Purpose: To examine the psychometric properties of the Communicative Effectiveness Survey-Revised (CESR), a patient reported outcome measure of communicative effectiveness for people with Parkinson’s disease (PD) and dysarthria using Rasch analysis, a one-parameter logistic probability model based on principles of scientific measurement. Methods: 60 individuals with PD and hypokinetic dysarthria rated their communicative effectiveness on a 27-item, 4-point equal interval scale. Exploratory Factor Analysis (EFA) was run to determine unidimensionality. Rasch analysis was completed using the rating scale model. The various analyses allow researchers to examine the item-level psychometric properties of the CESR, which result in measures of validity, reliability and sensitivity. Results: The EFA demonstrated a single linear construct (confirmed unidimensionality) which allowed Rasch analysis to proceed. Results indicated: the theoretical item difficulty hierarchy established a priori had a strong positive correlation with the CESR item difficulty hierarchy, indicative of construct validity; the 4-unit rating scale comprised equal intervals; the match between person ability and item difficulty was near the ideal range, indicative of content validity; person reliability (comparable to Cronbach’s alpha) was strong; 4) the CESR separated respondents into 4 statistically significant ability levels, indicative of respondent reliability and sensitivity of the measure; and no ceiling or floor effects existed, although participants demonstrated a wide range of ability. Conclusions: The CESR demonstrated strong item-level psychometric properties which were indicative of validity, reliability and sensitivity of the scale for use with individuals with PD and dysarthria. Further testing must be completed to determine cutoff and clinically meaningful difference scores.
Purpose: To examine the psychometric properties of the Communicative Effectiveness Survey-Revised (CESR), a patient reported outcome measure of communicative effectiveness for people with Parkinson’s disease (PD) and dysarthria using Rasch analysis, a one-parameter logistic probability model based on principles of scientific measurement. Methods: 60 individuals with PD and hypokinetic dysarthria rated their communicative effectiveness on a 27-item, 4-point equal interval scale. Exploratory Factor Analysis (EFA) was run to determine unidimensionality. Rasch analysis was completed using the rating scale model. The various analyses allow researchers to examine the item-level psychometric properties of the CESR, which result in measures of validity, reliability and sensitivity. Results: The EFA demonstrated a single linear construct (confirmed unidimensionality) which allowed Rasch analysis to proceed. Results indicated: the theoretical item difficulty hierarchy established a priori had a strong positive correlation with the CESR item difficulty hierarchy, indicative of construct validity; the 4-unit rating scale comprised equal intervals; the match between person ability and item difficulty was near the ideal range, indicative of content validity; person reliability (comparable to Cronbach’s alpha) was strong; 4) the CESR separated respondents into 4 statistically significant ability levels, indicative of respondent reliability and sensitivity of the measure; and no ceiling or floor effects existed, although participants demonstrated a wide range of ability. Conclusions: The CESR demonstrated strong item-level psychometric properties which were indicative of validity, reliability and sensitivity of the scale for use with individuals with PD and dysarthria. Further testing must be completed to determine cutoff and clinically meaningful difference scores.
INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSION
