Answers are grounded in this paper and cite their sources.
Objective: This study was aimed to propose an alternate method of scoring the intensity of disability in a patient, which will facilitate ranking and classifying a group of patients in mutually exclusive classes along with the quantification of progress made by a patient or the effect of interventions and also tracking the path of improvement of a patient and estimating the survival curve for drawing useful conclusions. Method: A nonparametric measure of disability intensity in terms of Cos θ i for the ith patient is proposed to avoid limitations of the usual scoring of modified Rankin scale and satisfies many desired properties. Two measures of interrater agreement in terms of standard error of measurement and coefficient of variation are also proposed and their relationships with Cos θ i are detailed. Illustration with hypothetical data involving 30 patients and 10 raters was undertaken. Results: Distribution of Cos θ was close to normal. Principal component analysis of Cos θ resulted in one factor explaining 64% of variance. Inconsistent raters were identified. Coefficient of variation appears to perform better than the standard error of measurement. Split-half reliability of raters in terms of angular association was higher. Similarly, split-half reliability of disability intensity (Cos θ) was higher. Average disability intensity [Formula: see text] showed similar empirical results with rigorous method of most preferred direction of a finite number of angles. Conclusions: The proposed simple measures will help the researchers and practitioners to make meaningful analysis and draw meaningful conclusions. Future studies may be undertaken to rescale Likert data to scales having properties of interval-level measurement and study the proposed measures in detail.
No takes yet. Share an insight, caveat, or question.
Satyendra Nath Chakrabartty (2019) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: