You're right, statistical significance is a valid concern but error bars is not the way to go. Check the paper for a rigorous analysis. We are not drawing our conclusions from a single plot.
I didn't think that you were (and having glanced at the paper now, there seems to be a lot more meat to it), but I think others will see only the plot. In my opinion, such a plot is misleading given the large variance I would expect in the rating.
