A Bayesian Psychometric Item-Level Resource for Korean Early Vocabulary and Developmental Measurement
Abstract
Purpose: Parent-report vocabulary measures are widely used to study early language development, but CDI-based analyses typically summarize children’s responses using total vocabulary scores. This approach provides limited information about how individual words differ in developmental timing and psychometric function. This study develops a Korean item-level vocabulary resource that jointly characterizes developmental emergence, item response theory parameters, and input frequency. Method: Using Korean Words & Gestures comprehension and Words & Sentences production data from Wordbank, we estimated Bayesian age of acquisition (AoA) at the 50% response threshold. We then examined associations between Bayesian AoA estimates, Bayesian two-parameter item response theory item difficulty and discrimination parameters, and corpus-derived Korean input frequency. The stability of Bayesian AoA and item response theory estimates was evaluated through repeated child-level resampling. Results: Bayesian AoA and item response theory estimates were stable across repeated resampling. Words that emerged earlier also tended to be psychometrically easier, and more frequent words tended to emerge earlier and show lower latent difficulty. However, AoA and input frequency were only weakly associated with item discrimination, indicating that developmental timing and input frequency do not fully explain how informative a word is for distinguishing children along the latent vocabulary scale. Conclusion: Early Korean vocabulary items vary along partially distinct developmental and psychometric dimensions. The resource characterizes Korean CDI words by AoA, latent difficulty, discrimination, and input frequency, providing a foundation for item-level analyses of Korean early vocabulary development to relate measurement properties to theories of lexical development.
Related
- Cross-cultural adaptation and comparative psychometric evaluation of four patient-reported outcome measures for breast cancer-related lymphedema: measurement properties and clinical discrimination
- A brief measure of life satisfaction: psychometric properties and measurement invariance of the Polish three-item Satisfaction with Life Scale (SWLS-3).
- Rapid Psychosocial Screening Instruments for Cancer Patients: A Scoping Review of Measurement Tools and Psychometric Evidence
- Psychometric properties and longitudinal measurement invariance of the Functional Assessment of Cancer-Gynecologic Oncology Group-Neurotoxicity subscale in the Chinese colorectal cancer chemotherapy population.
- A Spanish-Language Version of the Emotion Reactivity Scale: Psychometric Properties, Measurement Invariance, and Criterion Validity in a Large Adolescent Sample
BibTeX
@misc{chai2026bayesian,
title = {{A Bayesian Psychometric Item-Level Resource for Korean Early Vocabulary and Developmental Measurement}},
author = {Jun Ho Chai and Eon-Suk Ko},
howpublished = {PsyArXiv},
year = {2026},
month = aug,
doi = {10.31234/osf.io/28vjw_v1},
url = {https://osf.io/preprints/psyarxiv/28vjw_v1/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - A Bayesian Psychometric Item-Level Resource for Korean Early Vocabulary and Developmental Measurement
AU - Jun Ho Chai
AU - Eon-Suk Ko
PY - 2026
DA - 2026/08/31
PB - PsyArXiv
DO - 10.31234/osf.io/28vjw_v1
UR - https://osf.io/preprints/psyarxiv/28vjw_v1/
AB - Purpose: Parent-report vocabulary measures are widely used to study early language development, but CDI-based analyses typically summarize children’s responses using total vocabulary scores. This approach provides limited information about how individual words differ in developmental timing and psychometric function. This study develops a Korean item-level vocabulary resource that jointly characterizes developmental emergence, item response theory parameters, and input frequency. Method: Using Korean Words & Gestures comprehension and Words & Sentences production data from Wordbank, we estimated Bayesian age of acquisition (AoA) at the 50% response threshold. We then examined associations between Bayesian AoA estimates, Bayesian two-parameter item response theory item difficulty and discrimination parameters, and corpus-derived Korean input frequency. The stability of Bayesian AoA and item response theory estimates was evaluated through repeated child-level resampling. Results: Bayesian AoA and item response theory estimates were stable across repeated resampling. Words that emerged earlier also tended to be psychometrically easier, and more frequent words tended to emerge earlier and show lower latent difficulty. However, AoA and input frequency were only weakly associated with item discrimination, indicating that developmental timing and input frequency do not fully explain how informative a word is for distinguishing children along the latent vocabulary scale. Conclusion: Early Korean vocabulary items vary along partially distinct developmental and psychometric dimensions. The resource characterizes Korean CDI words by AoA, latent difficulty, discrimination, and input frequency, providing a foundation for item-level analyses of Korean early vocabulary development to relate measurement properties to theories of lexical development.
ER -