You almost never see all the data. Statistical inference is how you use a sample to draw careful conclusions about the wider population, and how you tell a real effect from random noise. It underpins A/B tests, model comparisons and any claim that one model beats another.
The lessons begin with the null hypothesis and p-values, then cover confidence intervals and choosing the right t-test. They finish with statistical power, which tells you how likely a test is to find a real effect, and bootstrapping, a flexible way to estimate uncertainty from the data you have.
Watch the lessons in order, then continue on to Bayesian Thinking.
Introduces hypothesis testing and the role of the null hypothesis.
Video 14minExplains what a p-value measures and how to read one correctly.
Video 11minShows how confidence intervals describe the uncertainty in an estimate.
Video 6minExplains statistical power and why it matters for finding real effects.
Video 8minIntroduces bootstrapping, a way to estimate uncertainty by resampling your data.
Video 9min