Statistical Power and Effect Size in Informative Retrieval Experiments

Authors

  • Michael J. Nelson University of Western Ontario

DOI:

https://doi.org/10.29173/cais437

Abstract

Statistical tests are used in information retrieval to test various hypotheses such as which indexing method is better or which retrieval system is better. Sometimes when using these statistical tests there is not enough evidence to reject the null hypothesis. Then either we have correctly discovered a true null hypothesis or made a type II error (probability denoted by b) and falsely accepted a . . .

Downloads

Published

2013-10-23

How to Cite

Nelson, M. J. (2013). Statistical Power and Effect Size in Informative Retrieval Experiments. Proceedings of the Annual Conference of CAIS / Actes Du congrès Annuel De l’ACSI. https://doi.org/10.29173/cais437

Issue

Section

Articles