They did look at it critically. Research 2000 was fired by Daily Kos before anyone noted any impropriety in the figures, simply because the numbers weren't matching up with reality. Shortly after this happened, Grebner, Weissman, and Weissman approached Markos with evidence of deliberate impropriety.
Does Daily Kos have a responsibility to not promote questionable information as truth? Of course, and they've apologized for the situation. But keep in mind that this information is only coming to light because someone with sufficient statistical background took the time to pore over the data. That sort of expertise is hard to come by, which is the reason why smaller media/news outlets contract out to firms like Research 2000 in the first place!
It's only relatively recently that there's been much interest in the science of polling. Before the emergence of aggregation sites like FiveThirtyEight or Pollster.com, it was extremely rare that you'd ever see this kind of statistical analysis of polling data. The traditional method of testing a pollster's reliability was simple trial and error over a period of several elections. Really, that's *still* the primary method. If anything, Research 2000 only got scrutinized in this case because of the issues with their accuracy that led to them being dropped in the first place.
For me, it's not really a partisan issue, despite the highly politicized nature of Daily Kos. It has more to do with the size of the media outlet. I would expect a major news organization with dozens or hundreds of employees, like Fox News or MSNBC, to be able to detect problems like these very quickly. A relatively small blog with maybe a dozen part-time employees like Daily Kos, or Red State, or whatever, I'm more willing to give a pass. At least at first: I'd expect Markos to learn his lesson from this and be more proactive in ensuring that it doesn't happen again.