This is in contrast to the RCP average of Clinton +9.2, and the last 4 polls being
Clinton +11
Clinton +10
Clinton +10
Clinton +11
RCP and others are simply not reporting this outlier, which is making right-wing heads espolode with conspiracy theories about the big bad media.
The really odd part about the poll is who conducted it: CEPEX Center for Excellence in Project Execution. It’s principle was quoted as saying :“One of the reasons of Trump’s surge is attributable to increasing support from the women”
Never heard of the CEPEX Center for Excellence? That’s because it’s based in Calgary, Alberta, and was only formed 5 months ago as a numbered company. It consists of two chemical engineers with no previous experience in polling whatsoever.
Who paid for this poll? Has anything this ridiculous ever happened (in terms of fake polls) during previous elections? The whole thing is very strange in my opinion.
Fake polling is now a strategy. If all the polls say you’re down by 11, publish one that say’s you’re up by 15 and pray that RCP includes you in their “analysis” or averages.
I will question their ethics. A brand new tiny firm in Calgary suddenly decides to prove their capabilities in management consulting by… Polling in the state of Pennsylvania about the US Election?
Not only does this not pass the smell test, it stinks to high heaven.
Margin of error. Their margin of error is 7.1% within a 95% confidence level. It’s saying there’s a 95% chance that the actual results are somewhere between Trump 34.8%-50% to Hillary 29.4-43.6, basically.
Which is in fact roughly consistent with other polling in that its 95% confidence interval (inclusive of a result of Clinton +8.8) overlaps with MOE ranges around other polls - which aggregate around Clinton +9.2 in the RCP rolling average, Clinton +9 by Wang’s median approach, and Clinton +7.5 in 538’s reckoning under polls-only method.
Hey outlier results happen to fairly well established houses. Remember a month ago Quinnipiac was reporting a PA result of Trump +2 even as the other houses were all giving Clinton the lead, with several by 9 or 11.
Maybe they are honest brokers (okay likely not) but minimally they are illustrating the Dunning-Kruger effect.
I do wonder if there is a factual answer to a question in the op; “Has anything this ridiculous ever happened (in terms of fake polls) during previous elections?”
Not that I recall. Anyone else have a recollection?
-If they want to make their mark as a new polling company, against stiff competition, then a poll in a battleground state in an election as…notorious as this one makes a certain kind of sense.
-“Never attribute to malice that which is adequately explained by carelessness.” - or in this case, inexperience. They state upfront that the margin of error (;)) is quite large, which gives them a point or two in my estimation.
-I also see that it was a poll conducted through land-line robocalling. This implies to me a bias towards older voters, who tend to be more conservative. Younger folk are more apt to have cells and no land-line, after all.
-Had it been commissioned by the Trump campaign, I’d expect it to be more favourable to Trump - maybe even showing him winning. The RNC may be smarter, but what’s the point?
Anyway, I’m not saying this poll is correct - I’m just saying there’s no evidence for it being crooked.
One of the reasons why Silver and 538 got famous is that he not only looked for the best polls, but as statistics propeller head he also had a hand in cleaning his sources. His group was involved in identifying pollsters like Research2000, Strategic Vision and others that the evidence showed that they were either making fraudulent polls or that they were very bad pollsters (like Zogby)
It seems that now they are not so active in routing bad sources, unless they are nowadays using a Trump like argument of pointing at guys that they know in their gut that they are nuts but making it sound like if they are “defending” them.
Bottom line, just adjust for the pollster bias and relax, but I sure miss the old “crush the [del]bastards[/del] bad sources” 538 of the past.
The thing is - they are NOT a new polling company. They are a couple of chemical engineers who started a company supposedly to do project management with the oil and gas industry in Calgary (according to their web page)
Why in God’s name would an oil and gas managment company in Canada do robocalling in Pennsylvania? It beggars the imagination.
It WAS favourable to Trump. +5 for Trump, versus + 11, +10 for Clinton in other recent polls. And rather curiously, it was spewed all over the web by right-wing sources, soon after it was put on their web page. Wonder how they found out about it?
In the last election, you had at least some semblance of normalcy, with a mostly conventional two-party nomination process and two conventional nominees, running two mostly normal campaigns. Breitbart was still just a few months past the death of its libertarianish founder.
Trump’s campaign has moved the needle on what is now considered normal. We’ve gone back to the era of 19th Century yellow journalism and we now have a ‘yellow campaign’, if you will. Part of the strategy of the Trumpists, whether it’s coming from Trump himself or not, is to throw so much bullshit into the information sphere that people can no longer tell the difference between reality and fiction.
There have always been rumors in politics. In previous years, when the mainstream media were the gatekeepers of information, there was some management of misinformation. In the age of social media, blogging, tweeting, sharing, and so forth…not so much. People are more likely to get their information from communities of like-minded people or from people who at least share some of the same views.
That brings us to the fence-riders, the undecideds, whatever we want to label them. In the past, they would have been influenced by coming home, turning on their televisions or reading online articles published by Reuters or AP, and influenced by whatever content they’re consuming. Now in 2016? That has changed. They might be influenced by what friends post and share. They might also be influenced by whatever is in their feed, and who knows how algorithms feed news?
Polls are important because they get attention. They also project strength. They can convince voters that there’s a reason to participate in the process and actually take time to vote. Fake polls get published, tweeted, and shared now just like real ones.
“Why is a Management Consulting Firm specializing in Oil/Gas/Resources getting involved in Polling? Should it not be left to regular Polling companies?
Response: As long as a firm is knowledgeable in Statistics and methodology of conducting the poll, along with strong control on the randomness of the sample set, any one can conduct a poll. This is an endeavor to show that our consulting firm has the capability to conduct surveys while sticking to an unbiased approach. nothing is more important to this firm than honesty and ethics.”
To quote Robert Heinlein: “When a man speaks of honor make him pay cash.”
I wouldn’t be surprised if these guys just did the poll to “get their name out there” as they are a startup consulting firm. Kind of silly, since they have no reason to think they’d be any good at it. "All you need is to know statistics and proper methodology "? Yeah, no shit but do you?
A few weeks ago I remember someone posting about a poll showing Ryan was in trouble in his primary race. Checking now, I see it coincidently was our OP here. In the end Ryan actually won with 80% of the vote. I think that one was more obviously a purposeful lie. http://boards.straightdope.com/sdmb/showthread.php?p=19525169&postcount=57
