Extending a Matrix Lie Group Model of Measurement Symmetries

This paper advances a Lie-group approach to measurement by identifying symmetry conditions that determine when effect sizes from different instruments can be meaningfully compared. Measurement transformations are modeled as elements of a two-parameter affine Lie group, and the associated Lie algebra describes the infinitesimal flow linking true scores and measurement-error variability across instruments. Within this framework, it is shown that the population standardized mean difference (SMD) is invariant across measures if and only if the transformation between them consists of a uniform affine transformation of true scores together with a uniform scaling of measurement-error standard deviations by the same factor. These symmetry conditions arise directly from the Lie algebra and ensure that the SMD remains constant along the exponential transformation flow; even slight departures from this symmetry produce a non-zero derivative of the SMD, marking a precise breakdown of invariance. A simulation study demonstrates how small nonlinear perturbations of the affine symmetry generate systematic distortions in the population true-score SMD. The results provide a mathematically grounded characterization of effect-size comparability and illustrate how continuous symmetries, Lie algebras, and transformation flows can clarify fundamental issues in measurement equivalence, meta-analysis, and longitudinal or cross-cultural research.