Mathematics
Inferential Statistics
Quick fact
Inferential statistics allows researchers to make claims about millions of people after studying only a few hundred, with a known margin of error.
Why this is interesting
You can know a lot about a whole group by studying just a small piece of it—but how do you know if what you see is real or just luck?
Read the full explanation
Understanding Inferential Statistics
Imagine you want to know if a new teaching method improves test scores. You can't test every student in the country, so you take a sample. Inferential statistics provides the tools to use that sample to make a reasonable guess about the entire population—and to calculate how likely your guess is to be wrong. Think of tasting a spoonful of soup to decide if the whole pot needs salt. The spoonful is your sample; your judgment is the inference. But sometimes you might get a spoonful that is not representative (e.g., a piece of salt). Inferential statistics tells you how much to trust your spoonful.
A deeper explanation
The mechanism relies on the concept of a sampling distribution. If you took many samples from the same population, their statistics (e.g., means) would form a distribution around the true population value. The spread of that distribution—the standard error—tells you how much sample results typically vary. With that, you can construct a confidence interval: a range that likely contains the true population value. Hypothesis testing formalizes this: you start with a null hypothesis (e.g., 'the new method makes no difference') and ask how likely your observed data would be if the null were true. If that probability (the p-value) is very low, you reject the null and conclude your finding is statistically significant. This process guards against being fooled by random chance and is why inferential statistics is essential for scientific discovery, A/B testing, opinion polls, and quality control.