Algorithm 2.

Posterior predictive power and coverage.

Input:
S simulated odds ratios ORm and geometric standard errors SEm for quantification methods m = 1, … ,M.
Level α at which to assess significance of statistical testing and to compute 100(1 − α)% confidence intervals.
Specification of a one or two-sided alternative HA to the null hypothesis H0 : OR = 1.
Reference odds ratio OR.
Output: Statistical powerm to accept the alternative hypothesis, and confidence interval coveragem of the reference odds ratio.
1:	if HA : OR ≠ 1 then
2:	α ← α2
3:	end if
4:	for m = 1 to M do
5:	powerm ← 0
6:	coveragem ← 0
7:	for s = 1 to S do
8:	lower ← 0	▻ Initialize interval lower
9:	upper ← ∞	▻ and upper bounds.
10:	if HA : OR ≠ 1 or HA : OR > 1 then
11:	$\mathit{lower}\leftarrow O{R}_m^{(s)}/{(S{E}_m^{(s)})}^{z_{1-\alpha }}$
12:	end if
13:	if HA : OR ≠ 1 or HA : OR < 1 then
14:	$\mathit{upper}\leftarrow O{R}_m^{(s)}\times {(S{E}_m^{(s)})}^{z_{1-\alpha }}$
15:	end if
16:	A ← {x : lower < x < upper}
17:	powerm ← powerm + 1Ā (1)/S
18:	coveragem ← coveragem + 1A(OR)/S
19:	end for
20:	end for