Answer. When is unknown and the sample is small.
Estimating by adds uncertainty, which the heavier tails of the -distribution account for.
Original questions written at exam standard. Work each one before you open the solution.
Answer. When is unknown and the sample is small.
Estimating by adds uncertainty, which the heavier tails of the -distribution account for.
Answer.
. One is lost because the sample mean was computed from the same data.
Answer.
SE , so .
Answer. Same shape but with heavier tails, so larger critical values.
The heavier tails reflect the extra variability from estimating . As the degrees of freedom rise, converges to the normal.
Answer. Do not reject .
The interval and the test are equivalent: a test rejects exactly when the corresponding interval excludes the claimed value.
Answer. The two measurements are not independent, and pairing removes between-subject variation.
Testing the differences isolates the treatment effect. Subjects may differ enormously from each other while their individual changes are measured precisely.