Power analysis for testing two independent groups of likert-type data

DOI

10.2991/iccsae-15.2016.8How to use a DOI?

Keywords

two-sample problem, nonparametric test, power of test

Abstract

In the one-sample location problem, it is tested whether the center of the whole is equal to a known value, otherwise whether there are significant differences between two samples is in consideration on pratical situation. In practical problems, there are many simulations where two general parameters are compared, instead it is tested whether the center of the whole is equal to a known value in the one-sample location problem. The aim of this article is to determine the goodness-of-fit of three different nonparametric tests, which being two sample rank test, Smirnov test ( two-sample Kolmogorov-Smirnov test) and two-sample Cramér-von Mises test. In the meantime the efficacies of their respective comparative analyses are also tested to choose their own two-sample test methods. Simulation results indicate that neither of the tests is the best for each sample distribution, but in most instances, the Cramer-von Mises test performs best. Moreover the Kolmogorov-Smirnov test is better than the Mann-Whitney test in term of distribution of samples, sample size and effect size .

Copyright

© 2016, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Zun-xiong Liu
AU  - Hao Chen
PY  - 2016/02
DA  - 2016/02
TI  - Power analysis for testing two independent groups of likert-type data
BT  - Proceedings of the 2015 5th International Conference on Computer Sciences and Automation Engineering
PB  - Atlantis Press
SP  - 34
EP  - 39
SN  - 2352-538X
UR  - https://doi.org/10.2991/iccsae-15.2016.8
DO  - 10.2991/iccsae-15.2016.8
ID  - Liu2016/02
ER  -