Effects of Self-Efficacy, Task Value, Test Anxiety, and Learning Strategies on Test Scores: A Washback Study

Washback is defined as the effects of tests on learning and teaching (Alderson & Wall, 1993). Washback has been researched in contexts such as China (Zheng, 2019) and Hong Kong (Stoneman, 2005), but only a few studies (H. Nguyen & Gu, 2020) have investigated washback in Vietnam. Moreover, all the studies in the Vietnamese context examined tests at the university level and used mixed methods or qualitative research methods. Therefore, this study aims at researching the washback of an English mid-term exam for middle school students in Vietnam using quantitative methods. This will fill an important gap in the literature of washback studies. This study adapted the washback framework of Hughes (1993), who stated that washback is mediated by participants, process, and product. Participants such as students and teachers first form attitudes towards the test, which influence the process, or the activities performed due to the test. The process then affects the product, or the outcome of learning and teaching. In this study, participants were 365 middle school students in Vietnam, and their attitudes were operationalized as their motivation, which comprises self-efficacy, task value, and test anxiety associated with the test. The process is the learning strategies students used to study for the test including metacognitive, cognitive, and memory strategies. The product is their test scores. To collect data, students' test scores and survey responses were obtained. The survey was adapted from the Motivated Strategies for Learning Questionnaire created by Pintrich et al. (1991) and the Strategy Inventory for Language Learning established by Oxford (1990). Test scores and survey responses were then analyzed via structural equation modelling. It was concluded that except for test anxiety (p < 0.43), self-efficacy and task values are variables of motivation. Moreover, learning strategies can strongly mediate the relationship between motivation and test scores (X2(8, N = 365) = 10.93, p < 0.21, CFI = 1.0, RMSEA = 0.03 (0.00 - 0.07), SRMR = 0.02).