Polls (Page 3)

ultimately I feel the polls are meaningless at this point either way because the debates and events are far more likely to determine the ultimate victor.  Trump could be sitting up 10 and I wouldn't feel any more comfortable. 

I think I mentioned earlier that I find it interesting how Hillary is up by 3%-5% in most polls, but she's spending a ton of money on ads to hold it.  Trump hasn't spent a penny and has been holding his own the past week or two and slightly gaining.
It will be interesting to see how things change when the debates kick in and Trump starts spending money to get his message out.

How much does, "believe me, it'll be great" cost?  LMAO

"believe me, it'll be great"

Pretty much what every politician says.  Chicken in every pot.......I will spend OPM on YOU!!

Zogby has it as a 2 point race

Clinton 38%
Trump 36%
Johnson 8%
Stein 5%
Not Sure 13%

The 13% not sure is interesting.  It shows a lot of votes still in play.

Originally posted by ejshowers

Originally posted by tuwood

Originally posted by ejshowers

Originally posted by ChineseDemocracy

Originally posted by ejshowers

Originally posted by tuwood

Zogby has it as a 2 point race

Clinton 38%
Trump 36%
Johnson 8%
Stein 5%
Not Sure 13%

The 13% not sure is interesting.  It shows a lot of votes still in play.

Why isn't that poll in the RCP list???

That's an easy answer...'cause it's not real clear.

ha ha ha :-) and not well regarded either.....

La Times +1 not well regarded either?

The LA Times poll is in there, but they only did Trump - Clinton poll, not a four way race. Not sure what you mean by +1? Gotta look harder Tony!

Clinton is up by only 1 point (+1) over Trump in a two way race by a reputable poll (LA Times) and Zogby (a middle of the road pollster).  Hillary typically does better in the two way races versus the 4 way races, so the +1 lead is likely lower if all four are considered.

Here's the trend on the LA Times poll over time.  You can pretend that Trump isn't in the process of erasing the post convention Clinton bump, but it doesn't change the reality of it.

Another interesting datapoint is the LA Times trend shows that both Trump and Hillary received similar convention bumps but are essentially in the same place they were prior to the conventions.

IMO polls are good for only one thing and that is showing which way the wind is blowing at any given time.

I saw a poll the other day that showed Clinton up by 6.1%.  Really?  You have a MOE of +/- 3% and you use 6.1%. 

Reminds me of the engineer who specified the length of a board to 3 decimal places....and the foreman who made the measurement with a tape measure....and worker who cut the board with an ax. 

The biggest question I have concerning polls this year is whether many of them aren't including Johnson and Stein, and how much inaccuracy that'll cause.  In most years, ignoring 3rd parties in polls is fine, but this year I question the effect of doing that, because of the larger percentage than usual who are already saying they're voting for them, the large percentage still undecided, and the effects of social media.  Anecdotally, I've been seeing posts on Johnson's facebook page by people who are Johnson supporters who were polled, but the caller insisted they could only select Trump or Clinton, and refused to accept the voter's answer that they're voting for Johnson.

It's hard to put a finger on the exact impact, but I do agree with you that a 4 way poll is a better representation.  The best we really have for comparison is the RCP averages because they show both:
http://www.realclearpolitics.com/epolls/2016/president/us/general_election_trump_vs_clinton-5491.html

Towards the top you can click on 3 way or 4 way to see the RCP average for each.  (note the skewing factor of the +13 monmouth poll from two weeks ago on the 4 way race)
Another dynamic that's pretty telling is the registered voters (RV) vs. likely voters (LV) which usually skews it another point or two.  So the ones that are typically the most accurate are the likely voter 4 way race polls.  The only ones the media likes to push with the big margins are the two way races with registered voters because they're typically the most favorable towards Hillary. 

If I were advising him I tell him to do exactly what Clinton does.  Speak only at rallies and read the speech her handlers wrote for her.  Don't do any press conference and don't take any questions.  Don't do any news shows unless you are being interviewed by a former employee of yours....like Stephenopolis and Clinton.

The RCP average is very laggy and typically lags the real numbers by a couple weeks.  I think we can all agree that the Clinton +13 poll that's still in the RCP average is nowhere near reality nationally today, but it's still in the average you're hanging your hat on.
Virtually all of the national polls have a tightening of the race approaching pre-convention levels and the most representative things are the overall trends and the latest numbers out.  The state polls are run far less often, and they typically lag the national numbers as well. 

I'm sorry if this reality doesn't fit the narrative you want, but it is what it is. 

538 rates LA Times as A-.  Monmouth has the race at +13 two weeks ago and LA Times has it at +1 today.  So what's that rating worth?

I mentioned a couple earlier, but I think it's pretty consistent across all the major polls (including RCP) that Trumps slide has arrested and he's started gaining.
I'll be curious to see what the next Monmouth poll shows because they're the outlier.

In general I try to look at trends within a single poll like the LA times trend as an example.  A single polling outfit uses the same methodology against similar demographics and captures trends better.  They may be off by 3-5%, but their trend should be on.  So if Monmouth was at +13 for Clinton and is now at +8 then it confirms the trend that LA times is showing with Clinton moving from +6% to +1%.

We can try to beat each other up on polls till the cows come home, but honestly it really doesn't matter that much right now.  If anything it just gives us something to argue about.   

Trump back in the lead in LA Times poll:
http://www.latimes.com/politics/

Trump: 44.2%
Clinton: 43.6%

That guys bias couldn't be seen from a mile away.

Anyways, every single poll has bias put in place because they have to modify the data based on their projected turnout models for the election.  For example, there was a poll in Ohio a couple days ago that had Trump +5 in the raw data of respondents, but they had more republican than democrats respond to the poll so they adjusted it down to their expected turnout on election day and it was Hillary +3 in the final results.  Did they put a bias in, or did they just adjust to what they think the turnout is going to be.

One big reason for the massive swings between polls is their turnout models they're using.  Many of the main stream polls are using Obama turnout numbers for Hillary, which would most definitely have an effect.  However, she's such a dud as a candidate with zero enthusiasm and Trump is a rock star (like Obama) on the Republican side it's really hard for me to believe the turnout would be 100% opposite the enthusiasm.

I am legitimately interested to see how the real numbers turn out on election day compare to the various polls.  One thing is for sure, with the wide spread in them "somebody" will get to say they were right.  hah