In rehabilitation, we need to continually assess those with disabilities, maintaining objectivity and consistency across time is challenging. In this paper, we describe our work in establishing objectivity in the widely used and accepted Action Research Arm Test on a group of actual patients. We concentrate on the grasp subtest which employs a cube into which we embed sensors. We feature a novel analysis of data by i) using singular spectrum analysis on the signals, for which the resulting eigenvalues are selected in a structured way to filtering the signals and ii) using receiver operating characteristics to set thresholds so as to objectively score the tests. These show promise to attain higher levels of objectivity in rehabilitative assessments.