Posts Tagged ‘Brexit’

Putin’s Peaks

June 25, 2018

The title of this post is identical to the title of an article by Dmitry Kobak, Sergey Shpilkin, and Maxim S. Pshenichnikov in the June 2018 issue of “Significance.” “Significance” is a joint publication of the Royal Statistical Society and the American Statistical Association. The subtitle of the article is “Russian election data revisited.”

The article states that the Kremlin wants a golden 70-70 win, meaning a win of 70% of the vote with a turnout of 70% to give it a clear mandate and provide it with a riposte to Western leaders who criticize Russia as an autocracy. What was actually achieved was a seemingly respectable 67.5%, with Putin securing 76.7% of the vote. But there have been criticisms of the election process, and doubts have been cast over the validity of the outcome. For instance, Golos, an election monitoring organization, has documented incidents of ballot stuffing at various polling stations, and multiple other violations both before and during the election ( At least since the mid-2000s Presidential and parliamentary elections in Russia have been accused of being fraudulent. From the Russian perspective, the two most important numbers that describe an election outcome are turnout percentage and leader’s result percentage. Although these percentages are not reported in the data sets from individual polling stations, they can be calculated from the information provided officially.

The authors (and others including HM) have argued that due to human attraction to round numbers, large-scale attempts to manipulate reported turnout or leader’s results would likely show up as frequent whole (integer) percentages in the election data. A previous “Significance” article gave the hypothetical example of a polling station with 1577 registered voters. Here election officials decide to forge the results and report a turnout of 85%. 85% was chosen as it is a round number which is more appealing than say 83.27%. To achieve a falsified turnout of 85%, this polling station needs to report 1755 x 0.85 = 1492 ballots cast. Other polling stations making similar attempts at fraud may also choose 85% as their target value, so that when we look at the turnout percentages for all polling stations, we see a noticeable split in the number of stations with turnout of 85%. In a previous article these integer peaks were found in elections from 2004 to 2012.

Since then two new elections were held in Russia, the 2016 parliamentary and the 2018 presidential elections. As with previous elections, sharp periodic peaks are clearly visible at integer values (91%, 92%, and 93%) and at round integer values (80%. 85%, and 90%) rather than fractional values (such as 91.3%).

The authors did Monte Carlo simulations of election results using the binomial distribution of ballots at every polling station. It strongly confirmed the hypothesis that results were being rounded to the benefit of the government. The authors note that integer peaks in the election data do not originate uniformly across all parts of Russia; they are mostly localized in the same administrative regions, providing additional evidence supporting that these are not natural phenomena Specific peaks can sometimes be traced to a particular city, or even an electoral constituency within a city, where turnout and/or leader’s results are nearly identical at a large number of polling stations. The most prominent example from the last two elections was the city of Saratov in 2016. Its plotting stations are the sole contributor to the sharp turnout peak at 64.3% and the leader’s result peak at 62.2%. These peaks are not integer and so are not counted towards the anomalies. Curiously, their product—showing the fraction of leader’s votes with respect to the total number of registered voters is a perfectly round 40%: 0.643 x 0.622 = 0.400.

One could regard these discrepancies, assuming that they do accurately reflect the underlying true vote as relatively innocuous. But the suspicion is that the results are significantly modified to get close to the Golden 70-70.

In the future it will be interesting to see if this integer bias persists in future voting summaries. It is disappointing to see this “rookie” flaw in a country noted for phony elections.

Russia’s newly developed strength is in influencing elections via technology. It has been discussed in previous healthy memory blog posts how Russian developed this new type of warfare. It began in homeland Russia. It was developed further in Russian speaking countries and in the Ukraine. And it has now been exported to Europe, where is it credited by some for the Brexit result, and to the United States were it is credited for Trump’s victory at least by some (Former DNI Director Clapper and HM at least).

