Anscombe's Quartet: When Numbers Lie
Anscombe's Quartet shows how four datasets can share identical summary stats (mean, variance) but look completely different when plotted. It's a classic reminder to always visualize your data before trusting numerical summaries.
The mental model
Anscombe's Quartet is a set of four datasets that are statistically identical but visually distinct. It's a powerful reminder that summary statistics alone can be dangerously misleading, and that you must always visualize your data before drawing conclusions.
How it works
Created by statistician Francis Anscombe in 1973, the quartet consists of four sets of eleven (x, y) data points. If you only calculate the simple descriptive statistics, they look the same. They share nearly the same mean of x, mean of y, variance of x, variance of y, correlation between x and y, and linear regression line. However, when you plot them on a graph, their true, different natures are revealed.
When to use it
The principle behind Anscombe's Quartet should be applied in any situation involving data analysis. Three key moments are: first, during exploratory data analysis, to understand the shape and structure of your data; second, before building any statistical or machine learning model, to check for outliers and non-linear relationships; and third, when interpreting the results of A/B tests or system metrics, to ensure an aggregate number isn't hiding a critical underlying pattern.
When not to use it
The principle of visualizing data is almost universally applicable. However, it is not a replacement for statistical rigor; it is a necessary complement to it. Relying only on graphs without understanding the underlying statistics can also be misleading, especially with complex, high-dimensional data that is difficult to plot. The goal is to use visualization and numerical calculation together, not to choose one over the other.
One canonical example
The four plots of the quartet reveal the truth the numbers hide. The first dataset is a simple, noisy linear relationship, the kind you'd expect. The second shows a clear non-linear, curved relationship; a straight line is a poor fit. The third shows a perfect linear relationship, but with a single, massive outlier that skews the regression line. The fourth shows that all data points except one have the same x-value, with that one outlier being so influential it single-handedly determines the slope of the regression line.
Interview question
What fundamental lesson does Anscombe's Quartet teach about data analysis?
- a.Outliers are the main reason why statistical summaries can be misleading.
- b.Linear regression is frequently an unsuitable model for diverse real-world data.
- c.Datasets with identical summary statistics can hide profoundly different underlying structures.Correct
- d.Statistical rigor should always take precedence over data visualization.
Why? this is the answer
Anscombe's Quartet explicitly shows that datasets can have identical summary statistics but vastly different visual patterns, underscoring the critical need to visualize data. While outliers are a factor in some of the quartet's examples, the overarching lesson is the general insufficiency of summary statistics alone, not just outlier detection.
Just read this? Test yourself on what you have been reading.
Read the original → en.wikipedia.org
You just looked this up. Could you explain it out loud?
That is the part interviews actually test. Tezvyn takes questions like this one and gives you what the interviewer is really checking, the answer that lands, and the mistake that ends the conversation, in the four minutes before your next meeting.
The iPhone app is on the way
We are building it. Until it lands, nothing here is held back from you: every interview card, your saved cards, streaks and the job board all work in Safari, plus hundreds of free practice quizzes of thirty questions each. Sign in and it all carries over to the app the day it arrives.
Want it as an icon? Tap Share at the bottom of Safari, then Add to Home Screen. It opens full screen and the cards you have read stay available offline.
We are hiring for this. Open roles that interview on statistics — each one lists the topics its interview covers.
See open roles