Skip to content

Research map

Automated essay scoring: the key papers, mapped

Automated essay scoring has been studied since the late 1990s, from early statistical systems such as e-rater and Bayesian scoring to neural networks, BERT and transformer models, with recurring comparisons against human raters and, recently, tests of ChatGPT and other language models as scorers. This map shows 40 papers on the topic from OpenAlex, grouped by what they study and linked to the papers closest to them in content.

  • 40 papers, 1998–2024
  • 2 topic groups
  • Updated

Topic groups

  • Educational Technology and Assessment21
  • Topic Modeling19

Most-cited papers

Click any circle for its details.

Bigger circle: more citationsDarker line: more similar contentColour: topic groupData: OpenAlex · November 15, 2026

// the papers

The papers, by topic group

Each group gathers papers that study similar questions. Within a group, the most-cited papers come first.

Group 1

Educational Technology and Assessment

21 papers · 1998–2024

Scoring systems and their use in assessment: e-rater, Bayesian and unsupervised scorers, comparisons with human scoring, nonnative English speakers, medical education, and AI language models, including ChatGPT on dental examinations.

  1. 1
    Exploring the potential of using an AI language model for automated essay scoring (opens in a new tab)

    Atsushi Mizumoto, Masaki Eguchi · 2023

    Research Methods in Applied Linguistics

    855citations
  2. 2
    An automated essay scoring systems: a systematic literature review (opens in a new tab)

    Dadi Ramesh, Suresh Kumar Sanampudi · 2021

    Artificial Intelligence Review

    692citations
  3. 3
    AUTOMATED ESSAY SCORING WITH E‐RATER® V.2.0 (opens in a new tab)

    Yigal Attali, Jill C. Burstein · 2004

    ETS Research Report Series

    641citations
  4. 4
    Automated Essay Scoring: A Cross-disciplinary Perspective (opens in a new tab)

    Mark D. Shermis, Jill C. Burstein · 2003

    Routledge eBooks

    515citations
  5. 5
    Automated Essay Scoring Using Bayes' Theorem (opens in a new tab)

    Lawrence M. Rudner, Tahung Liang · 2002

    Open Access Journals at BC (Boston College)

    260citations
  6. 6
    113citations

Group 2

Topic Modeling

19 papers · 2005–2024

Machine-learning models for scoring: neural and discourse-aware networks, BERT fine-tuning and transformers, string kernels, domain adaptation, two-stage learning, biased ratings, and surveys of the state of the art.

  1. 1
    A Neural Approach to Automated Essay Scoring (opens in a new tab)

    Kaveh Taghipour, Hwee Tou Ng · 2016

    Conference on Empirical Methods in Natural Language Processing (EMNLP)

    583citations
  2. 2
    Automated Essay Scoring: A Survey of the State of the Art (opens in a new tab)

    Zixuan Ke, Vincent Ng · 2019

    International Joint Conference on Artificial Intelligence

    287citations
  3. 3
    Flexible Domain Adaptation for Automated Essay Scoring Using Correlated Linear Regression (opens in a new tab)

    Peter Phandi, Kian Ming A. Chai, Hwee Tou Ng · 2015

    Conference on Empirical Methods in Natural Language Processing (EMNLP)

    178citations
  4. 4
    On the Use of Bert for Automated Essay Scoring: Joint Learning of Multi-Scale Essay Representation (opens in a new tab)

    Yongjie Wang, Chuang Wang, Ruobing Li et al. · 2022

    Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT)

    157citations
  5. 5
    Should You Fine-Tune BERT for Automated Essay Scoring? (opens in a new tab)

    Elijah Mayfield, Alan W. Black · 2020

    Workshop on Innovative Use of NLP for Building Educational Applications

    155citations
  6. 6
    Automated essay scoring with string kernels and word embeddings (opens in a new tab)

    Mădălina Cozma, Andrei M. Butnaru, Radu Tudor Ionescu · 2018

    Proceedings of the conference - Association for Computational Linguistics. Meeting

    139citations

// your topic

Map your own research topic

Type a topic, a paper title or a DOI, and the Literature Matrix builds a map like this one from real papers. It is included in Pro plans.

Sources and method

Where the data comes from

Titles, authors, venues and citation counts come from OpenAlex, an open catalogue of scholarly works (CC0). This snapshot was taken on November 15, 2026; newer papers and citations are not in it yet.

How the map is made

Each paper is read by SPECTER2, a model trained on scientific writing, and linked to the papers closest to it in content. Groups are named after the OpenAlex topic most of their papers share. How the Literature Matrix works

Spotted a problem?

A wrong detail, an important paper missing, or your own work shown in a way you would like changed? Tell us and we will look at it.

This map is a starting point, not a literature review: it shows a sample of the papers OpenAlex returns for this topic and can miss important work. Read the papers themselves before you cite them.