Academic journal article Modern Journal of Language Teaching Methods

An Investigation of Predictive Linguistic Features (Text Length, Lexical Sophistication, Syntactic Complexity, and Cohesion) in Scoring TOEFL Ibt Writing Tasks

Academic journal article Modern Journal of Language Teaching Methods

An Investigation of Predictive Linguistic Features (Text Length, Lexical Sophistication, Syntactic Complexity, and Cohesion) in Scoring TOEFL Ibt Writing Tasks

Article excerpt

1.Introduction

Independent writing has been widely used as a measure of second language writing ability. Since, independent writing tasks give the assessment of writing performance beyond morphological and syntactic manipulation (Camp, 1993), they provide a more valid representation of writing ability compared with indirect writing assessment. Unlike indirect writing assessments, independent writing tasks motivate test takers to produce an extended writing only on their prior knowledge and experience. However, concerns have been raised about writing assessments that only contain independent writing tasks because they risk de-contextualizing the writing activity (Hamp-Lyons & Kroll, 1996). Therefore, independent writing assessments may under-represent writing proficiency (Weigle, 2002).

To address these concerns, integrated writing tasks have been proposed as a proper alternative for standardized writing tests. Like academic writing, which is stimulated by outside sources (Carson, 2001), integrated writing tasks motivate test-takers to respond to source texts presented in oral or written format. Thus, integrated writing tasks may more authentically resemble the type of writing that is integral to academic contexts in higher education (Cumming et al., 2006).

The use of both independent and integrated writing as a means to assess academic writing abilities has been adopted in the new version of the Test of English as a Foreign Language (TOEFL iBT). The goal of including both writing tasks is to enhance the authenticity and validity of English as a Second Language (ESL) writing tests (Huff et al., 2008). In terms of testing validity, the combined use of integrated writing tasks and independent writing tasks can improve overall measures of writing ability because no single task can be only reliable to predict the writing ability of a test-taker (Cumming et al., 2005).

2.Literature review

Text length, lexical sophistication and syntactic complexity, have been investigated to find whether they vary with writing quality. In order to demonstrate similarities and differences between integrated and independent writing tasks, several studies have investigated the link between the textual features and writing proficiency in the two writing tasks.

Gebril and Plakans (2009) investigated on text length in relation to the holistic scores of the integrated essays. This study demonstrated that text length has a significant effect on the holistic score; therefore, they concluded that the longer essays were evaluated more favorably by raters in this writing task. Similar studies have been conducted for independent writing tasks, indicating a strong and direct correlation between text length and writing scores (Grant & Ginther, 2000).

Cumming et al. (2005, 2006) investigated lexical sophistication through type/token ratio (TTR) and average word length in independent writing tasks. They found that higher proficiency level correlates with higher type/token ratio but not with average word length. Gebril and Plakans (2009) investigated the effect of integrated essay scores on average word length as a measure of lexical sophistication. The result illustrated that in this point of view there are no significant differences between independent and integrated writing tasks.

Compared with study on integrated writing, independent writing research has investigated many more features of lexical sophistication and their effect on the essay scores. These features include average word length (Frase et al., 1999), lexical diversity (Crossley & McNamara, 2012; Grant & Ginther, 2000), specific lexical categories, and nominalizations (Connor, 1990). They found that higher scored independent essays are related to longer words, greater lexical diversity, more frequent use of nominalizations and certain lexical categories. The researchers concluded that lexical sophistication is an important predictor of the essay scores in independent and integrated writing tasks (Guo et al. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.