I know, many of us want to put the terrible 2016 presidential campaign behind us but there are some that still want to make issues with what happened. Contrary to the belief of some, from a purely statistical perspective, it doesn’t really appear that either candidate was really hurt or helped by the news in the last 2 months prior to the election.
The two big news events that many Democrats are hanging their hats on were the announcement by Comey that the FBI was going to reopen the investigation against Hillary Clinton, 2 or 3 weeks prior to the election and the constant flow of information by Wikileaks regarding the Democratic party dealings to make sure that Mrs. Clinton became the nominee and eventual president.
The measures that I’ll use to illustrate my point are:
- Google Trends information which illustrates Google search interest around a certain topic, specifically looking at the search term “wikileaks“
- Google Trends information which illustrates Google search interest around a certain topic, specifically looking at the search term “Federal Bureau of Investigation“
- Real Clear Politics poll averages of Favorability & Unfavorability ratings on both Hillary Clinton and Donald Trump
I’ve included links to all of the information that I used and charted in this overview.
One of the assumptions in most elections is that people vote for the people that they like, largely for the positions that they have on various issues. In many recent elections it could be argued that elections have come down to people voting for people that they hate less, this looks to be the case in 2016.
The chart on the left illustrates the Favorability ratings of both candidates in the weeks preceding the election. As you can see, Mr. Trump really never went above a 40% favorability rating and Mrs. Clinton never moved above 44% for any period of time. Right before the election, both candidates positive ratings dipped. This reflected the timing of the Comey/FBI announcement reopening the investigation on Mrs. Clinton. Interestingly BOTH candidates took a hit when this came out.
This chart illustrates that while positive views on both candidates increased in the 6 months leading up to the election, with Mr. Trump having more extended periods of lower ratings, the positive view of Mrs. Clinton held fairly steadily during the period with some moderate growth, mainly around the time of the Democratic and Republican conventions in the summer time.
Favorability Ratings for Trump & Clinton w/FBI Interest
The gray dotted line in the chart above illustrates the Google search interest around the term “Federal Bureau of Investigation”. There was and is always interest around the term, the interest spike significantly to 3 to 4 times the normal interest in the 2 weeks before the election. As you can see, the Favorable ratings for both candidates dipped when the news came out. The favorability of Mrs. Clinton did take a slightly bigger hit of 1.9% as compared to Mr. Trumps 1.1% — but they both had a negative impact from the news.
Now let’s take a look at the chart above. Interest around ‘Wikileaks’ had a longer impact on the peoples’ view on the candidates. The first news started to roll out in the middle of July, right before both parties had their conventions. There are two possible explanations for the fact that Mr. Trumps’ favorable rating went down and Mrs. Clintons went up. First, the Democrats could have very successfully distracted people away from the news or it’s possible that people just disregarded this information.
Wikileaks then had a quiet period but right as September was ending and for the 5 weeks leading up to the election, there was a tremendous amount of interest around this term (wikileaks) and what they were saying. Beginning in the first week of October the news and corresponding interest was not only steady but it was increasing up until the beginning of November. The impact on Mr. Trump was negative and there was virtually no impact on Mrs. Clinton other than a fairly steady increase. In the 4 full weeks of October her favorability ratings increased by an average of 0.25% per week.
During this same period Mr. Trump had an increase of 0.175% per week, but it was far different from Mrs. Clinton. The Clinton increase was steady, while the change in favorability for Trump was -3.2% in the first 2 weeks and +3.9% in the second 2 weeks, a very different picture. The chart below illustrates why this happened.
This chart illustrates the interest during 2014 on the search term “Billy Bush”. I’m sure that many remember the tape that was released by NBC on October 7th where Mr. Trump was recorded having a conversation with Billy Bush about women in general and about Nancy O’Dell in particular. In the 2 weeks that followed the release of this recording, Mr. Trumps favorability dropped that 3.2%. Interestingly in this same period Mrs. Clinton’s poll numbers didn’t really move at all with the first week having no change in favorability and the second week increasing 0.3%.
Quite honestly, the only thing that I think that you could conclude from all of this is that the polls didn’t really catch on what was going to happen. In the 8 weeks leading up to the election, Mrs. Clinton had an average favorability rating of 42.5875% and Mr. Trump had a rating of 37.0625%. Mrs. Clinton ended up with 48.2% of the popular vote (a difference of 5.6%) and Mr. Trump ended up with 46.1% (a difference of 9.0%). While the margin of “victory” was 2.1%, the favorability ratings were way off this time.
Out of the 3 big news stories that could have affected the election, it does appear that the FBI/Comey story could have had an impact, but it hurt both candidates. The drip-drip-drip of the wikileaks data, probably had little to no impact on the election results as both campaigns were scrambling to create offsetting or distracting noise. The only thing that I’m certain of is that the controversy around this election is not likely going to go away any time soon.