Revisiting Error Analysis in English Composition

Revisiting Error Analysis in English Composition

  Recently, productive aspects of English, especially writing, have received a lot of attention in that the communicative needs of English as an international language are more than emphasized. To meet these needs and offer some insights into a university level writing class, this study was conducted to investigate contributing variables to the writing performance and to integrate learning processes as well as outcome. The analysis of the multiple regression indicates that pretest (the ability of English composition in general), reading ability, sentence types, meaning, word choice, parallelism, and connectives account for 80 percent of participants’ writing abilities. Therefore, it can be concluded that these identified factors should be vital indications of what should be taught in an EFL writing class. However, the researcher postulates that the interpretation of this study needs to be considered not merely in the light of the vague notion of statistical generalization. Rather, one should take into account the contexts that the participants of the study were involved in. In addition, the possibility of using Nvivo 2 as a main tool of error analysis (EA) research is exploited and some promises in overcoming the limitations of previous EA studies are offered.