Skip to content

Mean Comparison & t-Test Lab

Two groups, one question: is the difference between the means large relative to how much such differences vary from sample to sample? Drag a mean handle (numerator) or a spread handle (denominator) and watch t move.
ObservationsGroup 1Group 2— mean┆ 95 % CI of the meandrag a point, the x̄ handle or the s handle · arrow keys when focused
t = -2.99, df = 18.0, p .008, difference -14.67

t = estimated difference ÷ its standard error

t =x̄₁ − x̄₂ = -14.67SE = 4.903= -2.99s₁²/n₁+s₂²/n₂=SE²(Welch)
x̄₁ − x̄₂ = -14.67, SE = 4.903, t = -2.99

Null distribution · t(18.0)

-4-2024t = -2.99p .008two-tailed, shaded = p
Null distribution t(18.0); t = -2.99; p .008

Synthetic scores (0–100, seeded). All statistics match the SPSS t-test rows (Student = equal variances assumed, Welch = not assumed) and SciPy; nothing here is an empirical finding.

Method: One-sample, paired and independent t-tests (Student and Welch rows as in SPSS); Hedges' g; mean-centred Levene test · Synthetic, seeded data. Illustrative, not an empirical finding.

Validation: t, df, p, CI and g match the SciPy reference (tests/unit) · numerically validated