The last 5 Iowa polls within the last half of October, post debate, without Biden.
Clinton (+7, +11, +3, +38, +41)
lol, I guess Web and Chafee's voters relocation really help out Clinton.
But boy, is this #s so varied.
Selzer has an A+ rating from 538, and is the +7 poll, which has 2 polls 4 points on either side of it. Hillary is up somewhere around 7 points. Lets not bother with these joke polls.
Are 538 ratings based on primaries or generals, or both? Is it just the last cycle, or a rolling average? Are ratings set in stone, or do they ever change? Is it possible that final results could be affected by random error despite sound methodology?
Or is just easier to say "538 rating lollllz" than to do actual research, acquire in-depth knowledge of industry practices, and judge methodologies accordingly?
Look, Selzer is obviously the gold standard of Iowa. They know the Iowa caucus. Some of these bozo pollsters probably think they're polling some sort of non-caucus primary.
Gallup was once the gold standard nationally. Zogby was once an excellent pollster. Scott Rasmussen was once highly respected in the industry. Who cares? It's no guarantee of future performance. Selzer bombed the last Iowa caucus they polled, despite good results in 2008. They could be close this year, or not so close. But win or lose, it will have little to nothing to do with what they've done previously.
Too often (and I see this everywhere, not just here) people look at polls as "This
seems right to me, so it's a good poll." Or "This pollster has done well before, so this poll must be accurate." Or "This is a partisan pollster, so none of their results can be trusted." This is the dilettante's substitution for critical analysis: Hyper-focus on individual polls and houses, and judge them on sh**t that doesn't matter.
