PlagScan's Author Metrics is designed to support professors, teachers and students in their daily quest to ensure the best education possible.
By adding Stylometry Analysis to PlagScan’s originality suite, safeguarding Academic Integrity as well as educating critical thinkers and creative writers can be taken to the next level.
The tool incorporates the latest research from Computational Linguistics. By analyzing different writing style measurements and patterns, the software can be used to verify the authorship of a document. The tool also allows one to make basic assumptions about the author's writing style.
PlagScan’s writing style analysis solution can be used in multiple ways. A few use cases are described below.
Peer Group Cross-Check:
Find the ones that stick out, then foster them individually.
The education of students is usually organized into groups, courses and classes. No matter on what kind of principle they are organized around, there is usually a common denominator/base when in comes to the writing style of each individual (i.e. age, education level, field of interest, etc.). Computational linguists call this the ‘peer-group similarity hypothesis.’
This is where the new feature, Author Metrics, comes into action: the software analyzes each assignment for a specific group, then compares the results to each other.
By getting more details on specific writing style characteristics like sentence length, formality, and overall writing style, educators and students can work together on developing the student’s writing style.
On the other hand, if a student’s text stands out from their peer group across many measures, it could be that the student has not written it.
Individual Analysis:
Become a better writer and get your message across.
A person's writing style is part of their personality, but it can also be influenced and improved. By tracking and analyzing the writing style of an individual, educators can work together with students on further developing academic strengths and improving upon weaknesses.
For more details on Ghostwriting detection and Stylometry, please also refer to following articles on the PlagScan blog:
Our research has evaluated more than 50 candidate metrics and found that those nine described below together are best suited for our analysis. Please note this section is still work in progress.
Sentence Length
This feature evaluates the average number of words per sentence in a text.
Reading Level
This feature evaluates the number of years of education needed to understand a text on first read. This value is also known as 'Gunning Fog'.
Formality
This feature evaluates how formal a text is.
Proportion of Rare Words
This feature evaluates the percentage of words that only appear once in a text. This value is also known as 'Hapax Legomena'.
Speech Diversity
This feature evaluates part-of-speech type-token diversity calculated by the 'Simpson's Index of Diversity'.
Punctuation Frequency
The average count of commas per 1,000 words in a document.
Character Distribution
This feature evaluates how distant texts are from each other based on characters.
Lexical Originality
This feature evaluates the percentage of unique words in a document out of the total number of words in a set of texts.
Word Length Vectors
This feature compares the count of word lengths between documents.