Intelligent selection testing
This article has multiple issues. Please help improve it or discuss these issues on the talk page. (Learn how and when to remove these template messages)(Learn how and when to remove this template message)
The Intelligent Selection testing model is a means of testing candidates on a subset of a high-volume repository of information and attaining a high level of accuracy of assessment of the candidate.
Whereas action selection in Artificial Intelligence attempts to determine a next action based on a predefined and fixed set of possible reactions to a current action, the Intelligent Selection testing model is more similar to a learning agent in that it selects a next question in the same academic category of questions based on the correctness of answers given in a previous question of the same category. However, Intelligent Selection takes advantage of the pattern and trends collected from previous questions answered in order to determine the next selection of question provided to the exam-taker.
This approach uses sampling of data for testing purposes when it is important to gauge an exam taker's understanding of all of the subject matter material but, practically speaking, when there is only enough time to test them on a sample of the subject matter population.
A practical example of this is technical testing of a Java developer. Java software development can be defined into more than 11 categories of technical expertise and more than 40 subcategories within these Java software categories. To technically test a Java developer in this broad a topic is not practical. Intelligent selection in the case of the eScreeningz test narrows the focus of the test for Java developers while retaining the depth of Java technical testing.
The Intelligent Selection testing model would test the exam-taker with one question in a particular category. Based on how well the exam-taker scores in that question, the next question selected in that same category of assessment would be easier, harder or at the same level of difficulty based on how the exam-taker performed. In this manner, the Intelligent Selection testing model is able to assess the candidate with few questions in each category of topic against a mathematically wider range of data due to an "intelligent selection" of following questions and their difficulty range.
Multiple industry examples exist that use the Intelligent Selection testing model, including The GRE General Test and the Graduate Management Admission Test.
- Special issue: An introduction to multistage testing. Applied Measurement in Education, 19(3).