Simon Rosenberg shared 5 current polls putting Biden pulling ahead. But polls are snapshots and most pollsters aren't all that good at their jobs. So take a breath but don't stop running.
Well, I think most pollsters are good at their jobs. I am biased, because my late mother was a pollster of sorts, mainly at the Census Bureau. Most are good, but the job is difficult, and getting more difficult with the changes in technology. Plus, you gotta read the small print. People say the polls were wrong in 2016 because Hillary lost. The polls were right to within their margin of error. They showed she might well lose in the Electoral College. She did win the popular vote.
In 2016, my daughter was working as a Democratic political operative with various senate campaigns and the Clinton campaign. She told me that weeks before the election she knew they would probably lose. It was no surprise, and it did not indicate the polls were wrong.
The worst polling in modern history was the Literary Digest poll predicting that FDR would lose the 1936 election. The reasons for that blunder were apparent to the experts. My mother said the pollsters learned their lessons from that. She said it is unlikely anyone will make such mistakes again.
Yes, they have many technical problems. Despite this, their predictions for the 2016 and 2020 presidential elections were close. They were within the margin of error, so they know what it is they don't know. People say "the polls said Hillary would win." No, the polls did not say that.
It would be a big mistake to ignore polls because it has become more difficult to take them, and because -- as you say -- they have to tweak them. Those people are good at tweaking. Modern computer technology lets them tweak in ways that my mother's generation could not have done.
Among themselves, professional pollsters say the tweaks no longer work very well. Most election polls report a 95% confidence level. Yet an analysis of 1,400 polls from 11 election cycles found that the outcome lands within the pollтАЩs result just 60% of the time. And thatтАЩs for polls just one week before an electionтАФaccuracy drops even more further out.
These changes in the polls post-conviction are being blown way out of proportion by both sides. They are miniscule shifts in both directions well within the margins of error.
Yep, because the possibility of conviction was already baked in. People who support Trump do so regardless of conviction, and people who oppose Trump opposed him (as do all patriots) also regardless of conviction.
The Biden campaign just told me:
The polls are shifting in our favor!
FIRST: Trump was convicted of 34 felonies.
THEN: New polling from Politico shows that Trump's convictions have turned independent voters against him!
NOW: President Biden has just taken the LEAD in two crucial battlegrounds: Michigan and Wisconsin!!
In the 1990's pollsters got a 36% response. Now it is 6%. I would not trust polls one way or another except maybe to reveal trends.
Simon Rosenberg shared 5 current polls putting Biden pulling ahead. But polls are snapshots and most pollsters aren't all that good at their jobs. So take a breath but don't stop running.
Well, I think most pollsters are good at their jobs. I am biased, because my late mother was a pollster of sorts, mainly at the Census Bureau. Most are good, but the job is difficult, and getting more difficult with the changes in technology. Plus, you gotta read the small print. People say the polls were wrong in 2016 because Hillary lost. The polls were right to within their margin of error. They showed she might well lose in the Electoral College. She did win the popular vote.
In 2016, my daughter was working as a Democratic political operative with various senate campaigns and the Clinton campaign. She told me that weeks before the election she knew they would probably lose. It was no surprise, and it did not indicate the polls were wrong.
The worst polling in modern history was the Literary Digest poll predicting that FDR would lose the 1936 election. The reasons for that blunder were apparent to the experts. My mother said the pollsters learned their lessons from that. She said it is unlikely anyone will make such mistakes again.
Because the the historically low response rate, the weighting tweaks are becoming less and less effective, making polling less and less reliable.
Yes, they have many technical problems. Despite this, their predictions for the 2016 and 2020 presidential elections were close. They were within the margin of error, so they know what it is they don't know. People say "the polls said Hillary would win." No, the polls did not say that.
It would be a big mistake to ignore polls because it has become more difficult to take them, and because -- as you say -- they have to tweak them. Those people are good at tweaking. Modern computer technology lets them tweak in ways that my mother's generation could not have done.
Among themselves, professional pollsters say the tweaks no longer work very well. Most election polls report a 95% confidence level. Yet an analysis of 1,400 polls from 11 election cycles found that the outcome lands within the pollтАЩs result just 60% of the time. And thatтАЩs for polls just one week before an electionтАФaccuracy drops even more further out.
The Biden campaign ads I'm receiving have a very different tone, saying he's behind in most swing states while asking for my support.
These changes in the polls post-conviction are being blown way out of proportion by both sides. They are miniscule shifts in both directions well within the margins of error.
Yep, because the possibility of conviction was already baked in. People who support Trump do so regardless of conviction, and people who oppose Trump opposed him (as do all patriots) also regardless of conviction.