https://joss.theoj.org/papers/10.21105/joss.03118

https://ctu-bern.github.io/presize/

Verbatim From the article – There are two approaches for each measure.

1. Firstly, based on a given sample size, e.g. for a retrospective data analysis, the precision of an expected measure can be calculated. Precision is expressed as the confidence interval around the measure. The level of confidence can be specified; the usual 95%-confidence interval (CI) is the default.
2. Secondly, based on a given precision (i.e. CI), the sample size can be calculated. This is mainly of use in the planning of prospective studies, where the aim is to estimate a measure of interest with enough confidence.

Proportion

prec_prop(p = 0.15, conf.width = 0.05, method = c("wilson"))

     sample size for a proportion with Wilson confidence interval. 

     p      padj        n conf.width conf.level       lwr       upr
1 0.15 0.1517077 783.4897       0.05       0.95 0.1267077 0.1767077

NOTE: padj is the adjusted proportion, from which the ci is calculated.


prec_prop(p = 4:8/20, conf.width = 0.05, method = c("wilson"))

     sample size for a proportion with Wilson confidence interval. 

     p      padj         n conf.width conf.level       lwr       upr
1 0.20 0.2011693  981.7281       0.05       0.95 0.1761693 0.2261693
2 0.25 0.2508324 1149.8753       0.05       0.95 0.2258324 0.2758324
3 0.30 0.3005949 1287.6200       0.05       0.95 0.2755949 0.3255949
4 0.35 0.3504120 1394.8294       0.05       0.95 0.3254120 0.3754120
5 0.40 0.4002604 1471.4388       0.05       0.95 0.3752604 0.4252604Code language: JavaScript (javascript)

Shiny Webapp

launch_presize_app()