The problem
One of the most pressing problems within academia in recent years has been the proliferation of 'content mills'. Students will buy completed papers from a third party and pass it off as their own work.
This project aimed to introduce new mechanisms that could complement our similarity detection and feel like another powerful tool instructors can use to find instances of academic misconduct.
What I wanted to learn
This project was a well-known area of concern for many of our power users with who we have established relationships. To validate it as a problem and find the scope of the work required to address their concerns, we decided on a multi-phase research project.
What we found
- Users are unlikely to find instances of this form of cheating immediately. It could potentially be months before they experience a problem with a paper.
- Instances of cheating can potentially lead to, at the very least, disciplinary action. It could potentially lead to legal action.
- This form of cheating is not fully understood, but all research participants were aware of instances of it happening once it was explained to them.
- Users saw this as an extension of a similarity check rather than something to stand on its own.
- Users want to make decisions themselves. We would provide the tools; they then use them to understand the submission.
What I did
- Every 'flag' we show should have its own explanatory text.
- We'd make the default state of 0 flags a positive state. There would be no empty states as that would imply the document wasn't checked at all.
- As this was considered a core part of academic misconduct, we would create the feature on both our major platforms and harmonise the language and voice used across both.
Every string on the screen was a decision
Open any one of these and I will tell you why it is worded that way.
The title doesn't assume guilt. It simply gives a prompt that the things within this page should be reviewed.