All 62 statistics calculators

Sixty-two tools is more than anyone wants to browse, so here is the shape of the collection.

If you are describing data you already have, you want the descriptive family: descriptive statistics computes everything at once, mean, median and mode explains the three centres, and standard deviation and percentiles cover spread. Draw it with a histogram or a box plot before you conclude anything — shape decides which methods are appropriate.

If you are asking whether a difference is real, you want a hypothesis test, and which one depends on four things: whether you are comparing means or proportions, how many groups there are, whether the observations are paired, and whether the data are roughly normal. The decision guide walks through exactly that, and it takes about a minute.

If you are planning a study or an experiment, start before you collect data: the sample size calculator says how many observations you need, and the minimum detectable effect says what you would be able to see with the sample you can actually get. Both are far more useful before the study than after it.

If you are working with probability rather than data, the distribution calculators — normal, binomial, Poisson — turn parameters into probabilities, and Bayes' theorem handles updating a belief when evidence arrives.

What every page here includes

The same eight things, in the same order: the head term as the heading, a one-sentence statement of what the tool does, the calculator itself, a plain reading of what the result means and does not mean, the formula with every term defined, a worked example with real numbers computed step by step, the assumptions along with links to the test to use when they fail, and the questions people actually ask.

Every number is verified against R, nothing you type is transmitted anywhere, and every calculation happens in your own browser.