Distorted claims about distortions, a response to reflections on the house of mirrors

In this response, we clarify the misunderstandings introduced by the paper “Reflections on the House of Mirrors,” concerning our article “The Psychometric House of Mirrors”. Our central claim is that caution is needed when incorporating self-report survey data into agent-based models (ABMs). Psychometric measures are not neutral carriers of information as distortions and scale properties can strongly influence ABM’s dynamics. While our correspondents suppose the possibility of using binomial distributions and “correct measuring methods,” we emphasize that distortions are inherent properties of many measurement systems, and that assumptions of independence among scale variables rarely hold in practice. Indeed, scales are usually built around the hypothesis of correlating variables. We further argue that seemingly simple solutions, such as adopting binomial models or assuming independence, require empirical testing against more complex state-of-the-art models before being recommended as best practice. Ultimately, our aim is to encourage debate and systematic study of how different measurement approaches affect ABM outcomes. We welcome further exploration but caution against prematurely endorsing single methods as definitive solutions. We further invite readers to consult our original paper for an undistorted presentation of our argument.