🔍 PaperTrail StyleMatch™  ·  $4.99/mo

This doesn't sound
like them.

StyleMatch compares a student's known in-class writing against submitted work across eight research-based stylometric metrics — the same signals forensic linguists have used for decades. All analysis runs in your browser. No text ever leaves your device.

Submitted work fills directly from the open Google Doc — one click Controlled sample pasted in by the teacher from any in-class writing All eight metrics computed locally — no text transmitted, ever Printable Authorship Consistency Report for documentation Built on peer-reviewed stylometric research, cited in every report
🔍
PaperTrail StyleMatch
Revision analysis & authorship comparison
Student Name
Emma Kowalski
Student ID
2841-K
Date
3/10/2026
Teacher Name
Optional
Course
Optional
Assignment
Optional
📄 Submitted Work 1766 words
✓ Fill from open doc
To what extent is interpretation a reliable tool in the production of knowledge? To say that all knowledge must be interpreted holds true in many regards. From the perspective of a knower, knowledge is not inherent in any object, speech, or piece of literature…
✏️ Controlled Sample 1861 words
…which that knowledge is produced and interpreted. Furthermore, it questions whether a deeper understanding is necessarily a better one. Context, here, extends beyond simply referring to background information and can be understood as the historical and methodological frameworks…
🔍 Run StyleMatch
?
Workflow

Two samples. Eight metrics. One report.

StyleMatch is designed around how teachers actually work — the submitted document is already open in Google Docs, and the controlled sample is whatever in-class writing you have on hand.

1

Open the submitted work

Open the student's submitted Google Doc. Click Fill from open doc in the StyleMatch sidebar — the submitted work populates automatically from the active document.

No copy-pasting. No file uploads. One click fills the field directly from the document you're reviewing.

2

Paste the controlled sample

Paste any in-class writing you have from the same student into the Controlled Sample field — a timed in-class essay, a Google Classroom response, a previous assignment.

The controlled sample is the anchor. The closer in genre to the submitted work, the stronger the signal on function words and discourse markers.

3

Run StyleMatch & read the report

Eight metrics compute instantly in your browser. A printable Authorship Consistency Report is generated with divergence ratings, the raw numbers, and plain-language research context for each metric.

The report is designed to be documentable — something you can save, print, or share with a department head before a student conversation.

The Research Basis

These aren't invented signals.
They're established science.

Every metric in StyleMatch is drawn from peer-reviewed stylometric and computational linguistics research. These are the same signals used in authorship attribution studies, forensic linguistics, and academic integrity research. The citations appear in every report we generate — not as decoration, but because you should know what you're looking at.

Stylometric analysis has been used in academic and legal contexts for decades. Function words in particular — the the, and, of, in that writers use without thinking — are among the most stable and reliable authorship signals in free prose precisely because no one writes them deliberately.

Grammatical Function Words
Burrows' Delta — the gold standard of authorship attribution
Burrows, J. "Delta: A Measure of Stylistic Difference." Literary and Linguistic Computing, 2002. Applied over 30 core function words. Lower Delta = more similar authorship.
Punctuation Fingerprint
Punctuation habits are largely unconscious and writer-stable
Koppel, Schler & Argamon. "Computational Methods in Authorship Attribution." JASIST 60:1, 2009. Comma, semicolon, em-dash, colon, !, ? rates per 1,000 characters.
Flesch-Kincaid Grade Level
Reading complexity is relatively stable across writing tasks for the same writer
Kincaid et al., "Derivation of New Readability Formulas." Naval Technical Training, 1975. A gap greater than 3 grade levels between samples is worth examining.
Discourse Marker Profile
How a writer structures argument reflects individual rhetorical habit
Hyland, K. Metadiscourse. Continuum, 2005. Delta over 29 connectives and stance markers — strongest signal for same-genre comparisons.
🔍 PaperTrail StyleMatch — Authorship Consistency Report
Controlled Sample: 1,861 words  ·  Submitted Work: 1,766 words  ·  Generated: 3/10/2026, 2:14 PM
Emma KowalskiStudent Name
2841-KStudent ID
IB English HLCourse
Stylometric Metrics — Side by Side FK Grade & Punctuation most stable · Function Words strongest same-genre
Grammatical Function Words
Top: of, in, that, for vs Top: the, and, is, as
18% Divergent
Δ 2.41 · Burrows' Delta over 30 core function words. Used unconsciously — among the most reliable authorship signals in free prose.
FK Grade Level
Grade 10.4 vs Grade 14.9
34% Divergent
Grade gap: 4.5 levels — above the 3-level threshold. Submitted work reads at a substantially higher complexity level than the controlled sample.
Discourse Marker Profile
because, since, thus vs however, furthermore, additionally
31% Divergent
Δ 1.87 · Submitted work relies on however/furthermore/additionally; controlled sample uses causal connectives. Same-genre — this divergence is notable.
Punctuation Fingerprint
C 6.1‰ · S 0.1‰ · D 0.0‰ vs C 11.4‰ · S 0.4‰ · D 0.6‰
67% Notable
Comma rate nearly double in submitted work; dashes absent in controlled sample but present in submission. Punctuation habits are largely unconscious.
Sentence Rhythm (SD)
SD 8.1 vs SD 13.7
61% Notable
SD gap: 5.6 words — submitted work shows considerably more varied sentence rhythm than the controlled sample.
Avg Sentence Length
21.4 w/s vs 22.9 w/s
84% Similar
Avg gap: 1.5 words/sentence — consistent with same-author variation.
Type-Token Ratio
31.2% vs 28.8%
86% Similar
Vocab range gap: 2.4% — comparable vocabulary density across both samples.

