An inferential model-based method for testing homogeneity of several variances against tree-ordered alternatives

Comparing several normal variances with a control group is very important in many fields. In this paper, we propose a new inferential model-based test for the homogeneity of several variances against tree-ordered alternatives. Simulation studies demonstrate that our new solution provides a competitive alternative to existing methods. Furthermore, our IM test is the only one that works in scenarios where the sample sizes are unequal. A real numerical example is presented to illustrate the flexibility of the proposed method.