Two-Sample Gradual Change Analysis

Abstract

We propose two-sample gradual change analysis motivated by gender differences observed in a real data set containing jumping speeds of 432 girls and 364 boys aged 6 to 19 years. Looking at this data set from the point of view of change-point analysis is more natural and it leads to more precise estimators than application of standard two-sample t-test in each age group. Apart of establishing the asymptotic distribution of the proposed two-sample change-point estimator, we also investigate its small sample properties in a simulation study.