Assessing the Dimensions of Source-Based Analytical Writing

• In Event: Multiple Lenses for Understanding Source-Based Analytical Writing Development

Abstract

A better understanding of high-quality source-based writing is of vital national interest given its increased emphasis in current educational standards and assessments. Typically, source-based secondary writing can be seen as a collection of interrelated skills.A multidimensional view of writing as a complex, strategic activity involving a number of cognitive, affective, and linguistic components (Berninger & Winn, 2006; Wagner et al., 2011; Hayes, 1984, 2012; Kim, et al., 2015) has been advanced, with researchers and practitioners using analytic rubrics to describe components of such writing (National Writing Project, 2005, 2010; Northwest Regional Educational Laboratory, 2011; Kim et al., 2014). However, there is a lack of empirical evidence that validates these rubrics’ analytic categories (i.e., organization, and content and ideas) as being distinct dimensions of source-based analytical writing. To address this lack of evidence, we examine the dimensionality of linguistically diverse secondary students’ source-based analytical writing.

The study used stratified random sampling to select 200 essays written by students from a single California school district that serves predominantly (80%) Latinx students and (73%) students eligible for the National School Lunch program. Essays were coded using an 18-item analytic framework that reliably measures distinct components of writing outlined in widely used analytic trait rubrics (National Writing Project, 2005, 2010). Next, we used statistical methods including Structural Equation Modeling (SEM) and Confirmatory Factor Analysis (CFA) to examine 1) the dimensions of text-based analytical writing for secondary students, 2) how these dimensions relate to each other, 3) how these dimensions of writing predict overall writing quality, and 4) how demographic factors predict performance in different dimensions.

Results indicate writing quality is neither purely unidimensional nor purely multidimensional; holistic and analytical views of writing quality complement each other. We find holistic scores provide a great deal of information about students’ writing, but other components of writing, such as language use and evidence use can also be seen as distinct components of students writing and may reasonably be addressed targeted writing instruction. Further, while linguistically diverse students’ writing is similar to other students in many ways, results show elements of their writing that are related to language use (e.g., sentence fluency and diction) uniquely predict their holistic writing scores. Findings can help writing researchers determine features of writing most amenable to improvement through interventions and how interventions may be adapted to linguistically diverse students.

Authors

• Jacob Steiss, University of California - Irvine

• Jenell Krishnan, University of California - Irvine

• Young-Suk Kim, University of California - Irvine