This paper proposes a general approach for handling multiple contrast tests for normally distributed data in the presence of partial heteroskedasticity. In contrast to the usual case of complete heteroskedasticity, the treatments belong to subgroups according ... ...

请注册登录后继续浏览