Abstract
Previous research on EFL and ESL learners’ writing has revealed that some advanced language learners may have difficulties in creating coherent texts. One of the essential textual elements for coherence is cohesion, which can be maintained via cohesive devices. Several succinct patterns of mistakes in learner production concerning the use of cohesive devices were diagnosed, such as run-on sentences and misuse of connectors. Related studies have examined the source of these problems and indicated that the major causes could be the misleading lists of connectors in textbooks as well as students’ insufficient knowledge of style differences between casual speech and formal writing. This requires well-designed instruction and practice to improve. Some suggestions have been offered by researchers to reinforce the teaching of conjunction; however, few of them have been examined empirically on their effects so far. With the development of Internet technologies, online instruction is advocated for advantages of self-access convenience, bilingual concordancing, rich resources of corpora, and sophisticated exercise types with immediate feedback. The study aims to investigate the effectiveness of four self-designed online conjunction units by modifying the traditional approaches and incorporating the advantages of the computer technology. The design featured textual instruction on the semantic properties and style differences of each cohesive device with corpus-based illustrations in context as well as exercises of various types (gap filling, sentence combining, run-on identification, and reformulation activity). Nineteen college freshmen of English majors from an intact writing class participated in the study in the fall semester, 2004. Before the study, a background questionnaire was given to gain information about students’ use of computers and previous learning behaviors. In the following, the online practice lasted for four weeks. A pretest, posttest, and delayed posttest three weeks after were administered to measure students’ learning of connectors in controlled contexts. Besides, the data of students’ writing before and after the instruction were collected to analyze their overall writing quality and the use of these devices in free production contexts. At the end of the study, all the participants’ perceptions about the effectiveness of the online conjunction instruction were elicited through an evaluation questionnaire. A Wilcoxon Signed Rank Sum test was run to investigate if there were any significant differences among scores in the pretest, posttest and delayed posttest. Students’ writing quality rated by the ESL Composition Profile was compared by the same statistic measurement and their conjunction errors in writing were analyzed in terms of the error rates. Qualitative observation of students’ texts was also included in the analysis. Through data analyses, it was found that learners’ use of conjunction not only improved significantly in a controlled context as demonstrated in both the posttest and delayed posttest but also in free production of writing. The conjunction errors reduced in students’ post-instruction writing and their overall writing quality improved as well. Responses from the evaluation questionnaire showed that students generally held positive attitudes toward the online conjunction units and endorsed the effectiveness of the online instruction in writing. On the other hand, they also recommended more concrete meaning explanations of conjunction and more varieties of exercise types. The results of the study supported that a web-based conjunction instruction module is able to assist EFL college students’ conjunction acquisition in both controlled and free production contexts. It helps students clarify the various meanings of each connector and perceive the important role of conjunction in English writing. With advanced computer tools, the online material design can be effective in improving students’ use of conjunctive devices and beneficial for their writing ability.