JOINT INTERNATIONAL IMEKO TC1+TC7+TC13 SYMPOSIUM: Foundational Imperatives for Measurement with Mathematical Models

摘要

A contemporary measurement problem is proliferation of mathematical models that purport to have objective measurement properties but without demonstration of meaningfulness or efficiency. This presentation emphasizes broad prevalence of this problem across nonphysical observations. A rigorous standard of practice was proposed that imposes explicit, well-established measurement principles on interpretations of mathematical measures. Principles proposed are objectivity, invariance, linearity, precision, and simplicity. Function of these principles during measurement was demonstrated during construct consolidation of scales calibrated separately for measuring health-related quality of life (HRQOL) and Spirituality, respectively, with a probabilistic measurement model. Data submitted for measurement consisted of questionnaire responses from 545 breast cancer patients to FACIT-General (28 items), which assesses general health HRQOL. Spirituality was assessed with FACIT-Sp (12 items). Separate and co-calibrated measures were compared, as well as ethnic differences (White, African-American, and Hispanic). Principal Component Analysis examined threats to dimensionality after co-calibration. In general, objectivity, invariance, and linearity demonstrated instrumental importance for establishing generality of measures, while precision and simplicity contributed to efficiency and effectiveness. Altogether, conformity to these principles established empirical foundations for asserting material grounds of a conceptual entity.