What it says about Perl is simply that it was not designed to do well in such studies...
I disagree. If it says anything, it's that if you give bad code to people who don't know anything about programming, they'll have trouble writing good code.
If you give a noob the task of writing a program in a language the noob has never seen before it makes sense the noob would write code that looks like random gibberish to a Perl programmer. Therefore, it's not surprising the random gibberish code would would resemble "a language design by chance" more than it would look like Perl.
The point he was making wasn't about the output, but about the fact that the learning material given to the test subjects was the worst perl imaginable.
I disagree. If it says anything, it's that if you give bad code to people who don't know anything about programming, they'll have trouble writing good code.