T-test: comparing two means

Dependent t-test (paired)

Independent t-test

Related data: same group different time, for instance: day 1 and day 2

Independent data: different group same time, for instance: female / male

Assumptions

IV: categorical with 2 groups (no more)

DV: interval/ratio

random sample

normal distribution N>30

T = mean paired difference - expected difference (0) / SE paired difference

test the difference

analyze

compare means

paired t-test

analyze

compare means

independent t-test

Values to interpret

compare means

t-test + significance

Is the difference meaningful?

Cohen's d

X1 - X2 / SD

Cohen’s d effect sizes:
0.2 = small
0.5 = medium
0.8 = large

Report:
A dependent t-test revealed that hygiene scores on day one (M = 1.65, SD = 0.64) were significantly higher than hygiene scores on day 3 (M = 0.98, SD = 0.71) when visiting a festival, t(122) = 10.59, p < .001, d = 0.95, 95% CI [0.55, 0.80].


We reject the null hypothesis that in the population there is no difference in hygiene scores between day 1 and day 3.

Values to interpret

Levene's test

variance in both groups are equal

Significant: no equal variance

Not significant: equal variance

mean difference + significance

Is the difference meaningful?

Cohen's d

(2*t) / square root df

Cohen’s d effect sizes:
0.2 = small
0.5 = medium
0.8 = large

Report:
An independent samples t-test was conducted to examine whether women scored hygher on hygiene than men after three days at a festival.
Levene’s test for equality of variances was significant, F = 13.27, p < .001. Results revealed that women had significantly higher hygiene scores after three days at a festival (M = 1.10, SD = 0.81) than men did (M = 0.83, SD = 0.54), t(115.58) = -2.21, p = .029, 95% CI [-0.51, -0.28], d = 0.41.
We reject the null hypothesis.