Moreover, Russia is perfecting this new form of warfare and is promising its continuance. There is much talk of the upcoming midterm elections in the United States, yet nary a word about Russian interference. Trump is not taking any actions to safeguard these elections, which is perfectly understandable as Russian interference benefits the invertebrates supporting him. Even if the Russians are not entirely successful in benefitting Trump, just a small amount of interference could call into question the validity of the elections.

© Douglas Griffith and, 2018. Unauthorized use and/or duplication of this material without express and written permission from this blog’s author and/or owner is strictly prohibited. Excerpts and links may be used, provided that full and clear credit is given to Douglas Griffith and with appropriate and specific direction to the original content.


Daniel Kahneman and the Stupidity Pandemic

December 26, 2016

In case you haven’t noticed there is a stupidity pandemic.  It’s a pandemic because it rages throughout the world.  Perhaps the most conspicuous example are the religious wars.  These wars are several centuries out of date.  Terrorism is a religious war being largely fought within the Islamic faith with some terrorists groups targeting the west.  Then there is Brexit, a phenomenon that was not predicted by professional politicians.  In general there is lack of faith in what is called the “establishment” and in bodies of knowledge such as science.

In the United States there is the phenomenon of Donald Trump.  When Trump began his campaign it was regarded as a joke and was quite funny.  It is still a joke, but one that is no longer funny.  If every vote had counted in the United States, the Trump problem would not exist.  But an archaic and stupid institution called the electoral college elected Trump, therefore nullifying the will of the majority of US citizens.

So what has Nobel Lauerate Daniel Kahneman have to do with this?  His two process theory of human cognition provides a means of understanding this pandemic.  System 1 refers to our normal mode of cognition.  It is very fast and allows for fluent conversations and skilled performance.  It is the default mode of cognition.  System 2 is called reasoning and corresponds to what we colloquially call thinking.  System 2 requires attention and mental effort.  One of the jobs of System 2 is to monitor System 1 for errors.  However, this requires mental effort and thinking.

Experiments have been run where statements are presented to the research participant.  The brain is monitored.  When a statement conflicts with a participant’s individual beliefs, a signature is reported from the brain.  The question is whether this statement will be ignored, or whether the participant engages in deeper thought to reconsider this statement.  There is a cognitive cost here and the simplest reaction is to ignore the statement and regard it as a mistaken belief.

Trump’s  victory was a victory for System 1 processing.  System 1 appeals to fears, emotions, bigotry, and so forth.  Trump is a genius at connecting with and exploiting the System 1 processes of people.  Trump himself rarely uses System 2 processing.  He does not read books, does not think he needs to attend briefings because he knows everything already.  His gut, his System 1 processing, tells him what is true.  However, Trump does not care what is true.  It is whatever he believes at the moment, and this does change from moment to moment.  This is one of the reasons he is such an effective liar.  He does not care what is true.  It is whatever is expedient for the moment.  When confronted with his lies, he denies the truth.  His promise to make America great again was predicated on the lie that the United States is not regarded throughout the world as a great country.  Enemies dislike the politics of Americans, but nevertheless respect its greatness.

Totalitarian countries have exploited the big lie, and so does Trump.  See the healthy memory blog “Sick Memory.”  Lying has become a profitable industry.  Dana Milbank had an interesting column in the 21 December 2016 Washington Post title “Hoping that he didn’t really mean it.”  Milbank pointed out that many areas of the country that went for Trump will suffer deeply from cuts in government spending that will occur if Trump acts on his promises.  The title of Milbank’s article provides the explanation of how these voters reconcile their vote with the adverse effects that will affect them personally.
It is clear that these people did not employ System 2 processing when they voted.  There is justification for believing that these people rarely engaging in System 2 processing.  Like Trump, they go with their gut feelings.  Unfortunately, there is some question if such people will ever realize that they have screwed themselves.  Trump can continue to exploit their fears and bigotry to keep them in line.

© Douglas Griffith and, 2016. Unauthorized use and/or duplication of this material without express and written permission from this blog’s author and/or owner is strictly prohibited. Excerpts and links may be used, provided that full and clear credit is given to Douglas Griffith and with appropriate and specific direction to the original content.

