Related Experiment Videos
Rasch modeling of revised token test performance: validity and sensitivity to change.
William Hula1, Patrick J Doyle, Malcolm R McNeil
1VA Pittsburgh Healthcare System and University of Pittsburgh, Pittsburgh, PA 15206, USA. william.hula@med.va.gov
Journal of Speech, Language, and Hearing Research : JSLHR
|March 15, 2006
Summary
This study validated the Revised Token Test (RTT) for stroke survivors. Rasch analysis showed scores detect group differences and change over time, though traditional scores also showed group differences.
Area of Science:
- Neuroscience
- Psychometrics
- Speech-Language Pathology
Background:
- The Revised Token Test (RTT) is used to assess language deficits in individuals with aphasia.
- Evaluating the psychometric properties of assessment tools is crucial for accurate clinical interpretation.
Purpose of the Study:
- To examine the validity of the 55-item Revised Token Test (RTT).
- To compare traditional and Rasch-based scoring methods for detecting group differences and changes over time in stroke survivors.
Main Methods:
- Administered the 55-item RTT to 108 left- and right-hemisphere stroke survivors.
- Applied Rasch analysis to the RTT data and compared traditional vs. Rasch scores using ANOVA and relative precision index.
Main Results:
- 48 items showed acceptable model fit; Rasch model explained 71% of response variance.
- Both scoring methods detected significant group differences; only Rasch scores detected significant change over time.
- Relative precision analysis did not confirm a significant advantage for Rasch scores in detecting change.
Conclusions:
- Findings support the RTT's general validity, but some items may measure a different construct.
- Within-subtest item difficulty variations warrant further investigation for impaired language performance.
- Rasch scores show a potential, albeit equivocal, advantage in detecting longitudinal changes in stroke survivors.