Kruskal–Wallis two-way ANOVAs were used to test the null hypothesis (i.e., to verify
if each factor independently considered has significant influence on flexural and
compressive strength responses, to determine the main contributions of each factor
to global variance, and to identify any eventual interaction effect across them). A
data rank transformation was made considering the entire set of observations from
smallest to largest, and the usual parametric procedure was then applied to the
ranks of the data instead of to the data themselves.