G. Elliot Morris left the Economist and they immediately made a new polling average that is much worse. Last week Morris indicated that his average would be about Trump +2. Maybe that's 1.5 now with the slight improvements in the trackers, but no sophisticated polling average would have Biden up.
Every news outlet wants to have their own polling average but most of them are just doing RCP But Smoothing The Line. That is arguably a worse approach to the polls than the vibes, which might more accurately weight higher-quality pollsters.
Eh, would disagree with this. Trump's lead would fall below a point in the RCP average simply by just taking out Rasmussen and HarrisX (both dubious pollsters with awful marks and track records and ones that I assume don't pass Economist's smell test), so it's not that hard to get a Biden lead depending on what the pollster quality grading scale is and how recent the average is (if they're just going off of the last week, or the last month, etc.)
That's my point. A sophisticated polling average will look further back in March and anchor to the high-quality data. When the best pollster has Trump +5 two weeks ago and you also have Fox and WSJ with Trump +2 at the same time, those should be dictating most of the average. The only good poll with Biden leading is Emerson.
A simple average with arbitrary guidelines of when polls are included or dropped is what RCP has and that's why it's bad. When any other outlet does the same, but this time in a way that would produce a Democratic bias when there's not much data, it's also bad.
