Statistical Considerations in the Interpretation of Functional Outcome Scales¹

Issue title: Outcome Evaluation

Guest editors: Karyl M. Hall

Article type: Research Article

Authors: Wright, Jerry

Affiliations: Santa Clara Valley Medical Center, San Jose, CA

Note: [1] The preparation of this manuscript was supported, in part, by grants (G0087C2015 and H133N00002) from the National Institute on Disability and Rehabilitation Research.

Abstract: Functional outcome scales occupy an important place in rehabilitation documentation and research. Because of the ordinal nature of most assessment data, interpretation of these measures is not always straightforward. Outcome measures are discussed in terms of the four levels of measurement (nominal, ordinal, interval, and ratio). A rationale is given for the designation of most assessment measures as being ordinal in nature. Many statistical and mathematical operations may not be appropriate for use with ordinal data. In particular, change scores, total scores, average scores, and most parametric statistics may be misleading when applied to ordinal data. Potential solutions and safeguards are commented on, including criterion scores, Rasch transformations, and nonparametric statistics.

DOI: 10.3233/NRE-1992-2405

Journal: NeuroRehabilitation, vol. 2, no. 4, pp. 21-30, 1992