Validity and reliability of the personal independence inventory

The purpose of this study was to measure the validity and reliability of the Personal Independence Inventory (PII), a check list of functional physical behaviors or activities of daily living (ADL). The PII is different than most ADL scales because it purports to measure performance instead of capability of performance and because it allows the subject to observe and record his own behavior . Four different evaluative arrangements were employed to converge on estimates of reliability and validity for the PII. Test-retest reliability coefficients ranged from .26 to .78, and percentages of agreement ranged from 33% to 84% for different PII response categories. For interobserver agreement, correlation coefficients ranged from .64 to .87 and percentages of agreement ranged from 22% to 84%. Mono-method, multi-trait comparisons for the PII yielded correlations ranging from -.62 to .90, indicating that the instrument measures two characteristically different types of functional behavior. Those are behavior involving some type of assistance and unassisted behavior. Multi-method, multi-trait comparisons of the PII and three other functional assessment methods (i.e. narrative behavioral observations, behavioral diaries, and a traditional ADL scale) showed varying degrees of external validity. For comparisons involving unassisted active behavior, correlation coefficients ranged from . 40 to .90. For those involving behavior with assistance of another person, correlation coefficients ranged from -.93 to.30. It was noted that due to a small sample size, a lack of validity criterion measures, and limited study scope, additional studies are needed to substantiate these results. The results converge to indicate that the PIT is anaccurate measurement tool, that it can be used repeatedly without distortion, and that it measures two basically different types of functional behavior by quadriplegic, spinal cord injured individuals. It also appears that the self observational measurement method used with the PII is comparatively accurate. These attributes when added to those of convenience and clarity make the PII a candidate for further development, evaluation, and possible application.