How Polls Get It Wrong (Statistics)

Crash Course: How Polls Get It Wrong (Statistics)

How Polls Get It Wrong: A Crash Course in Statistics

Introduction Imagine you're a presidential candidate, and you've just lost an election by a landslide. But here's the kicker: you won the popular vote by a whopping 3 million votes. Sounds crazy, right? That's exactly what happened in the 2000 US presidential election between Al Gore and George W. Bush. Welcome to the wild world of polls gone wrong!

The Core Idea Polls are supposed to give us a snapshot of public opinion, but sometimes they're way off the mark. In this crash course, we'll explore why polls get it wrong, from sampling errors to biases and beyond. Buckle up, folks!

Key Facts & Figures

• The Father of Statistics: Karl Pearson, a British mathematician, laid the groundwork for modern statistics in the late 19th century.
• The First Poll: In 1936, George Gallup conducted the first modern poll, predicting Franklin D. Roosevelt's landslide victory over Alf Landon.
• Sampling Error: Even with a sample size of 1,000, there's a 3% margin of error – that's 30 people out of 1,000 who might be wrong.
• Bias: In 1948, Gallup's poll predicted Thomas Dewey would win the presidential election, but Harry Truman pulled off a stunning upset.
• The Challenger: In 1980, a challenger poll predicted Ronald Reagan would win the Republican primary, but George H.W. Bush won instead.
• The Challenger Effect: In 2016, a challenger poll predicted Donald Trump would win the Republican primary, but Ted Cruz won in Iowa.
• The Polling Industry: In the 1990s, polling firms began to consolidate, leading to a decrease in competition and innovation.
• The Internet Effect: In the 2000s, online polling became more prevalent, but it also introduced new biases and errors.
• The 2016 Election: Polls predicted Hillary Clinton would win the presidential election, but Donald Trump pulled off a stunning upset.
• The Margin of Error: In 2018, a poll predicted a Democratic wave in the midterm elections, but Republicans actually gained seats.
• The Polling Industry's Response: In 2020, polling firms began to incorporate new methods, such as online panels and machine learning algorithms.

Thought Bubble Imagine you're a pollster, and you're trying to predict the outcome of a presidential election. You've got a sample size of 1,000, and you're using a combination of online and phone surveys. But here's the thing: your sample might be biased towards older, more affluent voters, who are more likely to respond to polls. And what about the people who don't have landlines or smartphones? They're not represented in your sample. You're also relying on self-reported data, which can be unreliable. And let's not forget about the margin of error – 3% might not seem like a lot, but it's enough to swing an election.

Why This Matters

• Democracy: Polls are supposed to give us a snapshot of public opinion, but when they're wrong, it can lead to a crisis of confidence in our democratic institutions.
• Elections: Polls can influence the outcome of elections, either by predicting a winner or by discouraging voters from participating.
• Policy: Polls can inform policy decisions, but when they're wrong, it can lead to bad policy choices.
• Media: Polls are often used in the media to create a narrative or to predict a winner, but when they're wrong, it can lead to a media frenzy.
• Public Perception: Polls can shape public perception of issues and candidates, but when they're wrong, it can lead to a distorted view of reality.
• The Polling Industry: The polling industry is a multi-billion dollar industry, and when polls go wrong, it can have serious consequences for the industry as a whole.
• Accountability: Polls can hold politicians and policymakers accountable, but when they're wrong, it can lead to a lack of accountability.

Crash Course Recap

• Polls are supposed to give us a snapshot of public opinion, but they can be wrong due to sampling errors, biases, and other factors.
• Karl Pearson is considered the Father of Statistics, and his work laid the groundwork for modern statistics.
• The first modern poll was conducted by George Gallup in 1936.
• Sampling error can lead to a 3% margin of error, which is enough to swing an election.
• Bias can lead to inaccurate results, and online polling has introduced new biases and errors.
• The polling industry has consolidated, leading to a decrease in competition and innovation.
• The 2016 election was a wake-up call for the polling industry, and firms have begun to incorporate new methods.
• Polls can influence the outcome of elections, inform policy decisions, and shape public perception.
• The polling industry is a multi-billion dollar industry, and when polls go wrong, it can have serious consequences.

Quiz Yourself

1. Who is considered the Father of Statistics? a) Karl Pearson b) George Gallup c) Thomas Dewey d) Harry Truman

Answer: a) Karl Pearson

1. What is the margin of error for a sample size of 1,000? a) 1% b) 3% c) 5% d) 10%

Answer: b) 3%

1. What is the name of the first modern poll? a) The Gallup Poll b) The Challenger Poll c) The 1936 Poll d) The Father of Statistics Poll

Answer: a) The Gallup Poll

1. What is the name of the 2016 presidential candidate who won the popular vote but lost the election? a) Hillary Clinton b) Donald Trump c) Al Gore d) George W. Bush

Answer: a) Hillary Clinton

1. What is the name of the method that polling firms have begun to incorporate to reduce bias and errors? a) Online panels b) Machine learning algorithms c) Phone surveys d) All of the above

Answer: d) All of the above