It can make predictions about behaviour based on observations that can be shown to be more likely via statistics. When the predictions don't work, different models/theories are used.

The psychological equivalent of ID would be to assume that we know all about human behaviour and that gaps in our knowledge are 'proof' of 'intelligent possession'.

IOW, ID is not a scientific theory at all. It's like looking at a blank spot on a map, and rather than exploring it to find what's really there, marking it with 'here there be dragons' or 'Atlantis'.