Fit parallel model to reference standard and test sample pair:
Log (RP) = ()
2. Back-transform using the same log base as was used for the dose:
RP = 2 Log 2 (RP)
RP = e Log e (RP)
RP = 10 Log10 (RP)
Confidence interval for relative potency
The variance, s2, of the estimate of Log (RP) is calculated from the variances and covariance of the model parameter estimates and using the Delta method.
The 95% confidence interval for Log (RP) can then be calculated as:
Log (RP) ± t(n- 5, 0.025)√ s2
where t(n- 5, 0.025) is the 97.5% percentile of the t-distribution with n degrees of freedom; n is the total number of data points in the Test and Reference samples.
The 95% confidence interval for RP is obtained by back-transforming the interval for Log RP using the same log base as was used for the dose.
) 
and 
using the Delta method.