There've been a few posts like this. I think the possibility of a software bug should be taken seriously.

Basically, combined scores should be higher than the average (or further from 100) of the component scores, since it is harder/rarer to score highly on several tests than just one.

The psychologist should realize this, and should be investigating. They should not automatically accept a computer output that is so overtly non-sensical.