In our field, you speak of an “at random test” when it involves the random construction of an exam. This results in a unique exam for each candidate. Currently, at random is the most commonly used digital test format. But there are still plenty of organizations that have not yet (fully) made the move from written to digital testing or are using digital (fixed) version exams. In this blog, we explain the benefits of at random testing.
In a fixed (version) exam, every candidate gets the same questions in the same order. In a fixed examination (version), each candidate receives the same questions in the same order.
However, at random is the most common digital form of testing. In the case of at random testing, questions are placed separately in a database according to a certain structure, also called a test matrix. For example, the structure of the subject matter or the book. This results in a question bank. Next, questions per topic are randomly selected from this question bank. As a result, each candidate takes a unique examination.
The questions in the database are divided into different topics. This makes it easy to see if enough exam questions are included for each topic. For example, should you need to change the questions of a particular topic due to a change in current events, you can easily look up these questions.
With at random tests, it is easy and obvious to have questions developed for each topic by multiple authors. This means a significant acceleration in the development process, improved control over the progress of the development process and a decrease in dependence on a single author. In addition, it forces authors to adhere to exam requirements. If you go outside it, there may be overlap in an exam. And not adhering exactly to the level of the exam requirement compromises the equivalence of tests. Long live that precision, but that takes some effort.
You can see exactly which exam requirements are or are not satisfactory or problematic, often during the development of the exam questions. An author has little slack and can hardly bypass exam requirements for which it is difficult to develop. Therefore, ask authors for their experiences in working with the exam requirements.
It’s a lot cheaper to maintain a few questions than to maintain entire versions of keys; why is that? If changes are needed to an exam, the exam will be temporarily placed on hold. Of course, it makes a huge difference whether one question or an entire exam is on hold. The former requires only a substitute question and the latter requires an entire exam.
By the way, setting up an at random question bank generally takes a bit more time. Think of it as a long-term investment.
Should the database still contain few questions, a small portion of the questions may be redrawn in a retake. In an at random test, questions are selected loosely from the database and automatically mixed with “new” questions. By also adjusting the order, the candidate is less likely to recognize the “old” questions in a retake. Moreover, by playing with different variables, one can put the same question in a different context, making it even less likely that the candidate will recognize the question.
Are you getting started with digital testing and want to develop an at random test? Bear in mind that it is not the intention to copy a written test to a digital version. In fact, the development of a digital test involves much more. So take the time to make good choices.
For example, what forms of questions would you like to use? What is the best relationship between knowledge, insight and application questions? Should you take other things into account in the question?
All in all, it makes sense to discuss the design of the question bank with experts well versed in the relevant field. They are good judges of how heavily an issue should weigh and how best to test one and other.
Want to know more about this topic? Please contact us for a free consultation.
If so, please contact us at 013 – 528 63 63.