The Relationship Between Clustering Variables and Meaningful Outcomes

Some variables you involve in cluster analysis, just really in the end don't separate the groups out. So I want to talk about the variables that are used for clustering as having some conceptual reason for being used. If we had groupings of people with respect to their alcohol consumption behavior, it would make sense that some of those groupings are related to particular outcomes more than others. It might also be the case that we wish to think about there as being variables that are precursors to this particular grouping. What are some things that might have led will say children to be in these particular behavior patterns or adults to be inthese particular alcohol consumption patterns?