Close

UCL Great Ormond Street Institute of Child Health

Home

Great Ormond Street Institute of Child Health

Section 2: Precision

If a sample size is too small the confidence interval obtained for the sample estimate of the population value will be too wide to be clinically useful. Increasing the sample size will always decrease the width of the confidence interval. One approach to the determination of a suitable sample size is to decide, prior to the study, how precisely we want to estimate the population value. In other words, how wide we would like the final confidence interval to be.

For all sample size calculation using this approach, the precision with which the estimate is desired needs to be given. 'Precision' is half the width of the confidence interval that it is expected to be obtained for the sample estimate after completing the study. The confidence interval is expected to be the sample estimate ± the precision specified. For all scenarios a larger precision leads to a smaller estimated sample size.

[2.1] Means

Below are links to excel spreadsheet (.xls) for calculating sample size for a single mean or for the difference between two means using the precision approach to sample size estimation.

Precision - difference between two means

[2.2] Proportions and Percentages

Below are links to excel spreadsheet (.xls) for calculating sample size for a single proportion or percentage, or for the difference between two proportions or percentages using the precision approach to sample size estimation.

Precision - difference between two percentages

[2.3] Rates

Below are links to excel spreadsheet (.xls) for calculating sample size for a single rate or for the difference between two rates using the precision approach to sample size estimation.

Precision - difference between two rates

[2.4] Measures of Agreement

Below are links to excel spreadsheet (.xls) for calculating sample size for numeric and categorical measures of agreement using the precision approach to sample size estimation.