Using novel word context measures to predict human ratings of lexical proficiency

Cynthia M. Berger, Scott A. Crossley, Kristopher Kyle

Research output: Contribution to journalArticlepeer-review

14 Citations (Scopus)


This study introduces a model of lexical proficiency based on novel computational indices related to word context. The indices come from an updated version of the Tool for the Automatic Analysis of Lexical Sophistication (TAALES) and include associative, lexical, and semantic measures of word context. Human ratings of holistic lexical proficiency were obtained for a spoken corpus of 240 transcribed texts produced by second language (L2) adult English learners and native English speakers (NESs). Correlations between lexical proficiency scores from trained human raters and contextual indices were examined and a regression analysis was conducted to investigate the potential for contextual indices to predict proficiency scores. Four indices accounted for approximately 42% of the variance in lexical proficiency scores in the transcribed speech samples. These indices were related to associative, lexical, and semantic operationalizations of word context. The findings demonstrate that computational measures of word context can predict human ratings of lexical proficiency and suggest that lexical, semantic, and associative context each play an important role in the development of lexical proficiency.

Original languageEnglish
Pages (from-to)201-212
Number of pages12
JournalEducational Technology and Society
Issue number2
Publication statusPublished - 2017

All Science Journal Classification (ASJC) codes

  • Education
  • Sociology and Political Science
  • Engineering(all)


Dive into the research topics of 'Using novel word context measures to predict human ratings of lexical proficiency'. Together they form a unique fingerprint.

Cite this