25 multiple-choice questions in College Board exam style. Each has four choices and a full explanation of why each option is right or wrong.
Use t whenever σ is unknown. Almost every mean problem on the exam is a t-procedure with df = n − 1.
Decide paired vs. two-sample first. Matched subjects ⇒ one-sample t on differences; independent groups ⇒ two-sample t.
Standard error is s/√n. Interpret intervals about the population mean, in context.
Never accept H₀ (fail to reject it) and read the explanations even when you got it right.