The too-much-of-a-good-thing (TMGT) effect occurs when ordinarily beneficial antecedents cause harm when taken too far. Statistically, this means that the seemingly monotonic positive relationship between the “ood thing”and the desired outcome ultimately reaches an inflection point after which the relationship turns asymptotic (no additional benefit), or even negative (undesirable, such as decreased performance), resulting in an overall pattern of curvilinearity. Although empirical evidence for the TMGT effect continues to grow, demonstrating curvilinear effects still proves to be a challenge. The current proposal aims to contribute to this vivid area of research by focusing on methodological issues that may have restrained the power to identify such curvilinear effects. Specifically, we focus on the way in which personality and leadership are typically measured in organizational research (i.e., dominance measurement models using Likert-type rating scales), and how this may contribute to measurement error and diminish the chances of finding curvilinear effects. We will investigate three innovative techniques to address these measurement problems, and related, to facilitate the identification of curvilinear relationships: a) the too little/too much rating scale (vs. Likert scale); b) ideal point item response theory (IRT) models (vs. dominance models); and c) the use of anchoring vignettes to control for response styles on Likert-type rating scales.