When a test-taking strategy is better? An approach from the paradigm of scheduling under explorable uncertainty

Abstract

[EN] In this article, we adopt the paradigm of scheduling under explorable uncertainty to explore test-taking strategies to solve standardized tests in terms of maximizing the correct questions answered. From this approach, a test taker considers a number of questions and has the possibility to read in order to obtain the difficulty of the question. Later, he/she has the option, for example, to answer the question or to skip the one that seemed difficult and read the next question in the test. Specifically, we state the problem definition by considering two test-taking strategies, formulate and implement a mathematical model, and generate computational experiments in order to determine the dominance of one strategy over another. The results show that the dominance depends directly on the design of the test and the maximum time to perform it, so knowing these parameters allow us to provide algorithmic insights to address this problem.