I think the graphs are significantly more telling for a show with high turnover, like SNL. There's an obvious shift in quality during certain periods; not only does the mean rating trend downward, but the individual episodes group together during certain periods indicating consistent quality problems. "Hating the lineup" probably sums it up perfectly.
I note that certain data points in some of the graphs lie exactly along the line representing the mean rating. Can we infer any sort of statistical value from that fact? Do those individual episodes truly represent the "average quality" of the show within the context of when the episode aired?