SummaryThe Quadriplegia Index of Function (QIF) was originally developed by the authors in 1980 because the popular Barthel Index was deemed too insensitive to document the small but significant functional gains made by quadriplegics (tetraplegics) during medical rehabilitation. The QIF has now been tested on a group of 30 complete quadriplegic patients at admission to and discharge from inpatient medical rehabilita tion. Resultant scores were compared to those simultaneously obtained by the Barthel Index and the Kenny Self-Care Evaluation. The QIF was found to be more sensitive (46 per cent improvement as opposed to 30 per cent by the Kenny Self Care Evaluation and 20 per cent by the Barthel Index). The QIF was also tested for reliability. Ratings by three different nurses, working independently, were found to be significantly positively correlated for all sub-scores (p < '001). We conclude that the QIF provides a useful option in choosing afunctional assessment instrument for use with quadriplegic patients.

No abstract

482 college students served as judges in the exploration of a method of determining a rational origin on an interval scale developed by Thurstone and Jones. This method is based on Thurstone's law of comparative judgment. It not only requires the judgment of single stimuli, but pairs as well. A basic assumption is that the scale values of pairs will equal the summation of scale values of corresponding components. The stimulus dimensions of birthday gifts, traffic violations, and bizarre behavior symptoms were studied. Only gifts and violations were predicted to be additive. Each dimension was presented in high, neutral, and low context sets. Paired-comparisons and categorical-assignments methods were employed. Each judge made 210 judgments involving instances of all 3 contexts, each represented by a different dimension. Support for the zero-point model was found in all 3 dimensions. In line with prediction, gifts were most satisfactorily additive. Evidence of additivity was found for behavior symptoms, contrary to prediction. Strong context effects were found under all conditions and modes of analysis. While a small proportion of this context effect could be interpreted as a random-error artifact, in bulk it seems a psychological effect. The simplest of 4 modes of analysis used, mean ratings, was as effective as any of the more complex methods in maximizing additivity and minimizing context effects.

No abstract