The article mistakes more ideas for better ideas. The benefit of working in a good group is not that you get more ideas, but that you take the ideas that you get, prune the bad ones and combine the good ones to come up with the 'best' idea.

If 'mutual respect' or 'fear of looking stupid' gets in the way of suggesting something while working in a group (as it quite often does) then you need to work on overcoming that fear (individually) and improving your rapport (as a group)

Now if they could repeat the study after having predetermined some criteria to judge the 'best' solution to a problem, my hunch is that they would find that a group with a good rapport (not a bunch of strangers put together) would come up with the better solution (all other things - e.g. familiarity with the problem set, problem solving skills etc being equal between groups). But yeah, it's just a hunch.