After revelations broke that Blumenthal had, that one time, accidentally exaggerated his military service, Rasmussen rushed out a poll showing a Democratic cataclysm in Connecticut:
Dick Blumenthal (D) 48
Linda McMahon (R) 45
Three polls followed, including ours, showing that Blumenthal had, actually, lost no support, and still had a massive lead.
His narrative-setting efforts busted and discredited, Rasmussen soon rushed out another poll showing that, oh well, maybe Bluementhal still leads big after all:
Dick Blumenthal (D) 56
Linda McMahon (R) 33
After the Kentucky primaries, with GOP nominee Rand Paul under fire for opposing the Civil Rights Act (among other things), Rasmussen rushed out a poll that showed Paul winning by a gazillion:
Rand Paul (R) 59
Jack Conway (D) 34
Wow! 25 points? That was odd. Our own pre- and post-primary polling showed Paul holding stead. Other polling soon followed showing that yes, Paul had a lead, but it was close:
His narrative-setting efforts busted and discredited, Rasmussen soon rushed out another poll showing that, oh well, maybe this is a tight race after all:
Rand Paul (R) 49
Jack Conway (D) 41
Remember, Rasmussen has two modes -- the narrative setting mode, which he's used to great effect this year to fuel the "Democrats are doomed" narrative, and the "get it right" mode he uses the closer we get to the actual election. Well, we're not that close to November yet, but apparently he felt threatened by having such massive outliers vis a vis other pollsters. So he rushed new polling more in line with the composite polling trendlines.
It's kind of hard to set narratives when you become a laughing stock in the political world.