Report generates instantly in the sidebar.
Print or save as PDF for documentation.

What StyleMatch Measures

Eight metrics. All of them established.

Every metric was selected because it's stable within a writer, difficult to consciously replicate, and grounded in published research. The report explains each one — so you understand what you're looking at, not just whether a number is red or green.

🔤

Grammatical Function Words

Burrows' Delta over 30 core function words — the, and, of, in, that and so on. These words are used unconsciously and are among the most reliable stylometric signals known. Writers cannot easily change their function word profile even when they try.

🗣️

Discourse Marker Profile

Delta over 29 connectives and stance markers — however, therefore, furthermore, suggests. Reflects how a writer structures argument and signals reasoning. Strong signal for same-genre comparisons; the report flags when genres differ.

Punctuation Fingerprint

Comma, semicolon, em-dash, colon, exclamation, and question mark rates per 1,000 characters. Punctuation habits are largely unconscious — most people don't know how often they use a semicolon, which makes it a surprisingly strong authorship signal.

📐

Sentence Length + Rhythm

Average sentence length and its standard deviation, separately. Average length catches obvious differences in sentence construction; SD catches rhythm — a writer who mixes short and long sentences has a signature distinct from someone who writes uniformly.

📊

Flesch-Kincaid Grade Level

Computed from sentence length and syllable count. Reading complexity is relatively stable across writing tasks for the same writer. A gap of more than 3 grade levels between samples is flagged as worth examining.

How divergence ratings work
Similar
Metric is consistent with same-author variation. No signal of concern on this dimension.
Notable
A measurable difference worth noting. May reflect genre, topic, or context — or may warrant a closer look.
Divergent
A substantial difference on a stable metric. This is the signal that warrants a student conversation with documentation.
How to read the report
No single metric is decisive. The report is designed to be read as a pattern — consistent Similar ratings alongside two Divergent ratings tells a different story than six Divergent ratings.
Genre matters. Function words and discourse markers are strongest when both samples are the same genre. The report flags when they're not.
The report is a data point, not a verdict. It is designed to support an informed educator conversation — never to replace it.
Honest Limitations

What StyleMatch can't tell you

StyleMatch gives you a structured, repeatable data point. It's not a verdict machine, and we've deliberately designed it not to be one. Here's what to keep in mind.

📝 Genre differences inflate divergence

Function words and discourse markers perform best when both samples are the same genre. A timed in-class response compared against a polished analytical essay will show divergence on these metrics regardless of authorship. The report flags this — but you need to account for it.

Use in-class writing of the same genre as your controlled sample when possible. FK Grade and Punctuation are more genre-stable.

📏 Short samples reduce reliability

A minimum of 500 words per sample is required. Below that, statistical noise dominates. Even above 500, shorter samples produce wider confidence intervals — a 600-word sample is less reliable than a 1,200-word one.

Longer controlled samples produce more reliable baselines. Multi-paragraph in-class responses work better than short answers.

⚖️ Divergence isn't authorship proof

A Divergent rating means the metric differs substantially between samples — not that the student didn't write it. Heavy editing, collaborative revision, significant time between samples, or genuine writing development can all produce divergence.

The report is a starting point for a conversation, supported by Inspect revision data. Never use it as the sole basis for an academic integrity decision.

When StyleMatch raises a question
you want to go deeper

StyleMatch gives you eight metrics and a printable report. When you want AI-assisted sentence-level analysis — or follow-up questions to use in the student conversation — Verify takes it further. Both tools are in the same extension.

🔒

No text transmitted

All eight metrics compute entirely in your browser. Student writing never leaves your device — not to PaperTrail, not anywhere.

👤

No student accounts

Students never interact with PaperTrail. No logins, no sign-ups, no data of any kind collected from students.

📋

Client-side only

StyleMatch is pure browser computation. The only thing that leaves your device is the report you choose to print or save yourself.

🏫

FERPA & PIPEDA conscious

Designed with North American school privacy requirements in mind. No advertising, no data resale, no third-party tracking.

Start with Inspect free. Unlock StyleMatch when you're ready.

Install the extension, run Inspect on any student document, and activate StyleMatch when you have an in-class sample to compare against. One extension. Three tools.

⬇ Add PaperTrail to Chrome
Inspect is always free StyleMatch $4.99/mo or $49.99/yr All analysis stays in your browser