I think the error would go both ways. Average users might fail where she succeeds, e.g. she recognized the Firefox icon immediately. They might succeed where

I think the error would go both ways. Average users might fail where she succeeds, e.g. she recognized the Firefox icon immediately. They might succeed where she fails, e.g. they won't search the web for .exe files. Just to be clear, I do think this experiment was worthwhile. It's important for Ubuntu to be usable by more savvy users. I just think a separate experiment would be needed to truly measure usability for computer-illiterate users.