A researcher's decision to use more than two levels of the independent variable in an experiment mainly depends on