When there are ______ in an experiment or quasi-experiment it is possible to partition variability into components associated with the two main effects and the interaction, then test these for statistical significance through omnibus F-ratios.
two dependent variables
two independent variables
two scaled contrasts
two independent samples
