How can you treat multiple observations that depict the same individual?

I usually identify only one observation too, but there is the risk that the other observation will be identified in another way and maybe both will get the Research Level.