If you, like much of our fake national news media, follow the RealClearPolitics.com average of recent polls, you likely believe that Hillary Clinton has a modest, but consistent lead in that measure of somewhere around 4-5 percentage points over Donald Trump in a head-to-head race. When polls add in the Libertarian and Green party candidates, the lead shrinks by about a percentage point. That's been pretty consistent over the last few weeks, though it has narrowed somewhat from a month ago.
What you need to know is this: In early June - June 4 to be exact - RCP, for whatever reason, chose to start including the Reuters/Ipsos daily tracking poll results in its average. As we detailed here several weeks back, that poll consistently and obviously intentionally over-samples Democrat voters in its' surveys by a huge amount. Because of this, it consistently produces a skewed result that heavily favors the Democrat candidate. This is not a secret - it is something everyone has been aware of for a long, long time.
The folks at RCP appeared to realize this reality from the outset of that tracking poll a year ago, and for ten months studiously refused to include its results in its average. That changed, as mentioned above, six weeks ago, just as Trump had closed what had been an 8-10 point Clinton lead. I inquired via email with RCP as to the reasons behind their decision, but received no reply.
This morning, another tracking poll - the NBC/SurveyMonkey daily tracking poll - shows the race has narrowed to a Clinton lead of 47-44. This compares to the Reuters/Ipsos result of Clinton 44, Trump 33.
Guess what? RCP does not include the NBC/SurveyMonkey poll in its average.
So let's play a little game: If one factors out the Reuters/Ipsos clear outlier from the current RCP average of polls going all the way back to June 24, you come out with an average Clinton lead of 3.5%, right in line with the NBC/SurveyMonkey poll.
Further, if one includes only those polls released in July, and thus far more reflective of current events, the average result comes out to a Clinton lead of 2.3%, well within the margin of error of all the polls taken.
Again, I have no idea why, after having ignored the Reuters/Ipsos poll for ten solid months, RCP chose to begin including it in its average at a particularly fortuitous time for Ms. Clinton. I've asked, and RCP chooses not to reply.
But if you rely on the RCP average for your info on where the polls currently stand, just be aware that its results are significantly skewed by the inclusion of a poll that intentionally uses a widely skewed sample.
Paragraph 4, last sentence. Is there a typo? Did you intend 43 instead of 33?
ReplyDeleteNo, 33 is right. R/I shows an 11 point lead. Ridiculous.
ReplyDeleteWould you help with a perspective on where past presidential contests' polling numbers sat in July?
ReplyDeleteDo the numbers historically move significantly over the next several months?
Thanks for your insights as always!
RCP and Gallup both have a bunch of historical info on their sites. The polls always move up and down as Election Day approaches. We'll see them move in favor of Trump during and after the GOP convention, and then move back in favor of Hillary after the DNC convention. After that, it will just depend on the course of events and how each campaign is managed.
ReplyDeleteAt the end of the day it will all come down to how much the general population truly wants a real change from the last 8 and 16 years. If that desire truly is pervasive, then Trump will likely win the election, because he is the only change agent present. If it isn't as pervasive as some of us think, then Hillary will probably win.
My view is that Trump has one job: to convince the voting public he is not dangerous, which is how Hillary and the fake media have tried to paint him. If he can do that, he'll win going away.