Analytic Techniques for Automated Analysis of Writing

Abstract

AbstractAnalysis of academic writing has long been of interest for pedagogical and research purposes. This involves the study of students’ writing products and processes, often enabled by time-consuming manual analysis in the past. With the advent of new tools and analytic techniques, analysis and assessment of writing has become much more time and resource efficient. Advances in machine learning and artificial intelligence also provide distinct capabilities in supporting students’ cognitive writing processes. This chapter will review analytical approaches that support the automated analysis of writing and introduce a taxonomy, from low-level linguistic indices to high-level categories predicted from machine learning. A list of approaches including linguistic metrics, semantic and topic-based analysis, dictionary-based approaches, natural language processing patterns, machine learning, and visualizations will be discussed, along with examples of tools supporting their analyses. The chapter further expands on the evaluation of such tools and links above analysis to implications on writing research and practice including how it alters the dynamics of digital writing.