I’m amazed that no one has mentioned the oldest known media bias: dog bites man is not a story. We’ve been talking about a blue wave for months, and it’s entirely predictable that the GOP will lose ground in the midterms in a complete absence of other information; that’s a century-old expectation.

So any story about GOP momentum is newsworthy, and it’s not in the least surprising if journalists and editors unconsciously bias in favor of printable news. (Same way that a scientist cheers for proof of hypothesis, not so they’re right, but so they can publish.)

The way this is solved in science is with meta-analyses, and arguably the equivalent here is aggregate sites like 538. I don’t know if I trust the exactness of their outcome (if we lose a 6 in 7 election, does that imply bad luck or nefarious doings?), but if the graph of the trend line ticks upwards or down, you can take that to the bank, and little else. Everything Shows a GOP Resurgence Except for the Evidence