LAUSR

ABSTRACT Language proficiency is assumed to play a role in achievement in tertiary education. When requirements for university entrance are low or demographic changes in the population take place, universities often employ post-admission language assessments to screen all incoming students. In this study, we will look at the predictive validity of a practical, low-stakes, web-based academic reading and vocabulary screening test. Additionally, we will investigate the predictive value of the screening test when other known predictors are controlled for. Our results confirm prior research that academic language proficiency is a small but meaningful predictor of achievement in that it can detect the students that are at-risk because of lower language proficiency. The result correlates modestly with our measure of achievement, credit completion rate, with a correlation coefficient of around .30. When other predictors are controlled for in a multiple regression analysis, demographic background variables do not seem to have a large impact on the predictive value of the language screening test. In our sample of mostly native speaking monolingual students, the academic language proficiency measure does appear to reflect students’ educational background variables. As a low-stakes instrument, it is useful as an early-alert signal, prompting further diagnosis or remedial activities.