T-test for two dependent samples - some practical notes

Abstract

This paper focuses on some important issues of the t-test for two dependent samples
(paired t-test). Emphasis is placed on the assumptions of the use of the test, on the
correct formulation of the tested hypotheses and on the selection of the correct method
in testing the averages of differences in pairwise observations. This is extended by
the calculation of the confidence interval for the mean difference and calculation of
the effect size by Cohend´s d. The paper also offers a demonstration of two different
solutions for the same example by means of correctly and incorrectly chosen
statistical test.
Key words: statistics, dependent samples, paired t-test

DOI: https://doi.org/10.24040/sjss.2021.7.1.48-55