The Dunning-Kruger Effect

July 22, 2016

HM had an embarrassing experience when his friend, a physicist, asked him about the Dunning-Kruger effect and he had to express ignorance.  HM was embarrassed because this effect is in the same field in which HM’s interests lay.  After learning about the effect, the relevance of the effect to the current phenomena known as Trump became evident.

There are two parts to the Dunning-Kruger effect.  The first refers to the cognitive bias in which relatively unskilled persons suffer illusory superiority.  The second part refers to a cognitive bias for highly skilled individuals to underestimate the relative competence of unskilled individuals and assume that tasks which are easy for them are also easy for others.

HM will address the second part first.  A fundamental difficulty HM has in teaching is to overestimate what students do and can understand.  HM is not implying that these students are stupid, although this might be the simplest explanation.  However, it is the teacher’s responsibility to teach to the level of what the student can understand.  As a result of repeated experience with students of a certain level, the teacher can and should identify the appropriate level to teach and proceed accordingly.

Dunning and Kruger were not the first to recognize this effect.  Confucius said, “Real knowledge is to know the extent of one’s ignorance.”  Bertrand Russell said, “One of the painful things about our time is that those who feel certainty are stupid, and those with any imagination and understanding are filled with doubt and indecision.”  This statement reminds HM of the phrase, “Ignorance is bliss.”  Charles Darwin wrote, “Ignorance more frequently begets confidence than does knowledge.”  Shakespearean “As You Like It”  wrote “The Foole doth think he is wise, but the wise man knows himself to be a Foole.”

Trump followers appear to be extremely confident in Trump.  How anyone can be confident in Trump given the content of the previous Healthymemory blog posts is completely baffling to HM.  But then, HM is assuming that Trump followers have knowledge that they don’t have.

It would be interesting to have discussion groups with Trump devotees.  The objective of these discussions would not be to try to persuade them to change their opinion, but rather to discuss how the different branches of government work, the role of the Constitution and the Supreme Court.  There would also be discussion regarding the economy, foreign trade, and the subtleties and intricacies of international relations.
I think the results of these discussion group would be extremely depressing.  But they would also be informative.

Palatable, informational presentations might actually urge these followers to think and to invoke their System 2 processes.  Arguing directly regarding the potential disaster Trump could cause the county will not work because people will become defensive.  However, for those who can actually be induced to think might change their minds on their own.

There is some evidence that the Dunning Kruger Effect might be specific to western cultures.  A number of studies using East Asian participants suggest that different social forces are at play in difference cultures.  East Asians tend to underestimate their abilities and see underachievement as a chance to improve themselves and to get along with others.  If only this attitude could be fostered in our culture.

Another Western culture showing the Dunning-Kruger Effect is Great Britain’s Brexit vote.  The Prime Minister assumed that a reasoned discussion of the benefits from remaining in the EU versus the costs in leaving the EU would result in a vote to remain, but just the opposite occurred.  One problem was that a reasoned discussion did not take place.  Rather it became a rowdy political contest in which lies and misrepresentation were made.  HM needs to bring Kahneman’s two process view of cognition into this discussion.  Remember that System 1, intuition, is fast, emotional, and our default mode for processing.  System 2, called reasoning, is slow and effortful.  It became clear that remain arguments had the flavor of System 2 processing.  They were well-reasoned and thought out and supported by data.  Unfortunately, exit arguments smack primarily of System 1 processes that were largely emotional.   They wanted to be British and they wanted to prevent immigration.

For more on the Dunning-Kruger effect and for more specific references see the Wikipedia.

© Douglas Griffith and, 2016. Unauthorized use and/or duplication of this material without express and written permission from this blog’s author and/or owner is strictly prohibited. Excerpts and links may be used, provided that full and clear credit is given to Douglas Griffith and with appropriate and specific direction to the original content.