to vary only in the specification of the groups. This can provide more specific
and useful information to a practitioner about how future observations for the different groups may
vary, rather than just an assessment of whether there is a statistically significant difference between
the regression models