"It is important to be critical of the statistics that are presented to us and consider the manipulation of the electorate" / Juan Carlos Erdozáin
"It is important to be critical of the statistics presented to us and carefully review how the data was collected, what methods were used to analyze it and what conclusions were drawn, also considering manipulation of the electorate prior to elections, ranging from handouts to extortion. moral, emotional and physical, which is reflected in distorted results" / Juan Carlos Erdozáin
When math challenges our first impressions, we often blame ourselves or a teacher who confused us in the past. In reality, what we need is to carefully review our basic ideas and understand how mathematical concepts are constructed. After all, in an environment where understanding is scarce, anyone with a little knowledge can look like an expert.
💡There is a very interesting book in English entitled "Coincidences, Chaos, and all That Math Jazz" by Edward B. Burger and Michael Starbird, which made me reflect on the way presidential polls are overrated in the world.
"There are three types of lies: lies, damned lies and statistics"
This phrase, often attributed to Benjamin Disraeli, highlights the power of statistics to manipulate and distort reality. While statistics can be a valuable tool for understanding the world, they can also be used to deceive and confuse, especially when they are selectively presented or misinterpreted.
💡 It is important to be critical of the statistics presented to us and carefully consider how the data was collected, what methods were used to analyze it, and what conclusions were drawn. Only then can we discern whether the statistics offer us an accurate picture of reality or whether they are simply another type of lie.
In addition, aspects of manipulation of the electorate prior to elections must be considered, ranging from handouts to moral, emotional and physical extortion, which is reflected in distorted results.
💡 Why statistics can be misleading?
✒️ Data selection: Sometimes, only the data that supports an argument is selected, ignoring those that contradict it.
✒️ Chart Manipulation: Charts can be designed to exaggerate differences or trends. As has been observed in the presentations made by the president in turn every day.
✒️ Correlation vs. Causality: Two things can be statistically related without one causing the other.
✒️ Misleading averages: Using averages can hide important variations within a data set.
💡 How can we protect ourselves from statistical "damn lies"?
✔️ Being critical: Do not blindly accept the statistics presented, but rather question their origin and methodology.
✔️ Find reliable sources: Use data sources recognized for their rigor and objectivity, being careful with some media that are sold at the orders of those who hire them.
✔️ Consider the context: Interpret the statistics in the broader context of the topic being discussed.
✔️ Learn about statistics: Knowing the basics of statistics helps us better understand how data can be used and misinterpreted and ignoring gossip, rumors and false assumptions.
💡Areas of everyday life where we find examples of misleading statistics
Advertising: Companies often use statistics to make their products seem more effective or attractive.
News: The news sometimes presents statistics in a sensational or simplified way to attract attention.
Politics: Politicians may use statistics to support their arguments or attack their opponents.
💡Presidential Polls: Digesting the data of life
🔺If you wanted to read well-digested literature at the beginning of the 20th century, then the Literary Digest was the publication you were looking for. However, his fame endures because of a statistical fiasco he inadvertently perpetrated in 1936, a fiasco that will keep the Literary Digest alive in the statistics textbooks of future generations.
🔺The setting was the 1936 United States presidential elections; the mission of the Literary Digest, to predict the outcome. Most people alive today did not vote in that election, but those who were present may remember that the two leading candidates were Franklin Delano Roosevelt, the incumbent, and Alfred Landon, the Republican opponent.
🔺In each of the previous five presidential elections, Literary Digest had correctly predicted the winner and had come within a few percentage points of getting the actual voting margin right. For the 1936 election, the Digest sent millions of surveys to voters across the country. The evidence was clear and the prediction was reliable: Landon would easily win the White House. In fact, the publication predicted that Alf would receive 57% of the popular vote and win the Electoral College vote by an overwhelming majority, 370 to 161.
🔺 Most people won't remember studying President Alfred Landon in their American history class, for the simple reason that Landon didn't win. Literary Digest's predictions were correct in only one respect: the election was a landslide, but it went the other way. Roosevelt was re-elected overwhelmingly with 62% of the popular vote and won the Electoral College by a remarkable vote of 523 to 8. How could the then-former Literary Digest statisticians be so completely wrong? Simple: they asked the wrong people for their opinions.
