Choosing appropriate methods and levels of scaffolding (see, e.g., Wood, Bruner, & Ross, 1976) is a crucial skill in second language instruction. The observation that too little or too much scaffolding for a task leads to an inferior learning outcome, known as the assistance dilemma (Koedinger & Aleven, 2007), has resisted quantitative analysis. However, it is now possible to take advantage of computerized tutors’ ability to precisely measure response latencies and accuracy rates to provide quantitative data to analyze the merits of different methods of scaffolding with regard to students’ performance on individual tasks. The present study describes a computer-aided language learning intervention in which 46 intermediate-level adult ESL speakers used a web-based vocabulary rehearsal program several times over the course of nine weeks. The tutor led participants in completing cloze exercises of the target words, with half of the exercises being presented with a hint in the form of a short definition of the target word and half of the exercises being presented without a hint. The results of the experiment indicate that the presence of the hint significantly increased participants’ accuracy on the task, but also significantly increased time on task. These results suggest that the form of support selected was an appropriate scaffold. However, L1 speakers of Arabic (N = 29) proved exceptional in a few ways: they expressed negative attitudes toward L2 writing tasks in general and did not show any increase in accuracy in the scaffolded condition, despite the fact that speakers of other L1s showed a very large and statistically significant improvement in accuracy in that condition. These issues may relate to Arabic speakers’ exceptional difficulties processing English orthography (Martin, 2011) and warrant future study.
展开▼