I think their observation was that there was more stereotyping in a disordered environment. That this was due to a need for internal order is their interpretation of the data.

I'll print the paper out for you. :)