The polls were not wrong.

[°Leprechaun]

In three key states (WI,MI,PA) the polls were not actually wrong. Clinton got about what was expected or better. The problem may have been margin of error and undecideds. It is difficult to say. Although Clinton got what was expected, she was leading Trump in all states, but that doesn't mean that the polls are wrong. To calculate the true margin of error you need to add undecideds to the margin of error.

[PresidentTRUMP]

LOL even after 306 electoral votes for Trump you stay strong with support of the polls. Americans weren't fooled by the ABC poll showing Clinton up +15 and they won't be fooled in 2020 either.

[°Leprechaun]

Facts are facts and you're not entitled to your own facts. The polls indicated a close election and that's what we got.

[TDwest6]

The polls and the media said Clinton would win.

They left out voters like myself - small town, Midwestern, moderates. I may be my no means a huge Trump supporter, but there's a reason the Democrats failed and the polls were wrong. My state nearly voted Trump, even though everyone assumed Clinton would win it.

The polls were definitely wrong, especially in this region. It's a fact that Democrat need to accept and prepare for. Without voters like me, the Democratic party's future is bleak.

[°Leprechaun]

I think Trump's polls showed that he would win. Why else would he put so much effort into PA and MI?

[ApatheticAustrian]

I think Trump's polls showed that he would win. Why else would he put so much effort into PA and MI?

lack of alternatives?

there wasn't any other way.

[°Leprechaun]

I think Trump's polls showed that he would win. Why else would he put so much effort into PA and MI?

lack of alternatives?

there wasn't any other way.

Actually, there were other paths without any of the three states, for example, firewall -NH and -ME2, but they were small states and it made sense to go after bigger ones. However, why go after states unless there is a chance to win them? I think his polls showed that he could win the states which he won and as I said in the OP, the poll averages showed that he could win them.

[ApatheticAustrian]

, for example, firewall -NH and -ME2, but they were small states and it made sense to go after bigger ones. However,

i don't want to go too theoretical but most of the other ways i have seen ran around NM/CO/NV and i think trump's own polls knew that they were blocked there anyway.

so they went big and won big....best kind of player.

[Beet]

The polls were definitely wrong. You can't argue margin of error when there are half a dozen polls that all show the wrong winner. The margin of error is supposed to be random.

[SirMuxALot]

And can we finally dispense with this idiotic idea that Trump had no internal polling (or dumped it in the last two weeks).

That would be despite her campaign chief being Kellyanne Conway, who RUNS HER OWN FREAKING POLLING COMPANY.

The Trump campaign visit to MN late (which also hit the SW Wisconsin media market) clearly tells us they had some internal polling that was closer to reality than the public MN/WI polls.

[Terry the Fat Shark]

Marist and RASMUSSEN look to be most accurate as far as popular vote is concerned, Trafalgar ( lol), Selzer, and I guess Gravis? look to be the most accurate state polls...

[SirMuxALot]

You can't argue margin of error when there are half a dozen polls that all show the wrong winner. The margin of error is supposed to be random.

Strongly agree.  Margin of error should follow an expected distribution of probability centered around the result, falling roughly equally to either side.

That almost all of the error fell on one side tells us this was not simple margin of error.  This was a systematic error.

Same was true in 2012.  Systematic error that slanted to Romney by 2-3 points.  The polls were roughly just as wrong in 2012 as they were now.  But we don't think about it that way since the outcome wasn't a surprise.

[Meclazine for Israel]

I say the polls were wrong.

At no stage did the polls predict a Trump win.

Every time Donald neared 50% chance his numbers dived.

The numbers were either fabricated with bias or the polling methods askew.

[TDwest6]

Marist and RASMUSSEN look to be most accurate as far as popular vote is concerned, Trafalgar ( lol), Selzer, and I guess Gravis? look to be the most accurate state polls...

Everyone missed out on the midwestern voters - stuff like this was bound to happen. Not like they're better - more skewed, really

[ApatheticAustrian]

we know that the campaign stopped classical polling 2 weeks ago, they have had some kind of modelling.

but i don't think that last-minute polling is able to change anything anyway.

[Storebought]

The collapse of Clinton in the marginal GOP states (GA, MO, IN) starting in October was a clear sign that something bad was happening, yet the AF Democrats chose to ignore it.

[SirMuxALot]

we know that the campaign stopped classical polling 2 weeks ago, they have had some kind of modelling.

How do we know this?  Give me a link.  And don't give me a link to a John Harwood tweet or a Matea Gold article.

[Likely Voter]

The national polls were off by 3 points, which isn't totally outside the MoE and all that different than previous cycles. And the polls in half of the swing states were off by less than 3 (FL, NM, VA, NV, VA, CO, GA) so those were also within the range, although again were mostly too D.

However the polls in NC plus the polls in the Upper Midwest/Rust Belt/New England (MN, WI, MI, OH, PH, NH, ME) were all off outside the MoE and all off in the same direction (too D). 

So the polls weren't all wrong but the polls in the NE were definitely wrong.

[Beefalow and the Consumer]

The national polls were off by 3 points, which isn't totally outside the MoE and all that different than previous cycles. And the polls in half of the swing states were off by less than 3 (FL, NM, VA, NV, VA, CO, GA) so those were also within the range, although again were mostly too D.

However the polls in NC plus the polls in the Upper Midwest/Rust Belt/New England (MN, WI, MI, OH, PH, NH, ME) were all off outside the MoE and all off in the same direction (too D). 

So the polls weren't all wrong but the polls in the NE were definitely wrong.

I came here to make exactly this point.  The PV result is not surprising at all given polling up to election day.  The state polls in the Great Lakes region OTOH were a massive clusterf**k.  Trump won states for which there was zero objective data pointing to his victory.  WI being the most obvious one.

I don't understand the Nate Silver bashing.  He was one of the few people who was relatively bullish on Trump in the final weeks, and said there's basically a 1 in 3 chance that polling error and a viable electoral path could make him President.  Everyone else wrote off his chances, or made it much more remote.  Nate, comparatively, nailed this.

[Erich Maria Remarque]

The national polls were off by 3 points, which isn't totally outside the MoE and all that different than previous cycles. And the polls in half of the swing states were off by less than 3 (FL, NM, VA, NV, VA, CO, GA) so those were also within the range, although again were mostly too D.

However the polls in NC plus the polls in the Upper Midwest/Rust Belt/New England (MN, WI, MI, OH, PH, NH, ME) were all off outside the MoE and all off in the same direction (too D). 

So the polls weren't all wrong but the polls in the NE were definitely wrong.

I came here to make exactly this point.  The PV result is not surprising at all given polling up to election day.  The state polls in the Great Lakes region OTOH were a massive clusterf**k.  Trump won states for which there was zero objective data pointing to his victory.  WI being the most obvious one.

I don't understand the Nate Silver bashing.  He was one of the few people who was relatively bullish on Trump in the final weeks, and said there's basically a 1 in 3 chance that polling error and a viable electoral path could make him President.  Everyone else wrote off his chances, or made it much more remote.  Nate, comparatively, nailed this.

His analysis were literally perfect.

- Hillary is favorite according to polls
- Trump has EC advantage
- Polling errors are correlated
- There is a lot of undecided/third party

[Erich Maria Remarque]

The national polls were off by 3 points, which isn't totally outside the MoE and all that different than previous cycles.

Single polls weren't outside the MOE, combination of them indeed was.

[Dumbo]

2012 RCP average Obama + 0,7 - he won 3,9, difference = 3,2
2016 RCP average Trump - 3,2 - he won - 0,2, difference = 3,0