Performance of CIM and FMA for outcomes generated from a Poisson regression with β = 0 (CR: 95% CI coverage rate, LER: left error rate, RER: right error rate, LCI: length of 95% CIs, AB: absolute bias, UL: upper limit) for the null model (β = 0) using doses generated under Scenario 1.

Results are summarized from 5,000 tests with sample size of 1,000 using different pre-selected 2DMC realizations of a cohort dose vector for each of the 5,000 tests. CR, RER, and LER were defined as the percentage of times the true value of the slope of the null dose response (β = 0) was inside the 95% CI, and below the lower limit, or above the upper limit, respectively.