To really the answer the question one would have to put in some qualitative question with the same sample group.
However, my reasoning is as follows: first you have to consider the type of people answering the questionaire and that question. It's not people who have these kind of languages on this question as their major "home" language, but have to go to it, for one or the other small issue.
Here the most determining factor on the first impression is how much it looks like your home language. However, I don't think this is the most important issue, the more important issue is, how much surprise-bites you get when things are suddendly subtly different than your home language.
If you are used to 0-based arrays in your language you use 90% of the time, 1-based arrays drive you crazy for the one job you have to (or better do out of other reasons) in the other language. Heck, many of the top "dreaded" languages in the questionaire are 1 based, or even worse, mixed, VB/A, Cobol, Matlab, Perl scoring worse than Lua.
Similar other things that are surprise bites.
PS: I would take the whole thing more seriously if PHP would lead by a fair margin...
