Plagiarism Detection Using Keystroke Logs
Saved in:
| Title: | Plagiarism Detection Using Keystroke Logs |
|---|---|
| Language: | English |
| Authors: | Scott Crossley, Yu Tian, Joon Suh Choi, Langdon Holmes, Wesley Morris |
| Source: | International Educational Data Mining Society. 2024. |
| Availability: | International Educational Data Mining Society. e-mail: admin@educationaldatamining.org; Web site: https://educationaldatamining.org/conferences/ |
| Peer Reviewed: | Y |
| Page Count: | 8 |
| Publication Date: | 2024 |
| Sponsoring Agency: | National Science Foundation (NSF) |
| Contract Number: | 2112532 |
| Document Type: | Speeches/Meeting Papers Reports - Research |
| Descriptors: | Plagiarism, Writing (Composition), Essays, Artificial Intelligence, Cheating, Keyboarding (Data Entry), Writing Processes, Persuasive Discourse, Adults, Prediction |
| Abstract: | This study examines the potential to use keystroke logs to examine differences between authentic writing and transcribed essay writing. Transcribed writing produced within writing platforms where copy and paste functions are disabled indicates that students are likely copying texts from the internet or from generative artificial intelligence (AI) models. Transcribed texts should differ from authentic texts where writers follow a process that includes monitoring, evaluating, and revising texts. This study develops a transcription detection model by using keystroke logs within a machine learning model to predict whether an essay is authentic or transcribed. Results indicated that keystroke logs accurately predicted whether an essay was written or transcribed with 99% accuracy using a random forest model. Authentic writing included a greater number of pauses before sentences and words, had a greater number of insertions and longer insertions, deleted more words and characters, and had a greater number of revisions than transcribed writing. Transcribers, on the other hand, produced a greater number of writing bursts because they were simply copying language. Overall, the results indicated that authentic writing is a dynamic process where writers monitor their writing and evaluate whether the writing needs to be changed if problems are identified. Transcribed writing, on the other hand is much more linear. The results may have important implications for plagiarism detection. [For the complete proceedings, see ED675485.] |
| Abstractor: | As Provided |
| Entry Date: | 2025 |
| Accession Number: | ED675615 |
| Database: | ERIC |
| Abstract: | This study examines the potential to use keystroke logs to examine differences between authentic writing and transcribed essay writing. Transcribed writing produced within writing platforms where copy and paste functions are disabled indicates that students are likely copying texts from the internet or from generative artificial intelligence (AI) models. Transcribed texts should differ from authentic texts where writers follow a process that includes monitoring, evaluating, and revising texts. This study develops a transcription detection model by using keystroke logs within a machine learning model to predict whether an essay is authentic or transcribed. Results indicated that keystroke logs accurately predicted whether an essay was written or transcribed with 99% accuracy using a random forest model. Authentic writing included a greater number of pauses before sentences and words, had a greater number of insertions and longer insertions, deleted more words and characters, and had a greater number of revisions than transcribed writing. Transcribers, on the other hand, produced a greater number of writing bursts because they were simply copying language. Overall, the results indicated that authentic writing is a dynamic process where writers monitor their writing and evaluate whether the writing needs to be changed if problems are identified. Transcribed writing, on the other hand is much more linear. The results may have important implications for plagiarism detection. [For the complete proceedings, see ED675485.] |
|---|
