Summary and practical aspects

Last updated on 2024-03-12 | Edit this page

Estimated time: 3 minutes

I hope this tutorial allowed you to notice a pattern in the tests you have seen.

Image of a cookbook for statistical tests
Cookbook (image adapted from kindpng.com)

The “cook book” for a test goes as follows:

  1. Set up a hypothesis \(H_0\) that you want to reject.

  2. Find a test statistic that should be sensitive to deviations from \(H_0\).

  3. Find the null distribution of the test statistic – the distribution that it follows under the null hypothesis.

  4. Compute the actual value of the test statistic.

  5. Compute the p–value: The probability of seeing a value as least as extreme as the computed value in the null distribution.

  6. Decide (based on significance level) whether to reject the null hypothesis.

Cook book in practice

In practice, you usually don’t have to perform steps 2-6 on your own – standard procedures, i.e. statistical tests, exist for most kind of data, and the computer does the calculation for you. So… what is left for you to do?

Computer screen executing wilcoxon test
In practice (image adapted from kindpng.com)

  1. Look at your data!

  2. Decide on a distribution that your data follow.

  3. Possibly transform your data to match a suitable distribution (suitable means: a convenient test is available for this distribution).

  4. Find a test that answers your question and is suitable for the distribution the properties of your data.

  5. Perform the test and decide. Report the p-value and the effect size.