"Which statistical test should I use?" is probably the single most common question in intro statistics — and the answer isn't about the numbers at all. It's about answering three questions about your data and your research question, in order, before you touch a formula. Getting this right is really the foundation of hypothesis testing: pick the wrong test and no amount of careful calculation afterward fixes it.
This statistical test decision tree walks through the three questions in order, so choosing statistical tests stops being a guessing game — the same process we use in statistics homework help sessions.
The decision tree: 3 questions to ask first
- What type of data do you have? Categorical (yes/no, categories) or continuous (numbers on a scale, like height or test scores)?
- How many groups or variables are you comparing? One group against a known value, two groups, three or more groups, or the relationship between two continuous variables?
- Are your groups independent, or paired? Independent means different people/items in each group; paired means the same subjects measured twice (before/after).
Quick-reference: matching your answers to a test
- Comparing one group's average to a known value (continuous data) → one-sample t-test
- Comparing two independent groups' averages (continuous data) → independent-samples t-test
- Comparing the same group before and after (continuous data) → paired-samples t-test
- Comparing three or more groups' averages (continuous data) → ANOVA
- Testing a relationship between two categorical variables → chi-square test
- Testing a relationship between two continuous variables → correlation or regression
Worked example
A researcher wants to know if average exam scores differ across three teaching methods (lecture, flipped classroom, online).
Step 1: Data type? Exam scores are continuous.
Step 2: How many groups? Three teaching methods = three groups.
Step 3: Since three continuous-data groups are being compared, this calls for ANOVA, not a t-test.
Common mistakes
ANOVA vs t-test: running multiple t-tests instead of one ANOVA. Comparing three groups with three separate t-tests inflates your chance of a false positive (the multiple comparisons problem). If you have 3+ groups, that's an ANOVA question, not three t-test questions.
Using a t-test on categorical data. If your outcome is "yes/no" or a category rather than a number, you likely need a chi-square test, not a t-test — check your data type first, every time.
Stuck on a statistics assignment right now?
Tell us what you're working on and get matched with a tutor who explains the reasoning, not just the answer — free quote by email or WhatsApp.
Get a free quote