Quick answer: In jamovi: T-Tests → Independent Samples T-Test, move your numeric outcome into Dependent Variable and your two-level group into Grouping Variable — done. The real skill is the check-list around it: pick the right t-test variant, verify assumptions (normality, equal variances), and read the effect size, not just p. This walkthrough uses a worked example; the underlying logic is in our hypothesis testing guide.
Which t-test do I need?
| Design | Menu choice |
|---|---|
| Two unrelated groups (drug vs placebo) | T-Tests → Independent Samples T-Test |
| Same people twice (pre-test, post-test) | T-Tests → Paired Samples T-Test |
| One group vs a known value (IQ vs 100) | T-Tests → One Sample T-Test |
Worked example: two teaching methods
40 students, exam scores, method A vs method B. Setup: Scores → Dependent Variable, Method → Grouping Variable. Tick Mean difference and Effect size (Cohen’s d) under Additional Statistics, and Normality + Homogeneity test (Levene’s) under Assumptions. Suppose output: A = 72.1 (SD 6.8), B = 66.4 (SD 7.5), t(38) = 2.46, p = 0.019, d = 0.78.
Reading it in order: the difference is 5.7 points with a 95% CI that excludes zero; p = 0.019 < 0.05 rejects “no difference” (what that means exactly: our p-value explainer); d = 0.78 is a medium-to-large effect — the finding matters beyond the p-value. The APA sentence writes itself: “Method A students scored higher (M = 72.1, SD = 6.8) than Method B (M = 66.4, SD = 7.5), t(38) = 2.46, p = .019, d = 0.78.”
The two assumption checks that matter
- Normality (Shapiro-Wilk): p > .05 is what you want. T-tests tolerate moderate violation with n ≥ 30 per group; for clearly skewed data use the Non-parametric pair (Mann-Whitney U) from the same menu.
- Equal variances (Levene’s): if p < .05, untick Student’s and tick Welch under Variance — jamovi prints the corrected row; degrees of freedom shift to decimals, which is expected, not an error.
Full first-session context (data types, variable setup, exporting) is in our jamovi first-analysis guide, and the biological decision logic behind test choice in hypothesis testing in biology. For one-to-one practice with an instructor checking your assumptions, Ampersand Academy teaches statistics with jamovi one-to-one.
Frequently asked questions
How do I run an independent t-test in jamovi?
Analyses ribbon, T-Tests, Independent Samples T-Test. Move the numeric outcome to Dependent Variable and the two-level grouping variable to Grouping Variable. Results appear immediately with t, df and p.
Should I use Student’s or Welch’s t-test?
Run Levene’s test under the Assumptions section. If it is significant (unequal variances), use Welch’s – in jamovi just tick the Welch option. With similar variances and similar group sizes, Student’s is fine.
How do I interpret Cohen’s d in the jamovi output?
d is the group difference in standard-deviation units: 0.2 is small, 0.5 medium, 0.8 large. It communicates how big the effect is, which the p-value alone never tells you.
What if my data are not normal?
With roughly 30 or more per group, t-tests are robust to moderate non-normality. For small, clearly skewed samples, use the Mann-Whitney U (independent) or Wilcoxon rank (paired) tests from the same T-Tests menu.
Why does my grouping variable show only one group in the t-test?
The grouping variable must be a factor with exactly two levels. Check its measure is Nominal and that both levels actually appear in the data – filter rows or typo-coded levels are the usual culprit.