⭐ Total Points: 0
Back to Lessons The Decision Tree 0 pts Module 2 · Lesson 6
Introduction

The Decision Tree

Pull it all together

The Finish Line

You've learned the four key questions. Now we combine them into a single decision tree that will guide you to the right statistical test every time.

No more guessing. No more "I think it's a t-test?" Answer four questions, get the right test.

By the end of this lesson, you'll have a systematic approach you can use for any analysis.

The Four Questions

Every test selection comes down to these four questions, asked in order.

1
What type is your OUTCOME variable?
Continuous · Categorical · Ordinal · Time-to-event
2
How many GROUPS are you comparing?
One · Two · Three or more
3
Are the groups PAIRED or INDEPENDENT?
Paired (same subjects) · Independent (different subjects)
4
Are the ASSUMPTIONS met?
Normal distribution · Equal variances · Sample size

Questions 1-3 narrow you down to a test family. Question 4 tells you whether to use the parametric or non-parametric version.

Note on Assumptions

Before using the decision tree, understand this important caveat.

Parametric vs Non-parametric

The decision tree gives you the parametric test by default. Parametric tests assume your data follow a normal distribution and have adequate sample sizes.

If these assumptions aren't met, use the non-parametric alternative instead:

t-test → Mann-Whitney U (independent) or Wilcoxon signed-rank (paired)
One-way ANOVA → Kruskal-Wallis
Repeated measures ANOVA → Friedman test

When in doubt, the non-parametric test is usually the safer choice. It works even when assumptions are met, just with slightly less power.

Interactive Decision Tree

Try it out. Answer the questions and see which test you get.

What type is your outcome variable?
How many groups are you comparing?
Are the observations paired or independent?
Parametric
Non-parametric

Quick Reference Chart

Keep this chart handy. It covers the most common scenarios.

Outcome Groups Design Parametric Non-parametric
Continuous 1 vs. standard One-sample t-test Wilcoxon signed-rank
Continuous 2 Independent Independent t-test Mann-Whitney U
Continuous 2 Paired Paired t-test Wilcoxon signed-rank
Continuous 3+ Independent One-way ANOVA Kruskal-Wallis
Continuous 3+ Paired RM ANOVA Friedman
Categorical 2 Independent Chi-square / Fisher's exact
Categorical 2 Paired McNemar's test
Categorical 3+ Independent Chi-square
Ordinal 2 Independent Mann-Whitney U
Ordinal 3+ Independent Kruskal-Wallis
Time-to-event 2+ Any Log-rank test / Cox regression

When in doubt: If outcome is continuous and normal, use t-test/ANOVA. If non-normal or ordinal, use Mann-Whitney/Kruskal-Wallis. If outcome is time-to-event, use survival methods.

Exercise: Choose the Right Test

Use the decision tree to select the appropriate test for each scenario.

Question 1 of 8

Lesson Complete!

0
Total Points Earned
Exercise (0/8 correct) +0 pts
Lesson Completed +100 pts
✓

The Decision Tree

Lesson 6 of 6 complete

Key Takeaways

  • Four questions: Outcome type → Number of groups → Paired or independent → Assumptions
  • Parametric vs non-parametric: Use non-parametric when assumptions aren't met
  • The decision tree: A systematic approach to choosing the right test every time

The Decision Process

1. What type is my outcome? → 2. How many groups? → 3. Paired or independent? → 4. Assumptions met?

Follow this process every time, and you'll choose the right test.