Objective scoring for computing competition tasks

From IOI Wiki
Jump to: navigation, search
Pub conference.png
  • Conference paper: Objective scoring for computing competition tasks
  • Author(s): Graeme Kemkes, Troy Vasiga, Gordon Cormack
  • Book: Informatics education – the bridge between using and understanding computers (Roland T. Mittermeir, ed.), Springer, Berlin, 2006, pp. 230-241
  • Series: Lecture Notes in Computer Science, vol. 4226
  • DOI: 10.1007/11915355_22
  • Conference: 2nd international conference on Informatics in Secondary Schools: Evolution and Perspectives (ISSEP)

Abstract: Computing competitions like the International Olympiad in Informatics (IOI) typically pose several problems that contestants are required to solve by writing a program. The program is tested automatically on several sets of input data to determine whether or not it computes the correct answer within specified time and memory limits. We consider the controversy of whether and how to award partial credit for programs that fail some of the tests. Using item response theory, we analyze the degree to which the scores from these automatic tests, separately and in various combinations, truly reflect the contestants’ achievement.

Download: http://dx.doi.org/10.1007/11915355_22 (requires subscription)