On this page
What this lesson covers
Parametric tests use distributional parameters and assumptions; non-parametric tests offer alternatives when ranks or departures from those assumptions make parametric methods unsuitable.
Choose a test family from its assumptions and the measurement scale. Examine outliers, ordinal ratings and the power trade-off when parametric assumptions do hold.
Key points
- Parametric: Tests like the t-test, the chi-square test of one variance, and the F-test are parametric.
- Non-parametric tests: Non-parametric tests do not require the data to follow any particular distribution.
- Outliers: Outliers add a second problem, because they can badly distort the observed mean and variance, which are the very inputs a parametric test uses.
- Power: If all the assumptions of the parametric test are met, the non-parametric test is almost always less effective at detecting a true effect.
For example, the consultant gives each of 200 external managers a rating from 1 to 5 stars. The gap between 3 and 4 stars is not a measured amount, so these ordinal ratings call for a non-parametric method that uses their order.
Common mistakes
- Treating star-rating labels as measured amounts assigns a measurable gap to what is only an ordering.
- Ruling out parametric inference solely because the population is skewed is incorrect.
Try a practice question
Choose an answer, then check it to read the explanation for every option.
Continue studying in Quizara
In Quizara the lesson comes with more practice questions on the topic, explanations for every option and an AI tutor for your follow-up questions.
Written by the Quizara Editorial Team under our editorial standards.