362
Views
0
CrossRef citations to date
0
Altmetric
Statistical Practice

Binomial Confidence Intervals for Rare Events: Importance of Defining Margin of Error Relative to Magnitude of Proportion

ORCID Icon & ORCID Icon
Received 07 Jun 2023, Accepted 21 Apr 2024, Published online: 24 May 2024

Figures & data

Table 1 Sample size comparison for ϵR=0.4.

Table 2 Wald-based sample size comparison – fixed ϵ.

Table 3 Wald-based sample size comparison – variable ϵ.

Table 4 ϵR thresholds as per (4) for a{5,10}.

Table 5 Coverage & relative margin of error tolerances.

Fig. 1 CPr versus ϵ˜R for p*=p=101,α=0.01. Dashed (gray) line represents the nominal CPr value. Dot-dashed (red) lines represent the target CPr and ϵ˜R tolerances from . Sample size range is from n = 100, to n = 800, in steps of 20. Labels shown adjacent to data points depict the first n where both CPr and ϵ˜R requirements are satisfied. Additional data labels are referred to within the text.

Fig. 1 CPr versus ϵ˜R for p*=p=10−1,α=0.01. Dashed (gray) line represents the nominal CPr value. Dot-dashed (red) lines represent the target CPr and ϵ˜R tolerances from Table 5. Sample size range is from n = 100, to n = 800, in steps of 20. Labels shown adjacent to data points depict the first n where both CPr and ϵ˜R requirements are satisfied. Additional data labels are referred to within the text.

Fig. 2 CPr versus ϵ˜R for p*=p=102,α=0.05. Dashed (gray) line represents the nominal CPr value. Dot-dashed (red) lines represent the target CPr and ϵ˜R tolerances from . Sample size range is from n= 1000, to n= 8000, in steps of 200. Labels shown adjacent to data points depict the first n where both CPr and ϵ˜R requirements are satisfied. Additional data labels are referred to within the text.

Fig. 2 CPr versus ϵ˜R for p*=p=10−2,α=0.05. Dashed (gray) line represents the nominal CPr value. Dot-dashed (red) lines represent the target CPr and ϵ˜R tolerances from Table 5. Sample size range is from n= 1000, to n= 8000, in steps of 200. Labels shown adjacent to data points depict the first n where both CPr and ϵ˜R requirements are satisfied. Additional data labels are referred to within the text.

Table 6 95% CI performance – p*=p=101, “small” n.

Table 7 95% CI performance – p*=p=101, “large” n.

Table 8 95% CI performance – p*=p=106, “small” n.

Table 9 95% CI performance – p*=p=106, “large” n.

Table 10 95% CI performance comparison – n.

Table 11 95% CI sample size comparison.

Table 12 ϵ˜R values corresponding to maintained coverage.

Table 13 95% CI performance – p*=p=102, small n.

Table 14 95% Wald CI – ϵR̂ comparison.

Table 15 Coverage comparison – fixed ϵ.

Table 16 Coverage comparison – ϵR=0.75.

Table 17 Comparison of classical and modern CI methods.