Statistic analysis of prof. J. Matulionis young mathematicians contest tasks

Abstract

Herein authors disclose the results of prof. Jonas Matulionis young mathematicians contest task’s statistic analysis. Research sums 160 tasks (one task contains 5 exercises) from past nine years, solutions by more than 3 thousand students. This analysis is done by examining exercises suitability for the contest according to few criterions: difficulty, resolution, correlation coefficient with task. Also, exercises were divided into types (algorithmic exercise, half algorithmic exercise and heuristic exercise) and topic (textual exercise; equations, equation systems; inequalities,function features and graphs; elements of number theory; plane geometry; stereometry; arithmetical and algebraical transforms). Description statistic’s methods and cluster analysis was used to summarize material. The analysis exposes that exercises prepared for prof. Jonas Matulionis contest stand out with excellent or very good resolution, reasonable difficulty, relevant correlation coefficient with task. Cluster analysis results indicate that all types of exercises are alike with their correlation coefficient, similar with their difficulty. However, algebraically and arithmetical transform exercises types separates in their correlation coefficient with whole task, it is higher here.