Validating GULS: An open-access Dutch language proficiency test

Previous post-entry language assessments (PELAs) research in higher education shows that academic language proficiency contributes to academic achievement. PELAs are particularly valuable for higher education systems with minimal or no admission requirements for identifying and supporting at-risk students. However, the availability of PELAs for languages beyond English, such as Dutch, is limited. Moreover, existing Dutch PELAs and their construct validity evidence are not publicly accessible, and predictive validity analyses typically do not reach the program-specific level. Therefore, the present study introduced the Ghent University Language Screening (GULS), an easy-to-administer, free, and publicly accessible Dutch PELA. More specifically, GULS evaluates first-year students' reading comprehension in higher education. First, we confirmed the construct validity of GULS at the model and item levels and its reliability using data from the two 3-year periods 2017-2018 to 2019-2020 (N₁ = 12,527) and 2020-2021 to 2022-2023 (N₂ = 17,204). Second, we examined GULS's predictive validity for academic achievement (i.e., Grade Point Average and study success) on data over the same two periods across 16 bachelor study programs (n₁ = 8244; n₂ = 10,891), followed by predictive validity analyses for each study program across the combined 6-year period. Results demonstrate that GULS is a valid and reliable PELA for assessing Dutch language proficiency, especially for first-year higher education students who require language support to ensure equal educational opportunities. As such, GULS functions as a predictor of first-year academic achievement. We discuss the potential application of GULS in future educational research and practice given its accessibility.