Pollsters got it wrong in last presidential election

Mar 2019
278
163
Portland, OR
There have been quite a few posts on this thread where I was told I didn’t understand what margin of error means and that the polls were right when they got it wrong but were within the margin of error. Now you are telling me margin of error doesn’t matter as long as the poll predicted the winner correctly.
You are mixing up the margin of error and margin of victory. The margin of error is the inaccuracy level of the poll, this very much does matter. The margin of victory is the expected amount the candidate will win and that is what I'm talking about, the margin of victory doesn't matter very much.

For example, lets say that a poll predicts candidate A will win with 51% of the vote, the margin of error is 5%, and he actually did win by 51%. Even though the poll's results perfectly matched the result of the election, because the margin of error was so wide, its prediction of who was going to win is mostly random and you can't give this poll too much credit.

Lets say poll B says the candidate was going to win with 55% of the vote, the margin of error is 2%, and the end result is that he wins with 51%. The poll was way off on the margin of error because of unexpected factors which unfortunately happens a lot today. But the poll was still essentially right about the end result which is all that really matters. That candidate will get elected equally with 51%, 55%, 60%, or 70% of the vote.

If poll C, says this candidate will lose with 49% of the vote, it has a margin of error of 3%, and its wrong and the candidate actually wins with 51% of the vote, you really can't blame the pollsters since they admitted that it wasn't an accurate predictor of the end result.

Also, the polls were not indicating a “complete toss up”. Clinton was expected to win because of the polls. The media reported the pollster results. Clinton was a huge favorite in Las Vegas also and that was because of the polls. The pollsters even fooled Clinton to believe she had a lock on states and as a result she skipped visiting those states.
She was leading heavily in the early polls, but toward the end of the election she was leading by 272 to 268 and even this was uncertain because 171 of these were tossup guesses. However the media was predicting Hillary winning with well over 300 electoral vote. The media just couldn't believe that Trump would win and focused on the earlier polls, while ignoring the recent ones coming out. The recent polls showed Hillary losing Ohio, Florida, and Iowa (still tossups) within the last few days of the election and the most recent polls in Pennsylvania and Michigan showed an uncertain Trump lead conflicting with earlier polling. Wisconsin however was a complete surprise. So they were only wrong about 1 state she had a lock on.
 
May 2007
5,283
2,749
your place
You are mixing up the margin of error and margin of victory...

No. I didn’t mix them up. If the margin of victory is way outside of the margin of error, it shows the pollsters very likely had some sort of errors in their methodology.


How Political Polling Works

“ The margin of error is one of the least understood aspects of political polling. The confusion begins with the name itself. The official name of the margin of error is the margin of sampling error (MOSE). The margin of sampling error is a statistically proven number based on the size of the sample group [source: American Association for Public Opinion Research]. It has nothing to do with the accuracy of the poll itself.
 
Mar 2019
278
163
Portland, OR
No. I didn’t mix them up. If the margin of victory is way outside of the margin of error, it shows the pollsters very likely had some sort of errors in their methodology.


How Political Polling Works

“ The margin of error is one of the least understood aspects of political polling. The confusion begins with the name itself. The official name of the margin of error is the margin of sampling error (MOSE). The margin of sampling error is a statistically proven number based on the size of the sample group [source: American Association for Public Opinion Research]. It has nothing to do with the accuracy of the poll itself.
The margin of victory is very often outside the margin of error for both right-wing and left-wing polls. This happens because fewer and fewer people are responding to pollsters and the assumptions behind margin of error calculations aren't always true. With a low response rate, the people who do respond are more likely to be politically engaged and engaged with the polling organization, which makes them different than the general population. Pollster will have to make some assumptions about the voting demographics which can be very unpredictable when people head to the polls.

Its gotten so bad that you can't completely trust individual polls anymore. More accurate sources are poll aggregators like Five Thirty Eight, RealClearPolitics, and PollingReport and looking at many polls from left, centrist, and right-wing sources.