Skip to main content

Table 1 Showing the level of cumulated type 1-error risk, if a threshold of 5% is applied constantly at each sequential significance testing, on an accumulating number of trial participants

From: Trial Sequential Analysis in systematic reviews with meta-analysis

Number of statistical significance tests The cumulated type 1-error risk in %
1 5%
2 8%
5 14%
20 25%
100 37%
Infinitely many 100%
  1. The resulting type 1-error risk will be larger than the nominal 5%, if a decisison is made to stop the inclusion of participants when P <0.05 and to continue when P ≥0.05 [22, 24]