Contract cheating signals: the essay was bought, and the file knows it
Metadata tells the first story
Word processors record the author name, the editing application, creation and edit times, and sometimes a company field from a template. A submission created by an account belonging to neither the student nor the institution, edited for twelve minutes total, and saved from a template used by an essay service deserves a question. Metadata can be explained: a borrowed laptop, a shared family computer. It cannot be ignored.
Version history is stronger than a snapshot. Platforms that keep drafts show whether a text grew over days, with deletions and rework, or arrived in one paste. Genuine writing has a messy middle. Purchased writing often has only a beginning and an end.
Style breaks inside one document
A bought essay is often assembled: an introduction from one writer, body sections from a template bank, a conclusion in the student's own hurried words. Read for the seams. Vocabulary, sentence length, referencing style, and even spelling conventions shifting between sections are the seams showing. One shift means little. A document in three distinct voices is a finding.
Compare against the student's supervised writing, not against a generic standard. A baseline from an in class task or a proctored exam gives the comparison fairness. Without a baseline, style arguments become an examiner's impression, and impressions lose appeals.
Institutions win these cases on process. Require and keep the artefacts of writing: proposal, draft, source notes, feedback responses. A student who produced the work can walk an examiner through its growth. A student who bought it can usually only discuss the finished text, and often only its surface.
Process evidence beats product arguments
Short oral follow ups, used routinely and not just for suspects, change the economics of buying work. If any submission might lead to ten minutes of questions about its argument, the purchased essay stops being a finished solution and becomes a liability the buyer must defend as their own.
Flags must trigger review, not automatic penalties, because every signal misfires: international students using editing help within the rules, students writing on shared machines, neurodivergent students whose drafting process is genuinely nonlinear. A process that cannot tell editing support from purchased authorship will punish the students least able to fight it.
Fairness is a design requirement
Publish what counts as authorised help, in plain words, before the assessment. Half the grey zone exists because the boundary was never stated. The other half is the market you are actually trying to close.
Keep reading: Proctoring flags need a human review queue, not an automatic accusation, Impersonation at test centres: the ID check is the exam before the exam.
Frequently asked questions
Can metadata alone prove contract cheating?
No. Metadata is a flag with innocent explanations, from borrowed devices to shared templates. It justifies requesting drafts and holding a discussion. Cases are decided on the combination of process evidence and the student's account of the work.
Can metadata alone prove contract cheating?
Because similarity and AI detectors answer different questions and both produce false positives. Draft history shows the work being made, which is the thing contract cheating cannot supply after the fact. Process evidence is harder to fake and fairer to challenge.
Can metadata alone prove contract cheating?
Not necessarily. Proofreading and language editing are authorised at many institutions, while having another person write or substantially rewrite the assessment is not. The institution's own stated boundary decides, which is why it must be published clearly.
Can metadata alone prove contract cheating?
Drafts with dates, source notes, and any feedback received, saved somewhere with version history. If authorship is ever questioned, the student who can show the work growing has the strongest and simplest answer.