Field note · August 4, 2026
How much statistics do I actually need?
Less theory than you fear, and more discipline than you would like.
A question that came in more than once, so the answer goes here.
How much statistics do I actually need?
Less theory than you fear, and more discipline than you would like.
The working set is small: what a distribution looks like and why the mean can mislead, sampling variation and confidence intervals, the difference between correlation and cause, confounding, and multiple comparisons. That is five ideas and it covers the overwhelming majority of what gets asked.
What is not on the list: hypothesis test taxonomies, distribution families, most of a first-year syllabus.
What matters more than any of it is the habit of asking "how sure are we, and what would change the answer" before presenting a number.
The two that catch people out in practice are confounding and multiple comparisons, and neither is hard — they are just easy to not think about. Confounding because the interesting comparison is almost never randomised, and multiple comparisons because a dashboard with forty metrics on it is forty tests, run daily, and something on it is always significant.
The statistics module is four lessons and covers the working set.