Guest post by Joe Hoft
Last week we described the results of a recent analysis of a sample of polls promoted by the mainstream media where Hillary Clinton was reported ahead of Donald Trump in the US Presidential race. (See ACTUARIAL REVIEW: Analysis of Recent Polls Shows Trump Win and Possible Landslide) We provided the following analysis –
With all the liberal distortions and dishonesty we decided to have a small team of actuarial and statistics professionals take a look at a couple of the recent polls to get their take on the reliability of these polls. They selected the recent FOX poll from October 14 showing Hillary up by 7 and the WSJ/NBC poll from October 16 showing Hillary with an 11 point lead.
The first observation is that both polls are heavily skewed towards Democrats. At a high level, the FOX poll consists of 43 Dems to 36 Reps to 21 Other while the NBC poll shows 44 Dems to 37 Reps to 19 Other. By selecting more Dems the polls are designed to provide a Dem result.
Our experts next analyzed the data and calculated results using the same data from the two surveys on a split of 40 Dems, 40 Reps and 20 Other. The results show that using either sets of data Trump comes out ahead with a larger margin of victory using the FOX data.
New Information Supports Our Analysis –
A couple of questions as a result of this analysis were ‘Why are the main stream media polls so different than our estimates and other traditionally reliable polls?’ and ‘Why would the main stream media distort their polls so drastically?’
A piece of information that answers these questions came in three emails released by WikiLeaks from the Clinton campaign over the weekend –
The first email is from the Clinton campaign where they “recommend oversamples for our polling before we start in February. By market, regions, etc. I want to get this all compiled into one set of recommendations so we can manximize we get out of our media polling.”
— Philip Schuyler (@FiveRights) October 23, 2016
The second email to Clinton Staff shows “Proscribed disruption vectors” where the “Technique – Poll-flagging” has “Emotional Violence – Rage” with a “Message” of “Barrage with high-frequency recent polls. If countered with pro-Trump polls counter with friendly (538) averages. Avoid “Brexit” arguments. If target is hooked, move to swing-states (FL/PA) and declare election over. See Poll-Elasticity folder for further statistics.”
— Sheila Cooper (@S_Cooper0404) October 23, 2016
The third email shows the Clinton camp knowing that the Michigan polls were bogus in their race with Bernie Sanders:
These three emails support the calculations we presented last week that the polling data promoted by mass media is not accurate. Oversampling of Clinton supporters is used to obtain a Clinton desired result.
To be clear, the analysis presented last week of polling data supports a Trump win.
A second piece of information supporting the analysis reported last week that was shared this weekend is the polling completed by reliable sources which also report a Trump win. One pollster, Dr. Helmut Norpoth, shared his expectation of a Trump win. He has a remarkable record in predicting election outcomes. The key point that he shared from his model is that primaries are heavily weighted in his analysis.
Dr. Norpoth’s calculations are consistent with the analysis we presented last week –
This year Republicans crushed their previous record in the primaries for number of votes by 150%. Their old record was 20 million and this year 31 million voted in the primaries. The Democrats on the other hand had 7 million votes less than their record year in 2008 with 30 million this year compared to 37 million in 2008.
It was also noted last week that –
The primaries were heavily contested on both sides resulting in factions from each party vowing not to vote for the party candidate. The impact of these two groups is difficult to judge. The percentage of these voters that change parties is probably limited. If anything, the Sanders people will probably be more likely to vote for Trump since he is an outsider and many of them will never vote for Crooked Hillary.
Finally, it is difficult to determine what the independent voters will do but many independent voters partook in the primaries to vote for Trump. Therefore it is more likely that Independent voters vote for Trump as well.
If more Democratic voters vote for Trump than Republicans vote for Hillary and more Independents vote for Trump than Hillary, both scenarios which are highly likely, then the results of the general election will likely be a Trump landslide.
An additional piece of information that was not shared last week is that our small team now believes that the WikiLeaks emails are beginning to have a noticeable effect.
This cannot be understated. As WikiLeaks continues to share Clinton emails that show a corrupt candidate, Trump is vilified for his remarks regarding the election being rigged and Hillary looks more and more corrupt. This factor may very well be the key factor resulting in a Trump landslide although it is difficult to measure at this time.
Joe Hoft called the primaries 100% accurately https://t.co/7P57TxZ7vq
— Jack Posobiec (@JackPosobiec) October 24, 2016