5.9: Computer software for sample size calculations

Last updated
Save as PDF

Page ID: 13610

\( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \) \( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)\(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\) \(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\)\(\newcommand{\AA}{\unicode[.8,0]{x212B}}\)

Most of the formulae given in this chapter are simple enough to do by hand, with the aid of a simple calculator. However, computer software is also available to carry out some of these calculations. This can be particularly helpful when a large number of cal- culations need to be carried out, for example, to explore sample size requirements for different outcomes or under different assumptions, or to produce power curves. Most statistical packages have some provision for sample size calculations. Here we mention three packages which readers may find helpful when planning field trials.

The sampsi command in Stata allows the user to obtain the required sample size for the comparison of means or proportions. Alternatively, if the chosen sample sizes are entered, the user can determine the power that these will provide. The command allows for different sample sizes in the two trial arms. The sample size formulae used by this package differ slightly from those presented in this book, but the results should be quite similar in most cases.

The POWER and GLMPOWER procedures in the statistical analysis program package SAS can handle sample size calculations for a range of situations, including survival analy- sis, as can the PASS module (a trial version of which is available at <http://www.ncss.com>).

A variety of free sample size calculators may be found on the Internet. These include the program PS which is described by Dupont and Plummer (1990) and available at <http://biostat.mc.vanderbilt.edu/wik...owerSampleSize>); and Open Epi which can be downloaded from <http://www.openepi.com/Downloads/Downloads.htm>.

Table 5.6 gives a spreadsheet which facilitates the calculation of the required size (number of clusters) for a cluster randomized trial, using the formulae given in Section 6 (as in Hayes and Moulton, 2009).

It is fairly straightforward to set up a spreadsheet, for example, in Excel, to apply any of the formulae given in this chapter. The freeware computer package Epi-Info has a useful component, called StatCalc, for calculating sample sizes for simple trials.

Table 5.6 Spreadsheet calculation of the number of clusters required in an unmatched cluster randomized trial. Table 5.6 shows the calculations, for some example situations, for (a) comparison of proportions and (b) comparison of rates. formulae are given which allow the calculation of required trial size for any (unmatched) cluster randomized trial in an Excel spreadsheet