384 followers
@arthur_alb1 Woods et al show how to derive the SE of the reference arm if you have the third "active-active" contrast - equation (7): https://t.co/jKWae2qnQ6 If you don't have the third SE (but you have sample sizes), you can approximate assuming propor