Schoolboy and his classmates studying on laptops while having a class in the classroom.
For educators and test creators, improving exam integrity is an ongoing process. As new technologies become available strategies like test question randomization become more valuable in ensuring exams are fair and cheating is minimized. This is especially important when exams are given asynchronously, where one student takes a test before another student.
While it may be impossible to stop cheating completely and guarantee exam integrity, online testing platforms have made this process easier for educators and test creators. Traditionally, providing randomized test questions was a chore for educators. They would need to take the questions that they had and provide 2 or 3 different copies of an exam along with 2 or 3 different answer keys. Scoring these was cumbersome and it was easy to make mistakes in grading. With digitized testing, educators can do all of this with a click of a button, ensuring a fair and equitable testing experience for all students while giving authentic feedback to the teacher.
Test question randomization is a process where questions in a test are presented in a random order to minimize the possibility of cheating and enhance test validity. It involves using technology to rearrange the order of questions and response choices so that no two tests are the same. The randomization process ensures that each test taker gets a unique test that has the same level of difficulty as the others.
The benefits of test question randomization include increased security, as it makes it difficult for students to copy from each other, and better test reliability, as it reduces the impact of test-taking strategies. Test question randomization is commonly used in computer-based testing, where questions are presented to students in random order. However, it can also be used in paper-based tests by creating multiple versions of the test, each with different question orders.
Test question randomization can be achieved through various EdTech tools and platforms, including TAO from Open Assessment Technologies, that offer the feature of randomizing questions. These tools use algorithms to generate random sequences of questions and response choices, ensuring that each student receives a unique test.
One approach to test question randomization is to create a question bank that includes a large number of questions covering all the topics in the test. The platform algorithm selects a random set of questions from the bank to create a unique test for each student and also allows test-authors to pin specific questions that they want all students to receive. This approach ensures that the difficulty level of the test is the same for all students and that each student has an equal chance of success.
Another approach is to use item response theory (IRT) to determine the difficulty level of each question and then use the IRT scores to generate a test with equivalent difficulty levels. This approach is more complex but can provide greater precision in test construction.
In addition to randomizing test questions, some platforms also offer the option of randomizing the order of response options for Multiple-choice, Gap-match, Order, Match, Inline choice and Associate interactions, further increasing the difficulty of cheating. All of this serves to increase the accuracy of a test and provide a real look into what students can do.
Improving the integrity of exams is central to ensuring that exams provide valid and reliable data for educators to use in their lessons. The reason that we do testing and exams in the first place is to give educators an idea of where students are at and what students can do. If students take an exam with loose security and low reliability and validity, it will not give us an accurate picture of the student’s abilities.
Two ways that improving the integrity of exams is beneficial are that it helps to prevent cheating and improves the accuracy of test results.
Test question randomization and other strategies that improve the integrity of exams such as question pools, limiting feedback, and mixing question types, all serve to prevent cheating. This is especially true for exams given asynchronously, including large high-stakes national exams. With these tests, students can share answers and feedback with other students in between testing times. However, by randomizing questions and pulling from different question pools each student takes a unique test.
Having a unique test makes it challenging to cheat because students cannot simply copy down a list of letters or answers; they would need to actually copy each question in full and then hope that the other test taker had the same list of questions. All of this effort for a potentially poor result deters many students from attempting to cheat.
One drawback to this method of cheating prevention is that educators must have a large bank of questions to pull from. This may lend it to larger-scale end-of-unit or semester tests rather than quick check-ins.
Another reason why it is beneficial to improve exam integrity is that, in the end, it produces more accurate results. By removing the element of cheating from the equation, educators get a real look into student achievement and growth. Additionally, as the question pool grows and test question randomization increases, we see more valid and reliable testing data.
This is critical because educators use testing data in their classes to determine student grouping, plan lessons, and determine what pathway to take for their students. All of this leads to increased learning, which is the goal.
Adaptive testing may seem similar to randomized test questions however there is a key difference. With adaptive testing the test changes based on how a student responds to a question, leading students down many different paths with potentially many different outcomes. This is different from randomized questioning because with randomized questioning the questions are given without rhyme or reason, they are random. How a student answers one question does not impact the next question that they are shown.
A test creator may opt to use test question randomization in conjunction with adaptive testing to give students a completely customized test that is designed to pinpoint their skills and abilities. Doing this greatly improves test integrity and makes it exceptionally challenging to cheat.
—
As an online, digital, testing and learning platform TAO allows educators to completely randomize test items from an entire item bank. In addition through TAO, educators may opt to implement test or question shuffling which greatly improves test integrity, reduces cheating, and provides educators with accurate testing data on individual students.
To learn more about how Open Assessment Technologies and TAO can improve your exam integrity and provide valuable learning for your students, click here.