How to create fraud-resistant unsupervised assessments

Guiding principles & background information

In order to give students grades that reliably represent how well they master the learning objectives (LOs) of a course, we must:

  1. Assess these learning objectives (and nothing more).
  2. Make sure that students deliver work that reflects their own level of  LO mastering.

In other words, we must prevent and detect fraud, and make sure that students can perform optimally during the exam.

Definition of fraud
This is the TU Delft definition of fraud:

“Fraud is taken to mean any act or omission by a student that makes it fully or partially impossible to properly assess the knowledge, insight and skill of that student or another student.”

Source: the model Rules and Guidelines of the Boards of Examiners, article 7.1

Fraud prevention measures versus student performance
Fraud prevention measures should not hinder student performance during the assessment. They should contribute to the quality requirements and values for assessment, respect the privacy of students, and the other eight requirements for remote assessment.

Fraud suspicions, decisions, and appeals

  • Suspicion: Examiners must be vigilant of fraud and report any suspicion of fraud to their Board of Examiners, following the procedure of their Board of Examiners (see the Rules and Guidelines that are applicable to your course)
  • Decision: The Board of Examiners (BoEx) of the student’s programme decides whether it is an actual case of fraud, and decides upon disciplinary measures for the student (i.e. it is not the examiner who decides this).
  • Appeal: Students can appeal against the decisions of the BoEx to the Examination Appeals Board (EAB, Dutch: College van beroep voor de examens, CBE), and they can appeal against the decisions of the EAB to the Council of State (Dutch: Raad van State; NB, there used to be a Higher Educational Appeals Tribunal until 1 January 2023, Dutch: College van Beroep voor het hoger onderwijs, CBHO). The examiner can be heard by the Board and Council.

Unsupervised assessments

  1. Grades should represent how well individual students master the learning objectives.
  2. TU Delft students want their degree to reflect how that they meet the final attainment levels, and they want their grades to reflect their personal performance that they can be very proud of, so they will not commit fraud.
  3. Students are more likely to commit fraud if they know that students who commit fraud will not get caught.

In unsupervised assessments, there is some level of uncertainty as to whether the delivered work is (solely) the product of the student(s) who handed it in. Examiners can take measures to ensure that this was the case. These measures are described below.

Use of (AI) tools
Lecturers are advised to discuss with their students whether and how students may use tools like ChatGPT and other sources like Studeersnel, and how students should report their use of these tools in their deliverables. Most importantly, explain the background of your standpoint. Beside AI tools, there are many other (online) tools and sources available that students can use to come up with correct answers, while they do not master the learning objectives (e.g. Studeersnel etc.).

1. Prevent fraud during the course (before the assessment)

  1. Make sure that students are prepared for the exam, so they do not feel the need to commit fraud. Have them practice with representative questions/assignments and give them feedback (e.g. model answers).
  2. Inform them of what fraud is, what the consequences are, and which measures you have taken. More information for students on fraud can be found here.


Overview of measures and their pros and cons during the assessment

Step 1: Find which fraud prevention measures are available for your assessment type (column) in the ‘summary of fraud prevention measures during assessments’ table.
Step 2: Choose balanced fraud prevention measures based on their effectiveness (column on the right) and ‘stress inducement’ (rightmost column) in the tables with ‘pros and cons per fraud prevention measure during assessments’

The fraud prevention measures during the assessment are grouped as follows:

2. Prevent fraud in question construction

3. Prevent fraud by creating unique assessments per student

4. Prevent fraud by authenticity and identity checks

5. Prevent fraud by intervening into the course of the assessment

6. Detect fraud during grading

7. What does an exam with question pools look like?

Per learning objective or topic, you will formulate a number of questions at the same levels of difficulty and of the same question type. This pool of interchangeable questions is called a question pool.

Examples of interchangeable questions in the same question pool:

  1. Fill in the blanks, automatically graded using regular expressions: Naming parts of a machine (if the answer can be copied from a book, this is only possible for proctored exams). The machine is different for each question.
  2. Short answer, automatically graded using regular expressions: Writing out the applicable formula for a situation shown in a figure. The situation is different for each question.
  3. Arithmetic question, automatically graded (all or nothing): Calculate the force on a beam in a construction. The construction or the beam is different for each question.

For each student, a unique exam will be formed with randomly drawn questions from the question pools.

Example 1

LO1: 3 question pools of unrelated questions 
o    Pool 1a: low difficulty: matching question
o    Pool 1b: low question: short open question
o    Pool 1c: medium difficulty: arithmetic question

LO2: 3 question pools of unrelated questions 
o    Pool 2a: low difficulty: open question, giving an explanation
o    Pool 2b: medium difficulty: arithmetic question
o    Pool 2c: medium difficulty: open question, analyzing a problem

LO3: 2 question pools of unrelated questions 

o    Pool 3a: medium difficulty: arithmetic question, analyzing
o    Pool 3b: difficult question: analyzing data and drawing a conclusion

Example 2 (solving engineering problems)

  1. LO1: 1 question pool that comprises questions with 5 sub questions each. The sub questions are of increasing difficulty.
  2. LO2: 1 question pool that comprises questions with 4 sub questions each. The sub questions are of increasing difficulty.

Can I randomly change the order of the questions?
The order of the exam questions should be logical, in order to enable students to perform optimally. Keep  questions on the same topic/learning objective together.