Do component weighting and testing method affect time management and approaches to test preparation? A study on the washback mechanism

This study utilized Structural Equation Modeling to investigate the washback mechanism, focusing on two design aspects of an English language proficiency test: component weighting (weight assigned to different test papers) and testing methods (item format), and their washback on test preparation. Two months before taking the test, a large sample of test-takers (N 1⁄4 1000) were surveyed regarding their perceptions of the two design aspects and their test preparation activities. Their official test scores were collected when they were available. Data was analyzed to estimate the washback effects of perceived changes on test-taker time management and approaches to test preparation, and their test performance. The study found that test-takers spent more time on the papers with higher weight and less on those with lower weight. Reporting component scores seemed unable to adjust this tendency. Meanwhile, favorable perception of test validity was associated with a higher level of engagement in both desirable language learning activities and focused test preparation (drilling and cramming). This suggests that favorable perceptions may not be able to reduce negative washback, but may be able to promote positive ones. © 2015 Elsevier Ltd. All rights reserved.