Many experiments on a system or process have multiple responses of interest, each with potentially different data types and hence underlying distributions. However, traditionally, designing the ...