We explain the difference between the scientific use of the terms accuracy and precision.
We show how to augment estimates of population means, standard deviations and proportions
with indications of their accuracy, known as confidence intervals. These ideas are extended
to the comparison of two populations. The closely related procedures of hypothesis testing,
that is assessing whether sample results are consistent with, or evidence against, some
hypothesis about the population are described. The construction of prediction intervals
and statistical tolerance intervals, for individual values rather than the population mean,
is explained.
|