Abstrakt
Electronic testing has become a regular part of online courses. Most learning management systems offer a wide range of tools that can be used in electronic tests.
With respect to time demands, the most efficient tools are those that allow automatic assessment. The presented paper focuses on one of these tools: matching questions in which one question can be paired with multiple response terms.
The aim of the paper is to identify how the types of questions used in a test can affect student results on such tests expressed as test scores. The authors focus mainly on the issue of the possible increase in scores that can occur with the use of closed questions, when students, after selecting the answers to the questions they know the correct answers to, then guess the answers to the remaining questions.
The authors show how the number of distractors included in a question influences the overall test score. The data on multiple-choice and alternative- response tests are well-known.
But not much is known about matching questions. Estimating formula scores for matching-question tests is important for determining the threshold at which students demonstrate they possess the required level of knowledge.
Here the authors will compare the scores obtained for three types of closed questions: multiple choice, alternative response and matching questions. The analysis of matching assignments in this paper demonstrates that they are a useful tool for testing skills.
However, this holds only if the assignment has at least two distractors. Then the informational value of this type of assignment is higher than that of ultiple-choice assignments with three distractors.
The results currently indicate that these types of assignment are not useful if the objective of the testing is to rank students or to distinguish between very good students - and this applies even if two distractors are used. In the case of such an objective, it is better to use multiple-choice assignments.