General Notes
Hypotheses: H0: mu2 - mu1 = delta Ha: mu2 - mu1 <> delta (two-sided hypothesis)
Ha: mu2 - mu1 > delta or Ha: mu2 - mu1 < delta (one-sided alternatives)
If delta = 0, then H0: mu1 = mu2
Test statistic: t statistic
P-value: the one- or two-tailed probability of a t-value at least as large in absolute value as the observed
Conclusion: Report P-value; reject H0 at alpha level if P <= alpha.
Also recommend reporting the standard error and the interval estimate.
Assumptions
The observations from both samples are randomly drawn from normally distributed populations.
Diagnostic: Normal probability plot with a test for normality.
The variances of both samples are equal.
Diagnostic: Levene's test
References