Effects Of Strategy Use On The Integrated Reading-to-write Test Performance

Universities and test institutions usually like to employ a timed impromptu essay task to test students’ writing ability. However, recently, it is criticized for lacking authenticity and involving construct-irrelevant variance in test scores. Thus, integrated writing tasks are designed as alternatives to theses writing-only tasks. Integrated writing tasks are a kind of tasks that integrate writing ability with reading and/or listening ability. They are considered more authentic and thus allow more accurate inferences on academic writing ability. As more and more integrated writing tasks are used for accessing students’ academic writing ability in many universities and colleges, a number of studies have been implemented to explore whether the score of integrated task could better interpret test takers’ academic writing ability. Some of them centered on test takers’ written responses of integrated writing tasks; some focused on their writing processes; only a few studies addressed both aspects of integrated writing tasks. To get more information about construct of the reading-to-write task, the study was conducted to explore the construct validity of integrated reading-to-write task through investigating the nature of strategy use and the relationship between strategy use and test performance. The investigation focused on the nature of strategies and their effects on writing performance in an assessment setting.The study employed quantitative and qualitative methods to validate integrated reading-to-write task.99 EFL Chinese students were recruited to complete a reading-to-write task in a test situation and fill out a strategy questionnaire immediately when they finished the test. Then, another 10 students performed integrated task through think-aloud protocols and took a follow-up interview. Four kinds of data were collected. The research used a series of exploratory factor analyses and confirmatory factor analyses to explore the trait structure of strategy use, and employed a structural equation modeling to investigate the relationship between strategy use and test performance on integrated reading-to-write task. In addition, the data of think-aloud protocols and interviews were transcribed and coded as specific strategies to compare the use of strategies across different performance levels, Meanwhile, these data were also used to triangulate the quantitative analyses of questionnaire data and test scores.The results have shown that strategies responding to the reading-to-write task can be divided into four major types:metacognitive, cognitive, affective, and test-wiseness strategies which are represented by several specific strategies including connecting, evaluating, organizing, selecting, monitoring, planning and so forth. Except for test-wiseness strategy use, the other three strategy use types have significant correlations with each other. Metacognitive strategy use executes a test-taker’s cognitive strategy use and affective strategy use. It has a direct and positive influence on them. Affective strategy use highly correlates with cognitive strategy use. And it is also found that metacognitive strategy use has an indirect influence on test-takers’ writing performance. Cognitive strategy use has a direct and positive influence on L2 writing performance on the reading-to-write task, but the influence is negligible. Affective strategy use has a relatively significant influence on students’ writing performance, affecting it directly and positively. Test-wiseness strategy use has a significantly negative impact on participants’ test performance. Additionally, the qualitative results of TAPs and interview protocols reveal that high-performance students are inclined to employ more various and global strategies, whereas participants in low performance tend to use less various strategies and more specific strategies. And based on these verbal reports data, a tentative working model of strategy use for integrated reading-to-write task is outlined. These findings will give several implications for second language writing instruction and assessment, specially on second language academic writing. Theoretically, the study confirmed that writers not only need L2 writing ability but also reading skill and discourse synthesis to complete an integrated reading-to-write task. Practically, the study found that test-takers’ strategy use was different across L2 proficiency and test performance levels. And methodologically, the study adopted a combination method of qualitative and quantitative analysis to explore the raised research questions. SEM was applied to examine the latent structure of integrated reading-to-write task through analyzing the collected data from strategy questionnaire.