Abstract
Generally speaking, grammar instruction still occupies a substantial portion of class time in secondary EFL classes in Taiwan. The way the target grammatical structures are taught is mostly output-oriented, with a focus on rule explanations from the teachers followed by immediate output practice for students. Close examination shows that the traditional output-oriented instruction assigns no role to both meaning-bearing input, which is the source for language acquisition, and instruction on input processing strategies. Motivated by the theoretical inadequacies of the traditional instruction, VanPatten developed Processing Instruction (PI), which attempts to facilitate learners' acquisition of the target form by directing more of their attention to the processing of input. PI is composed of three components - explanations of the target form, instruction in processing strategies, and structured input activities. The present study compares the effects of PI and those of traditional instruction, and also explores the potential for adopting PI in Taiwan. This study was conducted from April to November, 2000. A total of 51 first-year junior high school students from three existing classes participated in the study. An experiment using a split-plot design was administered. The three classes were randomly assigned to three treatments: PI, traditional instruction, and no instruction. Two one-hour instructional packets, based on the different natures of PI and traditional instruction, were developed by the researcher. Students' performance on the target item, English wh-questions, was assessed by means of a pretest, an immediate posttest, and a five-month delayed posttest. Interpretation and production tasks were designed to assess students' ability to comprehend and produce wh-questions. The raw and gain scores in both tasks of the three tests were submitted to one-way ANOVAs. Post-hoc Scheffe tests were carried out and Effect Sizes (ES) were also calculated for later comparisons. The immediate posttest results indicated that the processing group performed significantly better than the control group on the interpretation task and made significant gains on the production task. In the five-month follow-up, processing group vs. control group in production gain scores was the only comparison generating a significant difference. No significance could be observed between the processing group and the traditional group in either task of the two posttests. However, taking mean raw/gain scores and ES into consideration, students under PI outperformed those receiving traditional instruction in both tasks in the immediate posttest. This effect for PI also persisted in the second posttest production task. The above findings are encouraging in two ways: PI could achieve at least the same effect as the traditional instruction; PI could be beneficial for successfully affecting learners' developing system necessary for output production. To conclude, further research is needed for a thorough understanding of the effects of Processing Instruction.