Fisher-Yates Shuffle and Linear Congruent Algorithm in the Mini Challenge of Computational Thinking Task

Abstract

Computational Thinking (CT) is one of the crucial skills in supporting the processes of problem formulation and problem solving. Further, in exploring the skills of everyone from an early age, the Bebras Bureau of Bumigora University regularly organizes the Bebras Challenge as an extracurricular educational activity for students at various educational levels. Nevertheless, both conventional and computer-based tests still face challenges, particularly related to cheating due to the sequential presentation of questions. To address this issue, the Bebras Learning Management System (LMS) was developed, featuring online testing with randomized questions to minimize cheating. The system was designed using Computer-Based Testing (CBT) and implemented with the Fisher-Yates Shuffle Algorithm and the Linear Congruent Method (LCM), which function to randomize question order and thereby reduce the possibility of cheating among students. This research employed a methodology consisting of data collection, system design, data retrieval, algorithm implementation, testing, and evaluation. The results indicate that the application of the Fisher-Yates Shuffle and LCM algorithms in question selection and randomization produced variations with different levels of correlation, namely: No Correlation (38%), Weak Correlation (29.5%), Moderate Correlation (19%), Strong Correlation (12%), and Perfect Correlation (1.5%). We conducted the testing stage at the Bebras LMS, which achieved a final performance outcome of 86%.