Statistical Distributions
Mathematical functions describing the probability of different outcomes, forming the foundation of statistical analysis and decision-making.
Also known as: Probability Distributions, Distribution Types
Category: Thinking
Tags: statistics, mathematics, data-analysis, mental-models, probability
Explanation
Statistical distributions are mathematical functions that describe the probability of different outcomes in a dataset or random process. They are the foundation of statistical analysis, enabling us to model uncertainty, make predictions, and draw conclusions from data.
## Key Distributions
### Normal (Gaussian) Distribution
The famous bell curve. Most values cluster around the mean, with symmetric tails. Describes many natural phenomena: heights, test scores, measurement errors. The Central Limit Theorem explains why it appears so frequently — the average of many independent random variables tends toward normal distribution regardless of the underlying distribution.
### Power Law Distribution
A few items account for most of the effect (the 'long tail'). Examples: wealth distribution, city sizes, word frequencies, website traffic. Unlike normal distributions, extreme events are far more likely. The Pareto Principle (80/20 rule) is a manifestation of power law distributions.
### Uniform Distribution
All outcomes are equally likely. Examples: fair dice rolls, random number generators.
### Exponential Distribution
Describes time between events in a Poisson process. Examples: time between customer arrivals, radioactive decay, time between website visits.
### Binomial Distribution
Describes the number of successes in a fixed number of independent yes/no trials. Examples: coin flips, quality control pass/fail rates.
## Why Distributions Matter
- **Decision-making**: Understanding whether outcomes follow a normal or power law distribution fundamentally changes strategy. In normal distributions, extreme events are rare. In power law distributions, they dominate
- **Risk assessment**: Fat-tailed distributions (power laws) mean rare events are more common than normal distributions predict — crucial for finance, security, and engineering
- **Data analysis**: Choosing the right statistical test requires knowing the underlying distribution
- **Mental models**: Recognizing which distribution governs a situation prevents applying the wrong intuition
## Common Mistakes
Assuming everything is normally distributed is one of the most dangerous statistical errors. Many real-world phenomena (income, returns, earthquakes, viral content) follow power law or other fat-tailed distributions where extreme events are far more likely than a bell curve would predict.
Related Concepts
← Back to all concepts