The implementation in R software performs a robust hypothesis test comparing two samples, which takes into account measurement errors. This is a very unique approach to testing under the presence of measurement errors. The presented code performs a complete analysis, including the computation of the test statistic for a given dataset, the critical value, and evaluates the power of the test. Computing the critical value as well as the power is very tedious but feasible with this software for moderate sample sizes.

Feel free to use or modify the code.

No additional packages for R software are needed.

• The analysis must be performed in this order, starting with computing the test statistic, then the critical value, and finally power of the test: TestStatistic.R, CriticalValue.R, Power.R.

• Jan Kalina, The Czech Academy of Sciences, Institute of Computer Science
• Michel Broniatowski, Université Pierre et Marie Curie (Sorbonne Université)
• Jana Jurečková, Faculty of Mathematics and Physics, Charles University & The Czech Academy of Sciences, Institute of Information Theory and Automation

Do not hesitate to contact us (kalina@cs.cas.cz) or write an Issue.

Please consider citing the following:

Broniatowski M, Jurečková J, Kalina J (2018): Likelihood ratio testing under measurement errors. Entropy 2018, 20, Article 966.

This work was supported by the Czech Science Foundation grants GA19-05704S and GA18-01137S.