Objectives: This study performed item analyses of four working memory (WM) measures and constructed the validity of the WM measures. Methods: The present paper included two preliminary studies. The 1st preliminary study included data from three hundred twenty-nine children between 4–8 (NL 248, LI 81). Nonword repetition, matrix forward, matrix backward, and sentence repetition tasks were used. Item difficulty was analyzed using item response theory and classical test theory. In the 2nd preliminary studies, 84 children aged between 4–6 performed revised working memory tasks. Item difficulty and item discrimination parameters were analyzed. Lastly, construct validity was verified through the Pearson correlation between standardized vocabulary measures and four WM tasks. Results: Most items were appropriately placed in the order of difficulty, and some items with unacceptable difficulty levels were replaced, modified, or removed for the 2nd study. Most baseline items in forum WM tasks started with above 90% of the pass rate. The 2nd study results showed that item difficulty and discrimination parameters were adequate, but item parameters for few items were higher than expected, so that those items should be removed for the final version of tasks. The high level of concurrent validity for four WM tasks was demonstrated through significant correlations with receptive and expressive vocabulary scores. Conclusion: The present study systematically reviewed items in widely used WM measures. Based on the results, WM measures are expected to be used in the clinical setting. Further work is required to verify validity constantly.
Bibliographical noteFunding Information:
This work was supported by the Ministry of Science and ICT of the Republic of Korea and the National Research Foundation of Korea (NRF-2019R1A2C1007488).
© 2020. Korean Academy of Speech-Language Pathology and Audiology.
- Item analysis
- Item response theory
- Nonword repetition
- Sentence repetition
- Working memory