Related Experiment Videos
Improving the Action Research Arm test: a unidimensional hierarchical scale
Johanna H van der Lee1, Leo D Roorda, Heleen Beckerman
1Department of Rehabilitation Medicine, Vu University Medical Center, Amsterdam, The Netherlands. jh.vanderlee@vumc.nl
Clinical Rehabilitation
|October 24, 2002
Summary
The Action Research Arm (ARA) test functions as a single, strong scale for upper extremity motor function. A 15-item version is recommended for efficient adaptive testing in stroke patients.
Area of Science:
- Rehabilitation Science
- Neuroscience
- Psychometrics
Background:
- The Action Research Arm (ARA) test assesses upper extremity motor function through 19 items across four subtests.
- Empirical validation of the ARA test's multidimensionality was previously lacking.
Purpose of the Study:
- To empirically investigate the underlying dimensionality of the Action Research Arm (ARA) test.
- To determine if the ARA test functions as a unidimensional or multidimensional measure of motor function.
Main Methods:
- A cross-sectional study design was employed.
- Mokken scale analysis was utilized to analyze data from 63 chronic stroke patients.
- Item response theory principles were applied to assess scale properties.
Main Results:
- Mokken scale analysis confirmed a strong unidimensional scale encompassing all 19 items (scalability coefficient H = 0.79).
- High internal consistency was demonstrated (reliability coefficient rho = 0.98).
- A 15-item subset exhibited invariant hierarchical item ordering, suitable for adaptive testing.
Conclusions:
- The Action Research Arm (ARA) test is validated as a unidimensional scale, contradicting its original subtest structure.
- A 15-item version of the ARA test is proposed for efficient adaptive testing, reducing patient assessment time.
- Findings support the use of the ARA test as a single measure of upper extremity motor function in stroke rehabilitation.