3. What are its psychometric properties?
Overall, results indicate that the ALA is a psychometrically sound assessment tool built on the expressed needs and concerns of key stakeholders including people with aphasia, family members and speech-language pathologists.
Rasch analysis identified 12 questions for removal and four conceptual domains from focus groups which were not unidimensional. Therefore factor analysis was conducted to identify unidimensional sub-domains within conceptual domains, revealing sub-domains including 2 language, 2 participation, 2 environment, 2 personal sub-domains and a final sub-domain represented by a single question.
In addition, evaluation of test-retest reliability resulted in moderate to high intraclass correlations (ICC) for the ALA overall score (0.87) and individual sub-domain scores (0.66 to 0.84). Cronbach’s alpha demonstrated acceptable to high internal consistency of items. And lastly, evaluation of construct or concurrent validity resulted in significant and moderately strong correlation with measures believed to assess similar constructs including the Stroke and Aphasia Quality of Life scale-39 (SAQOL-39) (Pearson’s correlation coefficient = 0.65; p<0.001); the Visual Analogue Self-esteem Scale (VASES) (Pearson’s correlation coefficient = 0.63; p<0.001); the Communication-Associated Psychological Distress Scale of the Burden of Stroke Scale (BOSS CAPD) (Pearson’s correlation coefficient = -0.63 p< 0.001).