Micro Free Course — 100% Free Learning

All lessons in this module are free to learn. Sign in with Google to save your progress.

Reading Inference Without Doing the Math

P-values, confidence intervals, Type I/II errors, statsig vs shipsig, and the five stats traps that fool analysts daily.

Module Progress0% Complete

100 min total

15 Lessons

0 Completed

Module Content

P-values & Null Hypothesis - What They Actually Mean

Decoding 'p < 0.05' in plain English. The four most common misinterpretations. FIND_VIDEO: search 'StatQuest p-value null hypothesis' - recommended channel: StatQuest with Josh Starmer. Aim for 12 min or under.

12 minVideo

Start

Reading P-values Without Getting Fooled

P-values are the most misinterpreted number in applied stats. Learn what they do and don't mean.

6 minArticle

Start

Reading P-values Without Getting Fooled Quiz

Test what you just read.

4 minTutorial

Start

Confidence Intervals - The Most Useful Thing in Applied Stats

Why CIs beat p-values for communicating results to non-analysts. FIND_VIDEO: search 'StatQuest confidence interval' - recommended channel: StatQuest with Josh Starmer. Aim for 10 min or under.

10 minVideo

Start

Reading Confidence Intervals

A confidence interval tells you the size of an effect AND how uncertain you are about it. P-values can't do that.

5 minArticle

Start

Reading Confidence Intervals Quiz

Test what you just read.

3 minTutorial

Start

Type I vs Type II Errors - False Positives, False Negatives

Which error costs more depends entirely on the business decision - and most A/B tests are biased toward Type II. FIND_VIDEO: search '365 Data Science Type 1 error vs Type 2 error' - recommended channel: 365 Data Science. Aim for 9 min or under.

9 minVideo

Start

Which Error Costs You More?

Type I and Type II errors aren't equally bad - it depends on the decision. Learn the asymmetric cost frame.

5 minArticle

Start

Which Error Costs You More? Quiz

Test what you just read.

3 minTutorial

Start

Statistical vs Practical Significance

When a 'statistically significant' result has zero business meaning - and what to do about it. FIND_VIDEO: search 'statsig vs practical significance effect size' - recommended channel: Crash Course Statistics. Aim for 8 min or under.

8 minVideo

Start

Statsig vs Shipsig

With enough data, you can find a 'significant' lift of 0.1%. That doesn't mean you should ship it.

5 minArticle

Start

Statsig vs Shipsig Quiz

Test what you just read.

3 minTutorial

Start

Five Stats Traps That Will Fool You

Simpson's paradox, base-rate fallacy, regression to the mean, multiple testing, and selection bias - all in one short tour. FIND_VIDEO: search 'Simpson's paradox base rate fallacy regression to the mean' - recommended channel: MinutePhysics. Aim for 14 min or under.

14 minVideo

Start

The Five Traps

The five stats traps you'll actually run into. Memorize these - they show up in every interview and most data reports.

8 minArticle

Start

The Five Traps Quiz

Test what you just read.