Michigan
Related: About this forumTrump and Harris is virtual tie in Michigan (Poll)
https://www.woodtv.com/news/elections/just-1-point-separates-trump-and-harris-in-michigan-new-poll-shows/Trump 46 Harris 45. Last week we saw Trump up 7 points on Biden in Michigan and I think that poll was a big reason Biden dropped out. We're back to square at the start of the race and I think Harris has a lot more potential voters she can win.
riversedge
(73,132 posts)Johnny2X2X
(21,758 posts)This is a poll that for some reason has been an outlier in favor of Trump all cycle so far. So this is a good result.
gab13by13
(25,257 posts)What does TSF have left to run on? Polls are one thing. Did I mention he got caught paying for a favorable poll. Do you think he has stopped doing that since he got caught?
I day after Kamala replaced Joe there were polls out, how speedy those people are, and how reliable those polls must have been.
Johnny2X2X
(21,758 posts)They are pushing polls out hastily. And one of the polls everyone was talking about was 3 days of pre Biden dropping out and 1 day of post. So the mathcups with Harris were hypothetical until just now.
And let's not forget, Trump is still coming off the convention bump and sympathy from having his ear nicked. Plus he picked a VP in the middle of that.
Harris will pick a VP soon and then the convention is in late August. She will get bumps from those and then things will settle down until the debate (s). So I think we're still 6 weeks out from knowing the shape of the race.
Fiendish Thingy
(18,517 posts)That would be a huge scandal, and would put Emerson out of business.
gab13by13
(25,257 posts)The Washington Examiner has a front page article out about how TSF still leads in the battle ground states. It listed 4 states.
All TSF has right now is polls.
As far as paying for polling, that's just one option. I still remember back when Gallup was the king of polls, the top dog, and it predicted Romney over Obama. Now people always claim, but the margin of error. In Gallup's case, it actually put bad data in its formula which skewed the final result. If the king of polls can fuck up, so can the rest of them.
To me the only reliable polls are exit polling, they validate the results.
Fiendish Thingy
(18,517 posts)Never get too excited or panicked over a single poll, watch for polling trends (not averages); the trends from the few recent polls show little to no convention bump for Trump, and a narrowing of margins for Dems, definitely nationally, and less definitely in swing states.
Of course, its only been a few days.
waterwatcher123
(246 posts)"If political pollsters were physicians, large numbers of them would be getting hauled before medical disciplinary boards for misdiagnosing their patients."
https://www.bostonglobe.com/2024/04/03/opinion/why-are-presidential-polls-wrong-biden-trump/
"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 polls result just 60% of the time. And thats for polls just one week before an electionaccuracy drops even more further out."
https://newsroom.haas.berkeley.edu/research/election-polls-are-95-confident-but-only-60-accurate-berkeley-haas-study-finds/#
gab13by13
(25,257 posts)If they were wrong approximately 50% overestimating the Democratic candidate then I might have more faith in them. It seems odd to me that the polls mainly overestimate in favor of the Republican. I don't have any data to back that up, just what it seems like to me.
waterwatcher123
(246 posts)Polling is fundamentally broken in this country. The news media needs to treat these polls as simply a tool, albeit one that is wrong more often than right (not enough skepticism about the results).
A poll where people opt in is not a random sample. Neither are polls where they keep calling people until someone answers (who are these people who answer - skew older and more likely to trust unknown callers). I do not see any of these pollsters really accounting for geography either (other than by state). If the country were just one big homogeneous population, their current methods might work. However, these samples are too small to effectively account for these differences (cluster sampling of large samples would help).
getagrip_already
(17,436 posts)First, even the rrputable pollsters have no idea who will show up to vote, or how many of each demographic will be in the mix, and they have no way to determine it.
Those assumptions are key to designing an accurate poll.
Will it be a blue wave? A red wave? A roe wave? A youth wave? A black woman wave? An angry white wave? Some mix of all of those? Will it look like 2016 or 2020 or 2022 or 2023?
And an unscupulous pollster will bend those assumptions to their goals.
But dont look for big swings.
Votes have been locked in for months. There really arent many undecideds; just people unwilling to admit.
So there wont be huge swings in the polls unless pollsters play with the makeup of who they expect to turn out. Which they may do to spice things up a bit.