Measurement of confidence: the development and psychometric evaluation of a stroke-specific, measure of confidence

Objective: To design, develop and psychometrically evaluate a stroke-specific measure of confidence, the Confidence after Stroke Measure (CaSM). Design: Cross-sectional. Setting: Adults in the community. Participants: Stroke survivors and healthy elderly participants. Methods: Questionnaire items were generated based on the literature and qualitative interviews and piloted with expert groups to establish face validity. A 53-item CaSM was administered to stroke survivors and healthy elderly participants in the community. A second copy was posted four weeks later. Completed questionnaires were analysed for extreme responses, missing values, construct validity (factor analysis), convergent validity, divergent validity, reliability (internal consistency and temporal stability) and comparing responses according to age and gender. Results: Stroke (n=101) and healthy elderly participants (n=101) returned questionnaires. Eight items were removed that had extreme responses and large numbers of missing values. Six items had item total correlations <0.3 and were removed. A further item was removed demonstrating gender difference. An exploratory factor analysis was conducted on the remaining 38 items. A 27-item three factor solution was derived assessing Self-Confidence, Positive Attitude and Social Confidence, which explained 52% of variance. Cronbach's alpha coefficient demonstrated good internal consistency (=0.94). A test re-test on the 27 items indicated good temporal stability (r=0.85, P=0.001). Conclusion: The 27-item CaSM was a valid and reliable measure for assessing confidence in stroke survivors.