Premium
Psychometric properties of a revised version of the Assisting Hand Assessment (Kids‐ AHA 5.0)
Author(s) -
Holmefur Marie M,
KrumlindeSundholm Lena
Publication year - 2016
Publication title -
developmental medicine and child neurology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.658
H-Index - 143
eISSN - 1469-8749
pISSN - 0012-1622
DOI - 10.1111/dmcn.12939
Subject(s) - rasch model , psychology , reliability (semiconductor) , scale (ratio) , rating scale , psychometrics , goodness of fit , principal component analysis , clinical psychology , statistics , developmental psychology , mathematics , power (physics) , physics , quantum mechanics
Aim The aim of this study was to scrutinize the Assisting Hand Assessment ( AHA ) version 4.4 for possible improvements and to evaluate the psychometric properties regarding internal scale validity and aspects of reliability of a revised version of the AHA . Method In collaboration with experts, scoring criteria were changed for four items, and one fully new item was constructed. Twenty‐two original, one new, and four revised items were scored for 164 assessments of children with unilateral cerebral palsy aged 18 months to 12 years. Rasch measurement analysis was used to evaluate internal scale validity by exploring rating‐scale functioning, item and person goodness‐of‐fit, and principal component analysis. Targeting and scale reliability were also evaluated. Results After removal of misfitting items, a 20‐item scale showed satisfactory goodness‐of‐fit. Unidimensionality was confirmed by principal component analysis. The rating scale functioned well for the 20 items, and the item difficulty was well suited to the ability level of the sample. The person reliability coefficient was 0.98, indicating high separation ability of the scale. A conversion table of AHA scores between the previous version (4.4) and the new version (5.0) was constructed. Interpretation The new, 20‐item version of the Kids‐ AHA (version 5.0), demonstrated excellent internal scale validity, suggesting improved responsiveness to changes and shortened scoring time. For comparison of scores from version 4.4 to 5.0, a transformation table is presented.