariate Bayesian hierarchical modeling can be used to account for it (collaborator: Dr. Suzanne Hoogeveen). Building on this, I will discuss whether cognitive Bayesian hierarchical modeling can improve the reliability of individual difference measures of attentional control (collaborator: Michelle Hoogeveen). Finally, I will explore whether combining Bayesian hierarchical modeling with computational modeling can help us better understand the development of mathematics skills in elementary school (collaborator: Şeyma Nur Ertekin).