How Simpson's Paradox REVERSES Reality

Описание: Can data lie to you? Even though a graph might imply one thing, it can actually be the opposite.

In this video, we will explore Simpson’s Paradox. A statistical paradox where the core idea is this: when we aggregate data, trends disappear. This isn’t just a trick either, it is a hidden pattern which impacts medical trials, university admissions, and how we interpret the world around us.

Without looking for hidden groups of variables in the data, the initial presentation of data might lie to you without you even knowing. This is because there may be a hidden dominant group skewing the average.

In this video, I break down why these graphs can quietly mislead you, using both hypothetical and real world scenarios, and how to think clearer about the data before making a decision.

In this video you’ll learn about:
-The story of Edward H. Simpson and Bletchley Park
-Simpsons Paradox
-What a confounding variable is
-The UC Berkeley admissions Scandal

Timestamps:
0:00 The Impossible Medical Trial
0:34 The Story of Edward Simpson and Simpsons Paradox
2:16 The Impossible Medical Trial Explained
5:13 Correlation Versus Causation
6:33 UC Berkeley Admissions Scandal
8:53 Questions to Always Ask and Challenge

Here are some more videos of mine that you may find interesting:
Why the average does not exist -    • Why The Average Person Does Not Exist  
Normal Distribution -    • Are You Really in the Top 10%? (Percentile...  
Standard Deviation -    • Understand Standard Deviation Quick (Expla...  
P Values -    • Understand P Values Quick in Under 3 Minutes!  
Confidence Intervals -    • What Confidence Intervals Actually Mean  
Central Limit Theorem -    • Why Bell Curves Appear Everywhere (The Cen...  
Type 1 and 2 Statistical Errors Explained    • The Two Ways You Can Be Wrong With Data (T...  

Like, comment, and subscribe for more!


Relevant Sources and other further readings if you're interested!

(Stanford Encyclopedia): Sprenger, Jan and Naftali Weinberger. (2021). "Simpson’s Paradox." The Stanford Encyclopedia of Philosophy.

(Journal of Statistical Distributions): Selvitella, A. (2017). "The ubiquity of the Simpson’s Paradox." Journal of Statistical Distributions and Applications.

(Queens University Belfast) Edward Simpson Obituary (2019).
Accessed at https://alumni.qub.ac.uk/edward-simps...

UC Berkley Data Adapted from: (Berkeley/Original Paper): Bickel, P.J., Hammel, E.A., & O'Connell, J.W. (1975). "Sex Bias in Graduate Admissions: Data From Berkeley." Science.

UC Berkeley Dataset you can do your own analysis with: University of Illinois (Data/Dataset). "Simpsons Paradox." Basics of Data Science with Python. https://discovery.cs.illinois.edu/lea...

Disclaimer:
This video is for educational purposes only. To clearly illustrate Simpson’s Paradox, the scenarios and datasets used have been adapted and simplified. The conclusions presented and derived in this video are for statistical demonstration and are not intended to be a comprehensive sociological or medical analysis of any historical case mentioned.

#statistics #education #story #datascience #history #paradox #interestingfacts