Bulletproofing Online Assessments
Assessments that are properly designed and implemented indicate the degree to which a student met the related learning outcome for a course. For some assessment types, such as research papers, the absence of invigilation does not impact the validity of the assessment. Other assessment types that are open to easy methods of cheating may not yield a high confidence level in the exam results.
Choosing the Assessment Type
The types of assignments students say they are least likely to cheat on are reflections on practicums, vivas, personalized and unique tasks, and in-class tasks (Bretag et al., 2019).
The following assessments types are easy to implement online and have inherent or configurable settings that can reduce cheating.
Assignments where students have to prepare a presentation and present it, in person or virtually, can reduce the potential to cheat because the instructor can see the presenter.
“Must post first” discussions
When assessments include a discussion post, students often wait for other students to post so they can see the answer, or get ideas for their own answer, before posting. Selecting “must post first” requires that students start a thread before they can read and reply to other students’ threads. This makes the post more indicative of the student’s knowledge and skill related to the discussion topic.
The following information is relevant to any assessment including final exams, mid-terms, tests, and quizzes.
Open-ended quiz questions
Open-ended questions, especially those that require critical and creative thought, make cheating transparent because individual responses are rarely identical. Long answer and essay questions are examples of this.
Close-ended quiz questions
When students are given defined choices for the answer to a question, it is easier to cheat. A student can look up the answer quickly or ask a friend, “what is the answer to #5”, with a quick response of “B”. There is no way to tell that this student cheated as the answer is not unique, but there are ways to increase the integrity of close-ended questions.
Increasing the Integrity of a Quiz in an LMS
For online exams, tests, and quizzes, instructors should follow the same principles of assessment and question design as for face-to-face exams.
When creating questions for your quiz in Brightspace, do one or more of the following to increase the integrity of the quiz:
- Create unique quizzes for each student by generating a random subset of questions from the Question Library. In your Quiz (Quiz Questions page) create a Question Pool and enter the number of questions that you want for the quiz and the questions or sections that you want to draw from.
- Make a unique presentation of the same quiz by using Sections to randomize (shuffle) the questions’ order as they appear to each student.
- Randomize (shuffle) the response options for a question each time it is selected for a quiz.
- Avoid using “all of the above” or “none of the above”. Students need only identify more than one correct option or eliminate a single option to find the correct answer. In addition, randomizing the response options will not work when “all of the above” or “none of the above” are present.
- Include response options that require a high level of knowledge and thought to determine the correct answer versus response options where there is one obvious answer among implausible options.
- Create sections in Brightspace that contain questions based on topic, question type, difficulty level, etc. Then, create a quiz that randomly selects one or more questions from each section. This way, each student will receive a unique, but fair, quiz.
Change the question type
- Instead of asking questions that test lower order thinking such as recalling facts or recognizing items, write a question or present a problem to which a student must apply their knowledge in order to respond.
- Instead of using a multiple choice (MC) type question, use a written response (WR) type. Include the response options in the question and ask students to type in their response and a written explanation for choosing their response.
- Choose a WR question to allow students to solve a problem on paper and upload a picture of their solution. Asking students to sign the document or include a picture of their student ID is optional.
- Choose an arithmetic type of question when variables are required, resulting in a unique question for each student.
Configuring a Quiz to Reduce Cheating
Setting features of the Brightspace quizzes tool to decrease the ability to cheat:
- Paging prevents moving backwards through pages – students cannot go back to find an answer for a friend
- Shuffle questions at the quiz level
- Shuffle questions within a section
- Disable right-click to prevent students from printing quiz questions (this feature is browser-dependent)
- Disable Instant Messenger and alerts to prevent students from using the Messenger tool or alerts while taking a quiz
- Use submission views strategically (see below)
In the last section under submission views, we need to be careful not to include steps on how to do this, but advice on the options.
Submission Views display information for the user after they complete and submit a quiz. These views can provide valuable feedback to students but using them wisely is important as they can inadvertently provide information for cheating.
The Default View is useful for graded quizzes and includes the following settings:
- Displays immediately after submission
- It does not show the questions (or answers)
- Displays the user’s attempt score and overall attempt score (if more than one attempt was allowed)
- User statistics are not displayed
Create and use an Additional View if you want to customize a view for a specific type of quiz such as a practice quiz where feedback is important and cheating is not a factor. Options include showing questions that are answered incorrectly or correctly, the attempt score and overall attempt score, and class statistics.
- Bretag, T., Harper, R., Burton, M., Ellis, C., Newton, P., van Haeringen, K., . . . Rozenberg, P. (2019). Contract cheating and assessment design: exploring the relationship. Assessment & Evaluation in Higher Education, 44(5), 676-691. doi:10.1080/02602938.2018.1527892
- Deakin University. (2020). Ensuring academic integrity and assessment security with redesigned online delivery. http://dteach.deakin.edu.au/wp-content/uploads/sites/103/2020/03/DigitalExamsAssessmentGuide1.pdf