Developing the Occupational Self Assessment: the use of Rasch analysis to assure internal validity, sensitivity and reliability.

Abstract

This paper reports the results of three studies that examined the internal validity, sensitivity and reliability of the Occupational Self Assessment (OSA), a self-report of Occupational Competence and Value for occupational performance and participation (Baron et al 2006). All three studies used a Rasch measurement approach to explore the psychometric properties of iterative versions of the OSA. The first study showed that the 'Myself' Occupational Competence and Values items constituted a unidimensional construct, but did not discriminate optimally between participants. The results also suggested that the two 'My Environment' scales did not contain enough items to exhibit adequate measurement properties. Next, the rating scales were changed from three-point to four-point categories. The second study examined these changes and provided evidence to support the use of a four-point Occupational Competence scale. The Values rating scale was further revised because sensitivity did not improve. The third study confirmed that the OSA items in combination have good internal validity and measure the unidimensional constructs of Occupational Competence and Values. Further, both four-point rating scales resulted in improved person separation, indicated increased sensitivity, and could be used in a consistent manner by 90% of participants with a range of disabilities from a variety of contexts. The College of Occupational Therapists Ltd.