Vocabulary List Generator: A digital tool to generate frequency-based word lists adjusted for dispersion

• Type of Material: Article
• Authors: Cox Ashleigh, Georgia State University; Dixon Daniel, Georgia State University; Tulay Dixon, Emory University
• Language: English
• Date: 2025
• Publisher: Elsevier
• Publication Version: Preprint (Prior to Peer Review)
• Copyright Statement: © 2025 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
• Final Published Version (URL): https://doi.org/10.1016/j.rmal.2025.100180
• Title of Journal or Parent Work: Research Methods in Applied Linguistics
• Keywords: dispersion; corpus linguistics; vocabulary; English for Specific Purposes; corpus tools

Abstract

In second and foreign language teaching, helping students learn new vocabulary is an important goal, and many researchers aim to help teachers determine which vocabulary items should be prioritized by making predictions about the words that specific learner populations are likely to need. Such predictions can be made by examining the words that frequently appear in corpora that are representative of the learners’ target language use domains. While early vocabulary lists tended to consider only frequency and range when ranking important words in a domain, more recently, researchers have argued for more robust measures of dispersion, including evenness (i.e., how evenly spread a word is across texts) and pervasiveness (i.e., the proportion of texts using a word, which is similar to range) (Egbert & Burch, 2023). However, the current tools and methods that are commonly used to form corpus-based vocabulary lists often do not include these dispersion measures, especially the measure of evenness. To make it easier for researchers and teachers to inform language learning goals with vocabulary lists that consider both frequency and recommended dispersion measures, this report describes and demonstrates a new digital tool, the Vocabulary List Generator. The tool was piloted with a corpus of psychology journal articles, and the vocabulary list generated from the corpus demonstrates its use and output.