🔺The Literary Digest developed lists of people to send surveys to from several sources, including Digest subscription records, automobile registration records, and telephone records. Ten million surveys were sent and two million were returned. But 1936 fell in the middle of the Great Depression. Many households were cutting unnecessary expenses, and unfortunately (for the health of the Literary Digest), subscriptions to that esteemed publication may have been among the first victims of the budget cut.
Many families cut back further, giving up cars and phones. So the people on the poll list were not representative of the entire voting public. Additionally, only surveys that were returned voluntarily were counted. Who knows if the people who return the surveys are a representative sample of the public? In any case, the survey was tremendously biased and the deductions made were grossly erroneous.
🔺 GEORGE GALLUP. The Literary Digest fiasco had another interesting result: it catapulted a young statistician to lasting fame. George Gallup learned about the Literary Digest poll and its methods. Suspecting that the survey would be flawed, he himself surveyed 50,000 people using a method that would later become a standard feature of good surveys: pure randomness.
He chose his survey participants at random from the electoral rolls and discovered that his survey presented a radically different picture. Not only did he predict a resounding victory for Roosevelt, he also predicted what the Literary Digest's prediction would be. So he announced in advance that the Literary Digest prediction would be wrong, and he was right. George Gallup saw the flaw in collecting data from rich people (which led him to become rich himself). What he may not have foreseen is that the "Gallup poll" would become a commonly used term.
🔺The Literary Digest made the mistake of gathering its information from biased sources. Today, experts are more sensitive in their statistical studies, but they are still prone to drawing erroneous conclusions, so we must remain on our statistical guard.
The Literary Digest fiasco is an example of poor statistical methods. No professional today would make that particular mistake unless they did so intentionally. Suppose, for example, you wanted to write a compelling report supporting your view that almost all Americans drive dangerously. One way to gather data to support your position is to simply make your observations in the parking lot of a defensive driving school, but be careful where you stand.
In the contemporary era, it is known that there are random factors that, when voting, the voter can still reconsider:
🔳Vote of punishment for the official candidate and the number of voters who participate, among whom may be the resentful and undecided who exercise a vote of punishment for the official candidate.
🔳Fidelity and trust of voters / Collective Unconscious, how much does the voter identify emotionally with a candidate, with the history of the nation and its heroes?
🔳Sensitive and inclusive leadership (image), which somehow translates into charisma, histrionic capacity and empathy with voters.
🔳 Voter's gratitude to the candidate who provides him with handouts, a deception that in a figurative sense is "Giving fish to the Poor instead of teaching them to Fish"
💡Conclusions
✒️ It is important to be critical of the statistics presented to us and carefully consider how the data was collected, what methods were used to analyze it, and what conclusions were drawn. Only then can we discern whether the statistics offer us an accurate picture of reality or whether they are simply another type of lie.
Politicians can use statistics to support their arguments or attack their opponents.
✒️ Remember the Literary Digest case who was diametrically wrong because they asked the wrong people for their opinions.
✒️ Random factors
There are random factors that at the polling station and at the time of voting, the voter can still reconsider.
✒️Manipulation factors
Prior to elections, which range from handouts to moral, emotional and physical extortion, which is reflected in distorted results.
FUENTES:
Edward B. Burger y Michael Starbird. "Coincidences, Chaos, and all That Math Jazz"
Forbes: Violencia, la mayor preocupación de más de la mitad de los mexicanos: sondeo. https://www.forbes.com.mx/violencia-la-mayor-preocupacion-de-mas-de-la-mitad-de-los-mexicanos-sondeo/
New York Times: México se dispone a elegir a su primera Presidenta. https://www.nytimes.com/es/2024/05/30/espanol/claudia-sheinbaum-mexico-elecciones.html
El Pais: Elecciones Mexicanas. https://elpais.com/mexico/elecciones-mexicanas/2024-05-28/sheinbaum-llega-al-final-de-la-campana-como-clara-ganadora-y-maynez-supera-a-galvez-en-simpatia-ciudadana.html
Gallup: Mexico Votes: 5 Things to Know Ahead of Election: https://news.gallup.com/poll/645167/mexico-votes-five-things-know-ahead-election.aspx.aspx
Comments