NEW NONPARAMETRIC TESTS FOR LOCATION-SCALE TESTING IN A MIXED DESIGN

A mixed experimental design that combines Randomized Complete Block Design (RCBD) and Completely Randomized Design (CRD) can be structured to get advantages from both types of designs. In some experiments, researchers start with an RCBD plan but they might need to switch to a CRD due to circumstances beyond their control. Data from a mixed design can be analyzed using statistical methods that combine different statistics. In this paper, new tests are developed for testing the differences in mean and variances simultaneously for mixed designs of RCBD and CRD using nonparametric methods. These nonparametric tests are estimated and compared under a symmetric distribution via Monte Carlo simulation.