Very good question - the data themselves are discrete, so cannot be normal distributed. However, the log odds ratios are roughly normal distributed, hence you can calculate the 95% confidence interval of the log odds ratio. Note - the odds ratio is NOT normal distributed - that’s why you have to do the log-transformation.
Hi, what if the z score is negative? Do u just remove the 1- from this equation? And how many brackets, since I can’t get the right answer with my stats homework.
Does not make sense ED50 =1.6 Because from the graph it's clear the ED 50 must be in the middle of (between the concentration of drug 0.000001 and 0.000003